Datasets:

TaylorAI

/

pubmed_commercial

like 12

PMC010xxxxxx/PMC10241235.txt

==== Front Endocrinol Diabetes Metab Case Rep Endocrinol Diabetes Metab Case Rep EDM Endocrinology, Diabetes & Metabolism Case Reports 2052-0573 Bioscientifica Ltd Bristol 37185284 10.1530/EDM-22-0384 EDM220384 Paediatric Male Black - African Tanzania, United Republic of Adrenal Adrenal Unique/Unexpected Symptoms or Presentations of a Disease Unique/Unexpected Symptoms or Presentations of a Disease Novel likely pathogenic variant in NR5A1 gene in a Tanzanian child with 46,XY differences of sex development, inherited from the mosaic father R K Damji and others http://orcid.org/0000-0003-4536-4804 Damji Rahim Karim 1 Alimohamed Mohamed Zahir 2345 Claahsen-van der Grinten Hedi L 6 Westra Dineke 7 Hamel Ben 7 1 Department of Paediatrics, Regency Medical Centre, Dar es Salaam, Tanzania 2 Shree Hindu Mandal Hospital, Dar es Salaam, Tanzania 3 Department of Biochemistry, Muhimbili University of Health and Allied Sciences, Dar es Salaam, Tanzania 4 Department of Genetics, University Medical Center Groningen, Groningen, The Netherlands 5 Tanzania Human Genetics Organization, Dar es Salaam, Tanzania 6 Department of Paediatric Endocrinology, Amalia Children’s Hospital, Radboud University Medical Center, Nijmegen, The Netherlands 7 Department of Human Genetics, Radboud University Medical Center, Nijmegen, The Netherlands Correspondence should be addressed to R K Damji; Email: rahimdamji50@gmail.com 06 4 2023 01 4 2023 2023 2 22-038425 10 2022 06 4 2023 © the author(s) 2023 the author(s) https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License. Summary Pathogenic variants in the nuclear receptor subfamily 5 group A member 1 gene (NR5A1), which encodes steroidogenic factor 1 (SF1), result in 46,XY and 46,XX differences of sex development (DSD). In 46,XY individuals with a pathogenic variant in the NR5A1 gene a variable phenotype ranging from mild to severe is seen, including adrenal failure, testis dysgenesis, androgen synthesis defects, hypospadias and anorchia with microphallus and infertility. We report the clinical, endocrinological and genetic characteristics of a patient with 46,XY DSD with a novel likely pathogenic missense variant in the NR5A1 gene. A retrospective evaluation of the medical history, physical examination, limited endocrinological laboratory analysis and genetic analysis with DSD gene panel testing was performed. A 1.5-month-old individual was referred with ambiguous genitalia. The karyotype was 46,XY. The endocrinological analyses were within normal male reference including a normal response of cortisol within an adrenocorticotropic hormone test. A novel heterozygous missense variant c.206G>C p.(Arg69Pro) in the NR5A1 gene was detected. This variant was present in mosaic form (~20%) in his unaffected father. Because another missense variant at the same position and other missense variants involving the same highly conserved codon have been reported, we consider this NR5A1 variant in this 46,XY DSD patient as likely pathogenic in accordance with the ACMG/AMP 2015 guidelines causing ambiguous genitalia but no adrenal insufficiency. This variant was inherited from the apparently unaffected mosaic father, which might have implications for the recurrence risk in this family. Learning points The importance of performing trio (patient and parents) sequencing is crucial in pointing out the origin of inheritance. In a 46,XY differences of sex development patient, a normal adrenal function does not rule out an NR5A1 mutation. With the support of a dedicated overseas institute partnership, we could solve this complex clinical case by molecular diagnosis in a resource-limited setting. Patient Demographics Paediatric Male Black - African Tanzania, United Republic of Clinical Overview Adrenal Adrenal Publication Details Unique/unexpected symptoms or presentations of a disease April 2023 ==== Body pmcBackground Differences of sex development (DSD) have been defined as ‘congenital conditions in which the development of chromosomal, gonadal, or anatomical sex is atypical’. Therefore, the term DSD constitutes a spectrum of disorders that affect the genitourinary tract and the endocrine-reproductive system. Based on chromosome karyotype, DSD are usually classified into 46,XX DSD, 46,XY DSD and sex chromosome DSD (1). Heterozygous pathogenic NR5A1 variants account for 10–20% of 46,XY DSD cases (2). NR5A1 gene encodes for Steroidogenic factor 1 (SF1), which functions as a transcription factor for sex determination as well as regression of Mullerian structures (3). Since the description of the first 46,XY DSD patient with adrenal insufficiency and a pathogenic NR5A1 variant (4), the spectrum of phenotypes associated with pathogenic NR5A1 variants has greatly expanded. It became clear that adrenal insufficiency is a fairly rare feature of 46,XY DSD (2). Up to 2019, more than 180 pathogenic NR5A1 variants have been reported (3). There is no clear phenotype–genotype correlation and there is wide phenotypic variability between and within families (2). The spectrum of genital anomalies in 46,XY DSD patients carrying NR5A1 variants comprises partial to complete gonadal dysgenesis with female external genitalia, genital ambiguity, penoscrotal hypospadias, micropenis, cryptorchidism, anorchia and male factor infertility (5). Heterozygous pathogenic NR5A1 variants can also cause different types of ovarian insufficiency in 46,XX individuals (6). Here, we describe clinical and molecular findings in an individual with 46,XY DSD born with ambiguous genitalia and with a normal adrenocortical function and in whom a novel heterozygous NR5A1 variant was identified, inherited from the apparently unaffected mosaic father. Case presentation The infant was 1.5 months old when he first presented at our centre with ambiguous genitalia. He was born at term to nonconsanguineous parents, after an uneventful pregnancy with a birth weight of 3000 g. He was the only child in the family. As far as parents could tell, there were no cases of ambiguous genitalia and sub- or infertility in their families. Father was not known for hypospadias and from history, no complaints of adrenal insufficiency. Physical examination of the child revealed a small phallus (stretched length 2 cm), proximal hypospadias with the curvature of the phallus and single perineal orifice, bifid scrotum and bilateral palpable gonads in the scrotal sac (1 mL bilaterally) (Fig. 1A and B). No extra-genital anomaly was detected. Otherwise, the physical examination was unremarkable. Figure 1 (A) Bifid scrotum, small phallus. (B) Curved phallus, single perineal opening, proximal hypospadias and visible/palpable gonads. Investigation Karyotyping revealed a normal 46,XY karyotype and male gender was assigned. Mullerian derivatives could not be visualized on pelvic ultrasonography. Adrenal insufficiency was excluded with a standard dose adrenocorticotropic hormone (ACTH) stimulation test, which revealed a normal peak cortisol level. Other laboratory results were all within the reference levels of age for 46,XY individuals. The testosterone level at the age of 6 weeks (mini puberty) is expected to be higher in normal infants; however, testosterone here is low as expected in a child with gonadal dysgenesis and insufficient testosterone leading to incomplete virilization (Table 1). Table 1 Laboratory tests and their results at the age of 6 weeks. Laboratory tests Results Reference values* Follicle stimulating hormone, IU/L 4.6 0.1–11.3 Luteinising hormone, IU/L 0.1 0.02–8.0 Testosterone, nmol/L 0.1 0.03–6.14 Dihydrotestosterone, ng/mL 0.18 0.06–0.30 17-OH progesterone, ng/mL 1.3 0.4–2.0 Androstenedione (LC-MS), ng/mL 1 0.6–2.0 Sodium, mmol/L 135 135–145 Potassium, mmol/L 4.4 4.1–5.3 Random blood glucose, mmol/L 5.1 3.3–7.8 Peak cortisol level, µg/dL 23.3 >18 *Male reference values. After informed consent was obtained, blood samples were collected from the patient and his parents and shipped to Genome Diagnostics of the Department of Human Genetics, Radboud University Medical Center, Nijmegen, The Netherlands, for trio exome sequencing, as described previously (7). In short, exome enrichment was performed with the Agilent SureSelectQXT Human All Exon v5 Kit. Read alignment was done with BWA, and variant calling with GATK (SNVs) and CoNIFER (CNVs). After that, variants were annotated using an in-house developed pipeline. A bioinformatic filter for our ‘DSD’ gene panel (version DG-3.0) was applied and variants were selected and prioritized (information about the prioritization is available on request). Panel content and gene coverage can be found on the website of the Department of Human Genetics of the Radboud University Medical Center (https://order.radboudumc.nl/en/products/wes-disorders-differences-of-sex-development-dsd-primary-adrenal-insufficiency). A heterozygous variant in the NR5A1 gene was found in the patient: NR5A1 Chr9(GRCh37):g.127265396C>G NM_004959.5:c.206G>C p.(Arg69Pro). This exact NR5A1 variant has not been reported before and is not present in control populations of the Genome Aggregation Database (www.gnomad.org; v.2.1.1), but another missense change at the same nucleotide 206 and one at the same highly conserved amino acid arginine at position 69 in NR5A1 has been published in multiple patients (Table 2; 8, 9, 10) and functional analysis of these variants showed decreased protein expression (8). Therefore, we consider this variant as likely pathogenic according to the ACMG/AMP 2015 guidelines (11) (Table 3). With exome sequencing, the same variant was found in the father in mosaic form (~17%), which was confirmed with Sanger sequencing (present in ~20% of his DNA), whilst in the mother this variant was not identified. No other tissue was tested in the father to determine the mosaic status in there. Table 2 Previously reported NR5A1 variants involving amino acid 69. Reference Patient Nucleotide change Amino acid change Variant class Clinical data Family This study Present case c.206G>C p.Arg69Pro Likely pathogenic Bifid scrotum, small penis, curved phallus, single perineal opening, proximal hypospadias From mosaic father not exhibiting this phenotype Na et al. (8) Patient 6 c.206G>A p.Arg69His Pathogenic Bilateral dysplastic inguinal testes From father with likely normal phenotype Na et al. (8) Patient 5 c.205C>G p.Arg69Gly Pathogenic Bifid poorly developed scrotum, small penis, inguinal testes From mother Costanzo et al. (9) Patient from Family 5 c.206G>A p.Arg69His Pathogenic No data Familial case, so either from mother or father Kim et al. (10) Patient 15 c.205C>G p.Arg69Gly Pathogenic Clitoromegaly, atrophic testes in pelvic cavity Sporadic and no segregation analysis was performed Table 3 Applied ACMG/AMP 2015 criteria for the detected NR5A1 variant. Rule Description Evidence/justification to support the use of rule PM1 Located in a mutational hotspot and/or critical and well-established functional domain without benign variation Located in the Zinc finger, nuclear hormone receptor-type protein domain of Steroidogenic factor 1, in which many pathogenic variants have been described; no non-synonymous variants present in control populations in the Genome Aggregation Database (gnomAD v2.1.1) between amino acids 12 and 82. PM2 Absent from controls (or at low frequency if recessive) Absent from all control populations in the Genome Aggregation Database (gnomAD v2.1.1)) PM5 Novel missense at an amino acid residue where a different amino acid has been determined to be pathogenic has been seen before - c.206G>A p.(Arg69His) described in (8) and (9) - c.205C>G p.(Arg69Gly) described in (8) and (10) PP3 Multiple lines of computational evidence support a deleterious effect on the gene or gene product Align GVGD (v2007): Class C65 PolyPhen2: HDivPred: probably damaging (score: 1). HVarPred: probably damaging (score: 1) SIFT (v6.2.0): DELETERIOUS (score: 0.00, median: 2.98) MutationTaster (v2021): Deleterious. Treatment No pharmacological treatment was offered to this child since all the biochemical investigations were within normal range. Surgical correction of his external genitalia was advised. Outcome and follow-up Periodic follow-up was advised to monitor any new endocrine-related problems. The patient remained asymptomatic 12 months post-diagnosis. Discussion We describe a 46,XY DSD patient born with ambiguous genitalia with a novel heterozygous variant in the NR5A1 gene, which was inherited from the mosaic unaffected father. Pathogenic variants in NR5A1 are associated with 46,XY sex reversal 3 (OMIM 612965). In the majority of cases, these are heterozygous variants. The NR5A1 variant in our patient has not been reported before, but another missense change at the same nucleotide position 206 has been published in multiple patients (Table 2). More variants involving changes in the highly conserved amino acid arginine at position 69 in 46,XY DSD patients are described in the literature (Table 2). This variant is located in a functional domain, not present in the control population of the Genome Aggregation Database (www.gnomad.org; 12) and multiple lines of computational evidence support a deleterious effect on the gene product. Therefore, we consider this variant as likely pathogenic according to the ACMG/AMP 2015 guideline (11). As adrenal insufficiency is described in some but not all patients with NR5A1 mutations, adrenal testing is recommended. Our patient had a normal response to ACTH excluding adrenal insufficiency. Mosaicism of NR5A1 variants has been described before in two non-affected fathers (13, 14). Both fathers had normal genitalia and the mosaicism was confirmed by using another sequencing technique (13) or confirmed in another tissue (14). Identification of mosaicism in a parent has potential consequences for genetic counselling and the recurrence risk when this variant is also present in the germ cells. If present in a high percentage of germ cells, there is a high risk of having other affected sons (46,XY). There might also be a risk of primary ovarian insufficiency (POI) for affected daughters (46,XX). In about 20–30% of 46,XY DSD cases, NR5A1 variants are inherited from non-affected or later affected mothers (3, 15) and less frequently non-mosaic asymptomatic and sometimes symptomatic (hypospadias) fathers transmit the NR5A1 variant to their children (3). Since the first 46,XY DSD patient with a pathological NR5A1 variant was described by Achermann et al. (4), the spectrum of clinical presentation has evolved from the mildest to the most severe form including adrenal insufficiency as a potentially life-threatening complication. These cases, as well as their families, need a multidisciplinary team approach, with a focus on the child’s interest. Major aspects to be discussed with the families include gender assignment, endocrine as well as urological and sexual function. Long-term outcomes, risks of infertility and germ cell tumour also need to be discussed. Genetic counselling has to be part of the management, particularly in this case where the father is a mosaic carrier of the likely pathogenic NR5A1 variant. In conclusion, we describe a 46,XY DSD patient with ambiguous genitalia and normal adrenal function with a heterozygous likely pathogenic, novel missense variant, c.206G>C p.(Arg69Pro) in the NR5A1 gene, which was inherited from a mosaic apparently unaffected father. Long-term clinical and hormonal follow-up in this patient is needed to assess the gonadal and adrenal function. Declaration of interest The author declares that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported. Funding This study did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector. Patient consent Written informed consent was obtained from the parents of the patient for the publication of this case report and images. A copy of the written consent is available for the Editor of this journal. Patient’s perspective The parents expressed how the disease affected the child and the family since there is no similar history in the family. This will further create a family and social stigma since he was initially raised as a female until the karyotype results confirmed the male gender. I am very grateful to both local and international health professionals for making this medical mystery solved genetically. We have accepted the results and will raise him as a male. Author contribution statement R Damji and M Alimohamed developed the project design, described the case report and carried out a literature search and wrote the first draft of the manuscript. D Westra and B Hamel performed and described the molecular analysis and gave critical comments on the manuscript. H Claahsen-van der Grinten gave critical comments on the description of the case report and the interpretation of the data. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work. ==== Refs References 1 Hughes IA Houk C Ahmed SF Lee PA , LWPES Consensus Group & ESPE Consensus Group. Consensus statement on management of intersex disorders. Archives of Disease in Childhood 2006 91 554–563. (10.1136/adc.2006.098319)16624884 2 Martínez de La Piscina I Mahmoud RA Sauter KS Esteva I Alonso M Costa I Rial-Rodriguez JM Rodríguez-Estévez A Vela A Castano L , et al. Variants of STAR, AMH and ZFPM2/FOG2 may contribute towards the broad phenotype observed in 46,XY DSD patients with heterozygous variants of NR5A1. International Journal of Molecular Sciences 2020 21 8554. (10.3390/ijms21228554) 3 Fabbri-Scallet H de Sousa LM Maciel-Guerra AT Guerra-Junior G & de Mello MP . Mutation update for the NR5A1 gene involved in DSD and infertility. Human Mutation 2020 41 58–68. (10.1002/humu.23916)31513305 4 Achermann JC Ito M Ito M Hindmarsh PC & Jameson JL . A mutation in the gene encoding steroidogenic factor-1 causes XY sex reversal and adrenal failure in humans. Nature Genetics 1999 22 125–126. (10.1038/9629)10369247 5 Pedace L Laino L Preziosi N Valentini MS Scommegna S Rapone AM Guarino N Boscherini B De Bernardo C Marrocco G , et al. Longitudinal hormonal evaluation in a patient with disorder of sexual development, 46,XY karyotype and one NR5A1 mutation. American Journal of Medical Genetics 2014 164A 2938–2946. (10.1002/ajmg.a.36729)25160005 6 Camats N Pandey AV Fernandez-Cancio M Andaluz P Janner M Toran N Moreno F Bereket A Akcay T Garcıa-Garcıa E , et al. Ten novel mutations in the NR5A1 gene cause disordered sex development in 46, XY and ovarian insufficiency in 46, XX individuals. Journal of Clinical Endocrinology and Metabolism 2012 97 E1294–E1306. (10.1210/jc.2011-3169)22549935 7 Haer-Wigman L van Zelst-Stams WA Pfundt R van den Born LI Klaver CC Verheij JB Hoyng CB Breuning MH Boon CJ Kievit AJ , et al. Diagnostic exome sequencing in 266 Dutch patients with visual impairment. European Journal of Human Genetics 2017 25 591–599. (10.1038/ejhg.2017.9)28224992 8 Na X Mao Y Tang Y Jiang W Yu J Cao L & Yang J . Identification and functional analysis of fourteen NR5A1 variants in patients with the 46 XY disorders of sex development. Gene 2020 760 145004. (10.1016/j.gene.2020.145004) 9 Costanzo M Warman DM Guercio G Ciaccio M De Dona V Marino R Ramirez P Perez Garrido N Baquedano MS Berensztein E , et al. Mutations in the NR5A1 gene in patients with 46,XY disorders of sex development (DSD): high frequency of familial multi-generational occurrence. Hormone Research 2012 78 (Supplement 1 ) 122. Available at: http://www.karger.com/Article/Pdf/451040 10 Kim JH Kang E Heo SH Kim GH Jang JH Cho EH Lee BH Yoo HW & Choi JH . Diagnostic yield of targeted gene panel sequencing to identify the genetic etiology of disorders of sex development. Molecular and Cellular Endocrinology 2017 444 19–25. (10.1016/j.mce.2017.01.037)28130116 11 Richards S Aziz N Bale S Bick D Das S Gastier-Foster J Grody WW Hegde M Lyon E Spector E , et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genetics in Medicine 2015 17 405–424. (10.1038/gim.2015.30)25741868 12 Karczewski KJ Francioli LC Tiao G Cummings BB Alföldi J Wang Q Collins RL Laricchia KM Ganna A Birnbaum DP , et al. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature 2020 581 434–443. (10.1038/s41586-020-2308-7)32461654 13 Philibert P Polak M Colmenares A Lortat-Jacob S Audran F Poulat F & Sultan C . Predominant Sertoli cell deficiency in a 46,XY disorders of sex development patient with a new NR5A1/SF-1 mutation transmitted by his unaffected father. Fertility and Sterility 2011 95 1788.e5–1788.e9. (10.1016/j.fertnstert.2010.11.035) 14 Tuhan H Anik A Catli G Onay H Aykut A Abaci A & Bober E . A novel mutation in steroidogenic factor (SF1/NR5A1) gene in a patient with 46 XY DSD without adrenal insufficiency. Andrologia 2017 49 e12589. (10.1111/and.12589) 15 El-Khairi R & Achermann JC . Steroidogenic factor-1 and human disease. Seminars in Reproductive Medicine 2012 30 374–381. (10.1055/s-0032-1324720)23044873

PMC010xxxxxx/PMC10241305.txt

==== Front J Nematol J Nematol jofnem jofnem Journal of Nematology 0022-300X 2640-396X Sciendo 37284001 jofnem-2023-0025 10.2478/jofnem-2023-0025 Research Paper Impact of non-fumigant nematicides on reproduction and pathogenicity of Meloidogyne enterolobii and disease severity in tobacco Alam Md Shah Khanal Churamani ckhanal@clemson.edu Roberts Joseph Rutter William Clemson University, Department of Plant and Environmental Sciences, 105 Collings St., Clemson, SC 29634 Clemson University, Department of Plant and Environmental Sciences, Pee Dee Research and Education, Florence, 2200 E Pocket Rd, Florence, SC 29506 USDA-ARS, US Vegetable Laboratory, 2700 Savannah Hwy., Charleston, SC 29449 This paper was edited by Shaun D Berry. 2 2023 5 6 2023 55 1 202300259 3 2023 © 2023 Md Shah Alam et al., published by Sciendo 2023 Md Shah Alam et al., published by Sciendo https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. Abstract Meloidogyne enterolobii is a highly aggressive quarantine pathogen which threatens the multibillion-dollar tobacco industry and is not manageable with the currently available management methods in tobacco. There is currently no known host plant resistance in tobacco and previous studies have shown that the lower level of the currently recommended rate of non-fumigant nematicides does not provide satisfactory management of M. enterolobii. The current study was conducted with the hypothesis that M. enterolobii can be better managed using a single soil application of the maximum allowed rate of non-fumigant nematicides. Treatments involved three non-fumigant chemical nematicides (oxamyl, fluopyram, and fluensulfone), a biological nematicide derived from Burkholderia, and a non-treated control. Fluensulfone significantly suppressed the nematode reproduction relative to the control, the suppression being 71% for eggs and 86% for the second stage juveniles (J2). Fluopyram also suppressed nematode reproduction, although this was statistically insignificant, with the suppression being 26% and 37% for eggs and J2, respectively. Oxamyl significantly suppressed J2 (80%), but not eggs (50%) in relation to the control. The most significant reduction of disease severity was achieved by the application of fluensulfone (64%), followed by oxamyl (54%) and fluopyram (48%). Except for fluensulfone, which significantly reduced the root biomass, none of the nematicides significantly impacted root and shoot biomass. The biological nematicide did not significantly affect nematode reproduction, pathogenicity, or disease severity. The results from the current study suggest that while the non-fumigant nematicides provided a good level of the nematode suppression, more research is needed to improve the efficacy of non-fumigant nematicides through employing better application methods or finding better chemistries. Keywords biological Burkholderia fluensulfone fluopyram management Meloidogyne enterolobii nematicide non-fumigant oxamyl tobacco ==== Body pmcThe guava root-knot nematode (Meloidogyne enterolobii Yang and Eisenback, 1983) is a highly aggressive species of Meloidogyne that attacks a wide range of economically important crops. This nematode has been reported to severely infect guava and tobacco plants in Brazil (Gomes et al., 2008). According to Schwarz et al. (2020), M. enterolobii has been found in several tobacco producing fields in North Carolina, suggesting a growing threat to tobacco production. Because M. enterolobii is a quarantine species and is not manageable with currently available nematode management practices, several states in the US have imposed a quarantine on the movement of soil and planting materials from nematode-infested areas (Ye et al., 2021). Studies are needed to develop better nematode management programs, not only because M. enterolobii poses a threat to the 15-billion-dollar tobacco industry in the US, but also to protect the well-being of farmers who have been farming tobacco for generations as their only source of income. The current nematode management methods in tobacco are limited to the use of fumigants. Although effective in lowering the initial soil population of nematodes, fumigants do not provide season-long protection (Khanal et al., 2021). Additionally, they are cumbersome to apply and pose adverse impacts for the environment and human health (Khanal and Desaeger, 2020). Because fumigants work against a large range of organisms in the soil, they are also detrimental to beneficial microbiomes that play essential roles in maintaining soil health. In a quest to find safer yet effective alternatives to fumigants, some industries have developed and marketed non-fumigant chemical and biological nematicides. Non-fumigant chemical and biological nematicides are less harmful to the environment, have little to no off-target effect, and are easier to apply. Commonly available non-fumigant chemical nematicides include oxamyl (Vydate® L, Corteva Agriscience, Indianapolis, IN), fluopyram (Velum® Prime, Bayer CropScience, Research Triangle Park, NC), and fluensulfone (Nimitz®, ADAMA Agricultural Solutions Ltd., Raleigh, NC), while some biological nematicides include Burkholderia metabolites (Majestene®, Marrone Bio Innovations, Davis, CA) and Melocon (Certis® USA L.L.C., Columbia, MD). The efficacy of non-fumigant nematicides has been assessed for several crops; however, their efficacy on tobacco has not been studied extensively. A recent study of ours, which employed a single soil application of the lower level of recommended rate of oxamyl, fluopyram, fluensulfone, and Burkholderia metabolites found up to 99% suppression of M. enterolobii reproduction by the chemical non-fumigants while the effects of biological nematicide were similar to that of non-treated control (Alam et al., 2022). Because initial population of M. enterolobii even at a very low level can be highly damaging by the end of crop growing season and the non-fumigants at the lower level of currently recommended rate did not completely suppress the nematode reproduction, we hypothesize that a higher rate of nematicides could be helpful in further suppressing the nematode reproduction, pathogenicity as well as the disease severity. Therefore, the main objective of the current study was to evaluate the efficacy of a higher rate of non-fumigant nematicides on reproduction and pathogenicity of M. enterolobii as well as the disease severity in tobacco. Materials and Methods Nematode inoculum preparation Nematode eggs were extracted from pure cultures of M. enterolobii originated from South Carolina and maintained on a tomato (Solanum lycopersicum L., cv. Rutgers, Seedway, Hall, NY) in a growth room. Galled roots of 3–4 months old tomatoes were cut into 3–5 cm pieces and agitated in 0.6% NaOCl for 4 min to dislodge the nematode eggs (Hussey and Barker, 1973). The extracted eggs served as inoculum, and inoculation was conducted within two hours of extraction. Establishment of experiment Experiments were established in a Biosafety Level 2 growth room to comply with the biosafety requirements. Plastic pots with a 15 cm top diameter were filled with 1.5 kg sandy loam soil steam sterilized for four cycles of 45 min at 123°C. A waiting period 24 hr was maintained between pot filling and planting of tobacco seedlings to allow the release of possible toxic gases formed during soil sterilization. The soil-filled pots were placed in secondary containment on a bench to avoid floor contamination. Each pot received an 8-week-old tobacco seedling (cultivar K346, Gold Leaf Seed Company, Hartsville, SC) with four to six true leaves, and each pot was inoculated with approximately 10,000 eggs of M. enterolobii. The aqueous suspension of inoculum was pipetted into three 0.5 cm diam. × 5 cm deep depressions arranged into a triangular pattern and 2 cm away from the crown region. Treatments included a single drench application of the maximum allowed rates of three non-fumigant chemical nematicides (oxamyl, fluopyram, and fluensulfone) and one Burkholderia-derived biological nematicide plus a control that received nematode inoculum but not nematicide (Table 1). Nematicides were applied with a pipette four days after egg inoculation in depressions prepared in a similar way as described for egg inoculation. An exception with nematicide application was made for pots receiving fluensulfone, where tobacco seedlings were transplanted in egg inoculated soil a week after nematicide application because of possible phytotoxicity as stated on the label. Pots receiving nematicides were watered with 100 mL water a few minutes before nematicide application and with 50 mL water after application to facilitate nematicide mobility in the soil. The experiment was established as a randomized block design with five replications of each treatment. A 12-hr photoperiod was provided by four metal halide bulbs (1,000 W) hanging approximately 3 m above the table. Standard watering, fertilization, and insect management practices were conducted. The entire experiment was repeated once. The average temperature for the first experiment was 27.5°C (maximum = 30.8°C, minimum = 20.6°C) with relative humidity of 45.6% (maximum = 48.3%, minimum = 43.7%). Similarly, the average temperature for the second experiment was 22.7°C (maximum = 29°C, minimum = 20.6°C) with a relative humidity of 47.2% (maximum = 50.1%, minimum = 45%). Table 1. Non-fumigant and biological nematicides with their product application rates employed in the current study. Treatment Application rate (L/ha) Nematicide/pot (μl) Oxamyl 4.68 16.55 Fluopyram 1.00 0.06 Fluensulfone 8.18 0.48 a Burkholderia 18.71 37.85 Untreated control - - a Burkholderia rinojensis strain A396 secondary metabolites. The experiments were terminated eight weeks after inoculation. Gall index (GI) data in a range of 0 -10 was collected for disease severity analysis (Bridge and Page 1980). Nematode eggs were extracted from whole root systems in each pot on the day of termination using the method of Hussey and Barker (1973) described above. Eggs were enumerated within 24 hr of extraction using a compound microscope (Martin Microscope Company, Easley, SC) at 40x magnification. Soil samples were placed in a walk-in cooler at 4°C, and second-stage juveniles (J2) were extracted from a 100 cm3 subsample of soil from each pot using the centrifugal flotation technique (Jenkins 1964). Soil extraction and enumeration of J2 were conducted within two weeks of experiment termination. The aboveground (shoot) and belowground (root) plant parts were dried in an incubator at 45°C for 2 weeks, and dry biomass was recorded to determine the pathogenicity of the nematode on tobacco. Data analysis Nematode reproduction (eggs and J2) and pathogenicity (shoot and root biomass) data were subject to one-way analysis of variance (ANOVA) using the Generalized Linear Mixed Model in JMP PRO 16.2 (SAS Institute, Cary, NC). Residual analysis was performed to remove outliers. Subsequently, data were assessed for normality and any non-normal data were log transformed to fulfill the assumptions of ANOVA. Because of the absence of treatment by experiment interactions, data from two experiments were combined for analysis. Nematicide treatment was used as a fixed effect, and replication was used as a random effect. Tukey's HSD (p ≤ 0.05) was used for post-hoc mean comparisons. Untransformed values are presented in the Results section and figures. Results Effects of nematicides on nematode reproduction Nematode reproduction was significantly impacted by the application of nematicides (p = 0.0005 for eggs and 0.0001 for J2). The effects of nematicides on nematode egg production are presented in Figure 1A. Across all treatments, the number of eggs/root system ranged from 2,280 to 320,400. Tobacco plants treated with fluensulfone had significantly lower nematode eggs relative to the control, the reduction being 71%. Oxamyl and fluopyram, respectively, suppressed 50% and 26% of nematode eggs; however, the suppressions were not significantly different from the control. The number of eggs in Burkholderia metabolites treated plants was statistically similar to those in the controls. Figure 1: Reproduction of M. enterolobii expressed as number of eggs per root system of tobacco (A) and as number of J2 per 100 cm3 of soil (B). Data were combined over two experiments and are means of 10 replications. Data were log transformed and the treatment means followed by a common letter are not significantly different according to Tukey's HSD test (P ≤ 0.05). The treatment means in the figure represent untransformed values. Burkholderia refers to Burkholderia rinojensis strain A396 secondary metabolites. The effects of nematicides on the soil population of J2 are presented in Figure. 1B. The number of J2/100 cm3 soil across all treatments ranged from 10 to 8,100. Oxamyl and fluensulfone significantly suppressed the number of J2 relative to the control, the suppression being 88% and 86%, respectively. Fluopyram suppressed 37% of J2, although the suppression was not significantly different from the control. Effects of nematicides on nematode pathogenicity Nematicide application had a significant effect on plant shoot (p < 0.0001) and root biomass (p < 0.0001) as presented in Figures 2A and 2B. The dry shoot biomass across all treatments ranged from 4.5 to 11.3 g. Fluopyram treated plants had a significantly more, 27%, shoot biomass relative to the control. Although not significantly different, oxamyl and fluensulfone treated plants had 15% and 6% greater shoot biomass relative to the control. Burkholderia metabolites treated plants had 10% lower shoot biomass, although the reduction was not significantly different from the control. Figure 2: Dry weights of tobacco shoot (A) and root (B) inoculated with M. enterolobii. The plant materials were dried at 45°C for 2 wk. Data were combined over two experiments and are means of 10 replications. Treatment means followed by a common letter are not significantly different according to Tukey's HSD test (P ≤ 0.05). Burkholderia refers to Burkholderia rinojensis strain A396 secondary metabolites. The root dry biomass across all treatments ranged from 0.6 to 3.4 g. Only fluensulfone had a significant impact on the plant root biomass compared to the control. Fluensulfone treated plants had 39% less root biomass relative to the control. Fluopyram treated plants had 2% greater root biomass while oxamyl treated plants had 18% lower root biomass relative to the control. Plants treated with Burkholderia metabolites had slightly only 0.2% greater shoot biomass than the controls. Effects of nematicides on disease severity Nematode disease severity was significantly impacted by the application of nematicides (p < 0.0001) as presented in Figure 3. The GI ranged from 0 to 7 across all treatments. Plants treated with oxamyl, fluopyram and fluensulfone had significantly lower GI compared to that of controls. The greatest suppression of disease severity was achieved by the application of fluensulfone which suppressed 64% galls relative to the control. The suppression of disease severity by oxamyl and fluopyram was, respectively, 54% and 48%, relative to the control. The disease severity of plants treated with Burkholderia metabolites was statistically similar to that of controls. Figure 3: Root gall index as influenced by the application of nematicides. Gall rating was conducted on a scale of 0 -10 (0 = no galls, and 10 = completely galled roots) as described by Bridge and Page (1980). Data were combined over two experiments and are means of 10 replications. Treatment means followed by a common letter are not significantly different according to Tukey's HSD test (P ≤ 0.05). Burkholderia refers to Burkholderia rinojensis strain A396 secondary metabolites. Discussion Meloidogyne enterolobii has emerged as a significant constraint of modern agriculture. The current study evaluated the effects of non-fumigant chemical and biological nematicides on the reproduction and pathogenicity of M. enterolobii in tobacco. Although not always significant, a single application of the maximum allowed rate of the non-fumigant chemical nematicides was able to suppress nematode reproduction and pathogenicity. The Burkholderia-derived biological nematicide, however, was not able to suppress the nematode reproduction or pathogenicity. Nevertheless, results from the current study suggest that the non-fumigant chemical nematicides greatly suppress the reproduction and pathogenicity of M. enterolobii implying a possible alternative to fumigant nematicides in tobacco. Research on the efficacy of non-fumigant nematicides against M. enterolobii in tobacco is limited. Our previous study was the first published report assessing the efficacy of non-fumigant nematicides against M. enterolobii in tobacco (Alam et al., 2022). The previous study employed a single soil application of the lower level of recommended rate of nematicides and found that oxamyl, fluensulfone, and fluopyram suppressed, respectively, 99.9%, 93%, and 99.9% of nematode eggs. Similarly, the suppression of J2 was 99% for oxamyl, 98% for fluensulfone, and 94% for fluopyram. The current study involving single soil application of the maximum allowed rate of nematicides was able to suppress up to 71% of eggs and 88% of J2 suggesting a greater efficacy was not achieved by increasing the amount of nematicide. Nevertheless, oxamyl and fluensulfone in the current study provided a significant suppression of nematode reproduction similar to those reported in our previous study. Fluopyram in the current study suppressed nematode reproduction; however, the suppression was not significantly different from the control. In contrast, our previous study found a significant suppression of nematode reproduction by the application of fluopyram. While it is unclear why the results were different, one possible reason could be the inoculum. The inoculum level in the current study was 10,000 eggs/pot while those in the previous study was 1,000 J2/pot. It is likely that J2 are more sensitive to nematicides than the eggs, which may have greatly reduced the initial population of J2. Another reason could be the timing of the nematicide application. Nematicides were applied two days after inoculation in the previous study while it was four days in the current study. Inoculating with eggs in the current study may have resulted in less exposure of J2 to the nematicides, due to variability in the timing of egg hatching and potentially differential susceptibility of unhatched nematodes. Further studies are needed to understand the interactions between the timing of J2 hatch and nematicide application timing in both greenhouse and field conditions. Application of the non-fumigant chemical nematicides resulted in greater shoot biomass relative to the control, although not necessarily at a significant level, suggesting the pathogenicity of nematodes was reduced upon nematicide application. This finding is very useful for growers that are interested in moving away from the use of fumigants. The results from the current study align with the results from our previous study where the nematicides did not adversely affect the plant shoot weight (Alam et al., 2022). Interestingly, the plants treated with fluopyram in the current study had a significantly greater (27%) shoot biomass relative to the controls suggesting a higher dose of fluopyram possibly promotes shoot growth despite not being able to suppress the nematode reproduction significantly. Although the current study did not find adverse effects of the non-fumigant nematicides on shoot biomass quantity, further studies are needed to determine their effects on the quality of tobacco shoot biomass. The root biomass of tobacco was significantly reduced by the application of fluensulfone while it was statistically unaffected by oxamyl and fluopyram implying the phytotoxicity of fluensulfone as indicated in the label. Our previous study also reported a significant reduction of root biomass in tobacco plants treated with fluensulfone (Alam et al., 2022). Oxamyl in the current study also reduced the root biomass, but the reduction was not significant. This result is in accordance with the previous study, which reported a non-significant suppression of root biomass by oxamyl. Although the root biomass was reduced upon application of oxamyl and fluensulfone, both nematicides significantly suppressed the nematode reproduction suggesting a trade-off between nematode suppression and host root biomass. Similarly, despite the inability to significantly suppress nematode reproduction by fluopyram, plants treated with this nematicide had a greater shoot and root biomass. Results from the current study suggest that the non-fumigant chemical nematicides are a viable option for managing M. enterolobii in tobacco. Previous studies involving non-fumigant nematicides in other crops have reported inconsistencies in efficacies due to the nematode species, host crop, and environment. Oka and Saroya (2019) evaluated the in-vitro sensitivity of fluopyram and fluensulfone to M. incognita and M. javanica and found that the former was more sensitive to fluensulfone while the latter was more sensitive to fluensulfone. Fluopyram was effective in suppressing ring nematode in an in-vitro study (Kabir et al., 2021) while a field study did not find a suppressive effect of this nematicide on plant-parasitic nematodes (Schumacher et al., 2020). Ali and El-Ashry (2021) reported oxamyl to be effective against plant-parasitic nematodes while other studies reported it to be ineffective (Desaeger et al., 2004; Morris et al., 2015). A study by Khanal and Desaeger (2020) reported that the efficacy of non-fumigant nematicides depended on soil temperature, nematode species, the population density of nematodes, and crop type. An on-farm study found that fluopyram and oxamyl were able to suppress ring nematode populations in a peach field only for a few months citing their inability to provide season-long protection with a single application at currently recommended rates (Khanal et al., 2022). Looking into the results from previous studies, it has become apparent that the efficacy studies of non-fumigant nematicides conducted under controlled environments provide more consistent results than those conducted in the fields. Therefore, it is critical to conduct multiple on-farm efficacy studies to derive a more practical conclusion on the efficacy of nematicides. In most US states, however, on-farm studies on M. enterolobii are challenging because of the quarantine nature of the pathogen. The Burholderia metabolite derived nematicide in the current study was not effective in suppressing the reproduction of M. enterolobii. Moreover, the reproduction of nematode on plants treated with Burkholderia-derived nematicide was higher than those in the controls, although not at a significant level. Although this nematicide did not have adverse effects on nematode reproduction, root and shoot biomass were not significantly impacted suggesting the other ingredients in the formulation may promote plant growth to compensate for the losses from the nematode attack. Several studies conducted in the past have reported the Burkholderia-derived nematicide to be ineffective in suppressing nematode reproduction (Watson and Desaeger, 2019; Alam et al., 2022; Khanal and Desaeger, 2020) suggesting a need for the development of biological nematicides with better efficacy. The high levels of galling and reproduction of up to 320,400 eggs/root system supported by the tobacco cultivar K346 employed in the current study suggest the cultivar is susceptible to M. enterolobii. This reproduction is similar to the one reported by Khanal and Harshman (2020) involving 45-day experiments with 10,000 M. enterolobii eggs as initial inoculum on susceptible tomato cultivar Rutgers. In contrast, our previous study found a lower population of nematodes and indicated the possibility of the tobacco cultivar K346 carrying resistance against M. enterolobii (Alam et al., 2022). However, as explained earlier, the lower final population of nematodes may have resulted from the use of J2 as inoculum. Non-fumigant nematicides employed in this study seem to be a promising tool to manage M. enterolobii in tobacco. Results suggest that further research is needed to improve the efficacy of non-fumigant nematicides. The means to increase the efficacy could be employment of better application methods such as multiple applications or finding better chemistries. Until host plant resistance becomes available, better management of M. enterolobii can be achieved by combining non-fumigant nematicides with other compatible management options such as crop rotation and cover cropping (Khanal and Harshman, 2020). Acknowledgments The authors would like to thank Jeanice Troutman, David Harshman, and Ivan Alarcon Mendoza for technical assistance and the chemical companies that supplied us with the nematicides. Mention of a trade name, warranty, proprietary product, or vendor does not constitute an endorsement of a product and does not imply its approval to the exclusion of other products or vendors that also may be suitable. Funding for this study was provided by the South Carolina Tobacco Board Grant no. 2015272. ==== Refs Literature Cited Alam M. S. Khanal C. Rutter W. Roberts J. 2022 Non-fumigant Nematicides are Promising Alternatives to Fumigants for the Management of Meloidogyne enterolobii in Tobacco Journal of Nematology 54 Ali A. A. I. El-Ashry R. M. 2021 Potential effect of the nematicide oxamyl and surfactant combinations on root knot nematode Meloidogyne incognita infecting tomato plants Egyptian Academic Journal of Biological Science 13 159 176 Bridge J. Page S. L. J. 1980 Estimation of root-knot nematode infestation levels on roots using a rating chart International Journal of Pest Management 26 296 298 Desaeger J. Csinos A. Timper P. Hammes G. Seebold K. 2004 Soil fumigation and oxamyl drip applications for nematode and insect control in vegetable plasticulture Annals of Applied Biology 145 59 70 Desaeger J. A. Watson T. T. 2019 Evaluation of new chemical and biological nematicides for managing Meloidogyne javanica in tomato production and associated double-crops in Florida Pest Management Science 75 3363 3370 31074102 Gomes C. B. Couto M. E. O. Carneiro R. M. D. G. 2008 Occurrence of Meloidogyne mayaguensis on guava and tabacco in South of Brazil Nematologia Brasileira 32 244 247 Hussey R. S. Barker K. R. 1973 A comparison of methods for collecting inocula for Meloidogyne spp., including a new technique Plant Disease Reporter 57 1025 1028 Jenkins W. R. 1964 A rapid centrifugal-flotation technique for separating nematodes Plant Disease Reporter 48 692 Kabir M. F. Na H. Choi I. H. Cha Y. S. Mwamula A. O. Kim Y. G. Lee G. W. Lee G. Kim K. A. Lee D. 2021 Efficacy of nematicides against two destructive nematodes; Helicotylenchus microlobus (nematoda: tylenchida) and Mesocriconema nebraskense (nematoda: criconematina) in turfgrass The Korean Journal of Pesticide Science 25 212 220 Khanal C. Harshman D. Giles C. 2022 On-farm evaluations of non-fumigant nematicides on nematode communities of peach Phytopathology 112 2218 2223 35585720 Khanal C. Harshman D. 2022 Evaluation of summer cover crops for host suitability of Meloidogyne enterolobii Crop Protection 151 105821 Khanal C. Desaeger J. A. 2020 On-farm evaluations of non-fumigant nematicides on cucurbits Crop Protection 133 105152 Khanal C. Galbieri R. Timper P. 2021 Rotations with Crotalaria spp. do not suppress populations of Meloidogyne incognita in cotton Nematology 26 929 937 Morris K. A. Langston D. B. Dickson D. W. Davis R. F. Timper P. Noe J. P. 2015 Efficacy of fluensulfone in a tomato-cucumber double cropping system Journal of Nematology 47 310 315 26941459 Oka Y. Saroya Y. 2019 Effect of fluensulfone and fluopyram on the mobility and infection of second-stage juveniles of Meloidogyne incognita and M. javanica Pest Management Science 75 2095 2106 30843368 Schumacher L. Grabau Z. Wright D. Small I. Liao H. 2020 Nematicide influence on cotton yield and plant-parasitic nematodes in conventional and sod-based crop rotation Journal of Nematology 52 1 14 Tanner S. Li C. Ye W. Davis E. 2020 Distribution of Meloidogyne enterolobii in Eastern North Carolina and Comparison of Four Isolates Plant Health Progress 21 91 96 Watson T. T. Desaeger J. A. 2019 Evaluation of non-fumigant chemical and biological nematicides for strawberry production in Florida Crop Protection 117 100 107 Yang B. Eisenback J. D. 1983 Meloidogyne enterolobii n. sp. (Meloidogynidae), a root-knot nematode parasitizing pacara earpod tree in China Journal of Nematology 15 381 391 19295822 Ye W. Koenning S. R. Zeng Y. Zhuo K. Liao J. 2021 Molecular characterization of an emerging root-knot nematode Meloidogyne enterolobii in North Carolina, USA Plant Disease 105 819 831 32910724

PMC010xxxxxx/PMC10241306.txt

==== Front J Nematol J Nematol jofnem jofnem Journal of Nematology 0022-300X 2640-396X Sciendo 37283999 jofnem-2023-0022 10.2478/jofnem-2023-0022 Research Paper Morphological and Molecular Characterization of Talanema eshtiaghii sp. n. (Dorylaimida, Qudsianematidae) from Iran Vazifeh Nasir Niknam Gholamreza g_niknam@tabrizu.ac.ir Jabbari Habibeh Peña-Santiago Reyes Department of Plant Protection, Faculty of Agriculture, University of Tabriz, Tabriz, Iran Department of Plant Protection, Faculty of Agriculture, University of Maragheh, Maragheh, Iran Departamento de Biología Animal, Biología Vegetal y Ecología, Universidad de Jaén, Campus ‘Las Lagunillas’ s/n, Edificio B3, 23071-Jaén, Spain LSID code for this publication: urn:lsid:zoobank.org:act: 250EA066-7CC1-48DD-94C5-4ADD6490E015. This paper was edited by Majid Pedram. 2 2023 5 6 2023 55 1 2023002218 12 2022 © 2023 Nasir Vazifeh et al., published by Sciendo 2023 Nasir Vazifeh et al., published by Sciendo https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. Abstract A new species of the genus Talanema, recovered from the northwest of Iran, was described based on morphological, morphometric, and molecular data. Talanema eshtiaghii sp. n. was characterized by its 1.45–1.68 mm long body, lip region offset by constriction and 13–15 μm wide, odontostyle 15–18 μm long, double guiding ring, neck 312–362 μm long, pharyngeal expansion occupying 41–43% of the total neck length, uterus tripartite, and 111–189 μm long or 2.1–3.2 body diameters, vulva transverse (V = 55–58), tail similar in both sexes, conical with a dorsal concavity (30–44 μm, c = 33–56, c’ = 1.0–1.6), spicules 49–56 μm long, and 14–18 shortly spaced ventromedian supplements in front of the level of the anterior end of spicules, with distinct hiatus. It was compared to four closely similar species, with emphasis on the most relevant traits to distinguish them. Molecular phylogenetic studies using partial sequence of the 28S rDNA (D2–D3 segment) revealed that the new species forms a clade with other currently sequenced representatives of Talanema, tentatively supporting the monophyly of this genus. Keywords Bayesian inference D2–D3 rDNA phylogeny taxonomy ==== Body pmcAndrássy (1991) proposed the genus Talanema to accommodate four species transferred from Labronema Thorne, 1939. Other species were later on added by Vinciguerra and Clausi (1994), Shaheen and Ahmad (2005), and Andrássy (2011). More recently, the genus was matter of two monographic contributions (Imran et al., 2021; Jabbari et al., 2021), including the first molecular study of one of its representatives using 18S, D2–D3 28S rDNA data. Besides, Peña-Santiago et al. (2022) have described its thirteenth species from the Iberian Peninsula. Jabbari et al. (2021) studied one previously undescribed and three known species from Iran, where the genus apparently is well represented and diverse. Recently, another Iranian population was collected during a nematological survey conducted in the country. Its study confirmed that it belongs to an unknown species, which is described in this contribution. Material and Methods Sampling, morphological and morphometric study Several soil samples were collected from the Sufiyan district, East Azarbaijan province, northwestern Iran. The modified method of Brown and Boag (1988) was used to extract nematodes from soil samples. Nematodes were transferred to anhydrous glycerine according to De Grisse's (1969) method and mounted on glass slides. Morphological observations were made and morphometrics were taken using an Olympus BX41 microscope with a drawing tube device. Micrographs were taken using a DP50 digital camera attached to the same microscope, powered with differential interference contrast (DIC). Drawings were made using the CorelDRAW® software version 12. DNA extraction, PCR and sequencing A single nematode specimen of the new species was picked out and transferred to a small drop of distilled water or worm lysis buffer (WLB) and crushed by a sterilized scalpel. The suspension was transferred to a microtube containing 25.65 μl ddH2O, 2.85 μl 10X PCR buffer, and 1.5 μl proteinase K (600 μg/ml) (Promega, Benelux, the Netherlands). The microtube was incubated at 65°C (1 h), then at 95°C (15 min). The resulting DNA extract was stored at −20°C until used as a template for polymerase chain reaction (PCR). For DNA amplification, 1 μl of the extracted DNA was transferred to a microtube containing: 0.75 μl of each primer, 12.5 μl Taq DNA Polymerase 2× Master Mix RED, 2Mm MgCl2 (Amplicon-Denmark) and ddH2O to a final volume of 25 μl. The thermal cycler was programmed as follows: denaturation at 94°C for 2 min, followed by 35 cycles of denaturation at 94°C for 30 sec, annealing at 55°C for 45 sec, and extension at 72°C for 3 min. A final extension was performed at 72°C for 10 min (Archidona-Yuste et al., 2016). Primers for 28S rDNA D2–D3 amplification were D2A (5′-ACAAGTACCGTGAGGGAAAGT-3′) as forward primer and D3B (5′ TGCGAAGGAACCAGCTACTA-3′) (Nunn, 1992) as reverse primer. The PCR product was sequenced in both directions using the same primers used in PCR with an Applied Biosystems® 3730/3730xl DNA Analyzer in South Korea. The newly generated sequence of the studied species was deposited in GenBank database (accession number OP870361). Phylogenetic analyses The recently obtained LSU rDNA sequence was edited/trimmed and compared with other dorylaimid sequences available in the GenBank database using the BLAST homology search program of the National Centre for Biotechnology Information (NCBI). The selected DNA sequences were aligned using the Muscle software implemented in MEGA6 (Tamura et al., 2013). MrModeltest 2.3 (Nylander, 2004) was used to select the base substitution model supported by the Akaike criterion in conjunction with PAUP* v4.0b10 (Swofford, 2003). Bayesian analysis (BI) was performed using MrBayes 3.1.2 (Ronquist and Huelsenbeck, 2003) running the chains for 10 million generations. After discarding burn-in samples, the remaining samples were retained for further analyses. Posterior probabilities (PP) are given on appropriate clades. The tree was visualized using FigTree v1.4.3 and was digitally drawn in CorelDRAW software version 12. Results Systematics Talanema eshtiaghii sp. n. (Figs. 1, 2) Figure 1: Talanema eshtiaghii sp. n. (A) neck region; (B) anterior region in lateral median view; (C) anterior region in lateral surface view; (D) female, posterior genital branch; (E): spicules; (F) female, caudal region; (G) female, entire; (H) male, entire; (I) male, caudal region; (J) lateral guiding piece. Figure 2: Talanema eshtiaghii sp. n. (A) anterior region in lateral median view; (B) anterior region in lateral surface view; (C) female, entire; (D) female, caudal region; (E) vagina region: (F) female, posterior genital branch; (G) sperm cells; (H) lateral guiding piece; (I) male, caudal region; (J) male, entire; (K) pharyngo-intestinal junction; (L) spicules. (Scale bars: A, B, D–I, K and L = 10 μm, C and J = 100 μm). Morphometrics See Table 1. Table 1. Morphometrics of Talanema eshtiaghii sp. n. from Iran. Measurements in μm except L in mm, and in the form: average ± sd (range). Locality Province City of Sufiyan, Roodghat area, Zeinabad village East Azarbaijan Holotype Paratypes n ♀ 5♀♀ 3♂♂ Character L 1.50 1.56 ± 0.08 (1.45–1.68) 1.49 ± 0.04 (1.45–1.53) a 32 29.2 ± 2.1 (27–32) 27.6 ± 2.1 (27–30) b 4.3 4.5 ± 0.2 (4.3–4.9) 4.5 ± 0.1 (4.4–4.6) c 44 42.1 ± 7.7 (33–56) 39.7 ± 3.6 (37–44) c’ 1.2 1.3 ± 0.1 (1.0–1.6) 1.1 ± 0.1 (1.0–1.3) V 55 56 ± 1.2 (55–58) - Lip region diameter 13 13.7 ± 0.1 (13–15) 13.4 ± 0.1 (13–14) Odontostyle length-dorsal side 18 16.4 ± 1.3 (15–18) 17.3 ± 0.5 (17–18) Odontophore length 24 25.1 ± 1.7 (24–27) 26.1 ± 0.7 (25–27) Neck length 344 349 ± 10 (341–362) 322 ± 11 (312–337) Pharyngeal expansion length 144 141.0 ± 7.6 (131–152) 133.0 ± 2.1 (124–138) Body diameter at neck base 45 51.1 ± 4.3 (45–56) 52.0 ± 3.3 (49–56) mid-body 47 52.3 ± 2.8 (47–59) 53.4 ± 3.5 (50–58) anus/cloaca 27 28.3 ± 1.2 (27–30) 32.3 ± 1.5 (31–34) Prerectum length 61 67 ± 10 (56–81) 73.4 ± 8.3 (59–97) Rectum/cloaca length 34 34.2 ± 2.8 (30–38) 59.0 ± 4.1 (58–66) Tail length 34 38.7 ± 3.6 (30–44) 36.1 ± 3.8 (33–40) Spicules length - - 52.2 ± 3.5 (49–56) Ventromedian supplements - - (14–18) Description Adult Moderately slender to slender (a = 27–32) nematodes of medium size, 1.45–1.68 mm long. Body cylindrical, tapering towards both ends, but more so posteriorly, as the tail is conical in both sexes. Upon fixation, habitus slightly curved ventrad, to an open C shape. Cuticle almost smooth under light microscopy, two-layered, its total thickness 1.5–2.0 μm at anterior region, 2.5–4.0 μm in mid-body, and 4.0–4.5 μm on dorsal side of tail. Lateral chord 10.5–14.5 μm or 21–24% of maximum body diameter. Lip region offset by a weak but distinct constriction, 2.5–3.2 times as wide as high and less than one-third (24–32%) of body diameter at neck base; lips moderately separate and slightly angular, and perioral liplets might be present, labial and cephalic papillae visibly protruding. Amphidial fovea cup-shaped, its aperture 7.5–9.0 μm long or up to two-thirds (57–67%) of lip region diameter. Cheilostom a truncate, thick-walled cone. Odontostyle robust, hardly but appreciably shorter at its ventral side, 6.4–7.4 times as long as wide, longer (1.1–1.2 times) than lip region diameter, its aperture 5.5–7.0 μm or one-third to two-fifths (32–41%) of the total length. Guiding ring double, fixed ring situated at 9–11 μm or 0.7 times the lip region diameter from the anterior end. Odontophore rod-like, 1.3–1.6 times longer than odontostyle. Pharynx entirely muscular, gradually enlarging into the basal expansion that is 5.9–6.9 times longer than wide, 2.6–3.2 times longer than body diameter at neck base, and occupies less than one-half (39–43%) of the total neck length; gland nuclei located as follows: DO=59–60, DN=62–68, S1N1=70–75, S1N2=77–79, S2N=85–87. Nerve ring located at 126–146 μm distance from anterior end or 34–43% of the total neck length. Pharyngo-intestinal junction consisting of a short and rounded conoid cardia enveloped by intestinal tissue, all together forming a longer conoid, 14–19 × 12–15 μm structure, bulging into the intestinal lumen. Female Genital system diovarian, with equally developed branches, the anterior branch 260–322 μm long or 16–20% of body length, the posterior one 233–322 μm or 16–20% of body length. Ovaries moderately developed, often reaching and surpassing the sphincter level, 60–74 μm the anterior and 60–79 μm the posterior, with oocytes first arranged in two or more rows, then in only one row. Oviduct joining subterminally the ovary, 62–89 μm or 1.1–1.5 times the body diameter long, consisting of a slender distal region made of prismatic cells and an often well-developed proximal pars dilatata with lumen inside. Sphincter present between oviduct and uterus. Uterus tripartite, usually convoluted, 111–189 μm long or 2.1–3.2 body diameters, consisting of a distal, almost spherical, pars dilatata with visible lumen, a long and slender intermediate section without visible lumen and somewhat refractive lining, and a proximal dilated region with wide lumen. Uterine egg not observed. Vagina extending inwards 23–25 μm to 38–44% of body diameter: pars proximalis 15.6–17 × 10–11.2 μm, with almost straight walls surrounded by weakly developed musculature, pars refringens consisting of (lateral view) two close together, often mostly trapezoidal, sclerotized pieces measuring 4.5–5.5 × 3–3.5 μm and with a combined width of 8–9 μm, and pars distalis 3–4 μm long. Vulva a slightly post-equatorial, transverse slit. Prerectum 2.0–2.7, rectum 1.1–1.3 anal body diameters long. Caudal region conical with finely rounded terminus, ventrally almost straight, dorsally first convex, then with a more or less (usually well) distinct concavity, giving a digitate aspect to the tail; inner core reaching 52–59% of tail length, leaving an appreciable hyaline portion; caudal pores one pair, at the middle of tail, one subdorsal, another lateral. Male Genital system diorchic, with opposite testes. In addition to the ad-cloacal pair, situated at 7.5 μm from the cloacal aperture, there is a series of 14–18, almost contiguous or shortly spaced, 6.0–7.5 μm apart, ventromedian supplements, the most posterior of them situated at 49–51 μm from the ad-cloacal pair, appreciably in front of the anterior end of spicules, then with a distinct hiatus. Spicules dorylaimid, 4.9–5.5 times as long as wide, 1.5–1.8 times the body diameter at level of the cloacal aperture: head 10.0–16.5 μm long or 18–32% of spicule length, 2.2–2.4 times longer than wide, and its dorsal side much longer and more curved than the ventral one; median piece comparatively slender, occupying 25–28% of spicule maximum width; posterior tip 3.5 μm wide; ventral hump located at 18–22 μm or 36–42% of spicule length from its anterior end; curvature 134º. Lateral guiding piece 13–14 μm long, coarse, 3.4–4.4 times as long as wide, conspicuously tapering at its posterior third. Prerectum 1.8–3.7, cloaca 1.8–1.9 body diameters long. Caudal region basically similar to that of female, but more straight or even slightly curved ventrad at the end. Molecular characterization Sequencing the D2–D3 region of the 28S rDNA resulted in one sequence 751 bp (OP870361) long. The BLAST homology search showed it has an 89.09% identity with Talanema baqrii (Khan et al., 1989; Imran et al., 2021) (MT645228), 91% identity with T. ibericum (Peña-Santiago et al., 2022) (OP793646) and 94% identity with Talanema sp. (OP870362). Diagnosis The new species is characterized by its 1.45–1.68 mm long body, lip region offset by constriction and 13–15 μm wide, odontostyle 15–18 μm long, guiding ring double, neck 312–362 μm long, pharyngeal expansion occupying 41–43% of the total neck length, uterus tripartite and 111–189 μm long or 2.1–3.2 body diameters, vulva transverse (V = 55–58), tail similar in both sexes, conical with a dorsal concavity (30–44 μm, c = 33–56, c’ = 1.0–1.6), spicules 49–56 μm long, and 14–18 shortly spaced ventromedian supplements in front of the level of the anterior end of spicules, with distinct hiatus. Relationships Morphologically and morphometrically, the new species resembles three Talanema species showing no sexual dimorphism of tail shape, namely, T. ibarakiense (Khan and Araki, 2002; Andrássy, 2011), T. pararapax (Ahmad and Jairajpuri, 1982; Andrássy, 1991) and T. sphinctum (Mohilal and Dhanachand, 2001; Imran et al., 2021), but it can be distinguished from these in its shorter odontostyle (15–18 versus 20 μm long or more). Besides, it differs from T. ibarakiense in its larger general size (body length 1.45–1.68 versus 1.0–1.2 mm), more posterior vulva (V = 55–58 versus 50–52), longer tail (30–44 versus 21–24 μm, c’ = 1.0–1.6 versus 0.8–0.9), and less (14–18 versus 21–22) ventromedian supplements with distinct (versus without) hiatus. From T. pararapax in its shorter spicules (49–56 versus 57–63 μm) and different arrangement of ventromedian supplements (ending in front of versus at level of the anterior end of spicules). It can be separated from the type population of T. sphinctum in its more posterior vulva (V = 55–58 versus V = 52–55), higher number of ventromedian supplements (14–18 versus 8) with very different arrangement (shortly versus widely spaced, with distinct versus no hiatus). It is also similar to T. saccatum (Jabbari et al., 2021), an Iranian monosexual species, from which the new species can be separated in its more slender body (a = 27–32 versus a = 22–26), somewhat shorter odontostyle (15–18 versus 18–20 μm), and slightly longer female tail (30–44 versus 24–31 μm, c’ = 1.0–1.6 versus c’ = 0.8–1.0) with no (versus abundant) saccate bodies. Finally, the new species also resembles Labronema digiturum (Vinciguerra, 1984), a taxon with a questionable belonging to Labronema, but it can be distinguished from this in its much more slender body (a = 27–32 versus a = 16–20), shorter odontostyle (15–18 versus 19–20 μm), tripartite (versus apparently simple) uterus, absence (versus presence) of cuticular irregularities (wrinkles) at both sides of vulva, and shorter spicules (49–56 versus 67 μm). The evolutionary relationships of the new species, as derived from the molecular analyses, are presented in the tree of Figure 3. The 28S rDNA sequence of T. eshtiaghii sp. n. formed a maximally supported clade (1.00 Bayesian posterior probability) with other Talanema representatives, namely T. baqrii, T. ibericum and Talanema sp. This means that, at present, Talanema is a monophyletic taxon. Nevertheless, its external relationships remain unsatisfactorily resolved. Thus, on the one hand, it forms part of a larger clade also including two sequences of Labronema montanum (Peña-Santiago and Abolafia, 2019) as a sister group, but with comparatively low support (0.84 BPP). On the other hand, this (Talanema + Labronema) clade is only one out of six Dorylaimina subclades, whose evolutionary relationships are neither well resolved. Figure 3: Bayesian 50% majority rule consensus tree of Talanema eshtiaghii sp. n. as inferred from D2–D3 expansion segments of 28S rRNA gene sequence under the GTR+I+G model. Bayesian posterior probabilities are given for each clade. Newly obtained sequence is indicated by bold letters. Type locality and habitat Northwestern Iran, East Azarbaijan province, Sufiyan district, Roodghat area, Zeinabad village (38°17′46″N, 46°07′37″E, elevation 1528 m a.s.l.), where the new species was collected in soil from the rhizosphere of common wheat (Triticum aestivum L.). Type material Female holotype, three female paratypes and two male paratypes deposited with Collection of Nematology Laboratory, University of Tabriz, Iran. One female paratype and one male paratype deposited in the Nematode Collection of the University of Jaén, Spain. The LSID for this publication is urn:lsid:zoobank.org:act:250EA066-7CC1-48DD-94C5-4ADD6490E015 Etymology The new species is named in honor of Dr. Hassan Eshtiaghi, the late Nematologist in the Department of Plant Protection, University of Tehran, Tehran, Iran. Acknowledgements The Spanish author (RPS) thanks the University of Jaén, Spain, for the financial support received through the research program ‘POAIUJA 2021/2022: EI_RNM02_2021’. ==== Refs Literature Cited Ahmad W. Jairajpuri M. S. 1982 Some new and known species of Dorylaimoidea Nematologica 28 39 61 10.1163/187529282X00501 Andrássy I. 1991 The superfamily Dorylaimoidea (Nematoda): a review. Family Qudsianematidae, II Opuscula Zoologica Budapestensis 24 3 51 Andrássy I. 2011 Three new bisexual species of Labronema Thorne, 1939 (Nematoda: Qudsianematidae) Opuscula Zoologica Budapestensis 42 107 120 Archidona-Yuste A. Navas-Cortés J. A. Cantalapiedra-Navarrete C. Palomares Rius J. E. Castillo P. 2016 Cryptic diversity and species delimitation in the Xiphinema americanum-group complex (Nematoda: Longidoridae) as inferred from morphometrics and molecular markers Zoological Journal of the Linnean Society 176 2 231 265 10.1111/zoj.12316 Brown D. J. F. Boag B. 1988 An examination of methods used to extract virus-vector nematodes (Nematoda: Longidoridae and Trichodoridae) from soil samples Nematologia Mediterranea 16 93 99 De Grisse A. T. 1969 Redescriptionou modification de quelques techniques utilisées danslétude des nematodes phytoparasitaires Mededelingen Rijksfaculteit Landbouwwetenschappen, Gent 34 351 369 Imran Z. Abolafia J. Ahmad W. 2021 On the identity of Labronema baqrii Khan, Jairajpuri and Ahmad, 1989 (Nematoda, Dorylaimida) and analysis of the relationships between the genera Labronema Thorne, 1939 and Talanema Andrássy, 1991 Zoologischer Anzeiger 291 103 112 10.1016/j.jcz.2021.02.003 Jabbari J. Niknam G. Vazifeh N. Fallahi A. 2021 Description of a new species of the genus Talanema Andrássy, 1991 (Nematoda: Qudsianematidae) with additional data on three known species of the genus from Iran Biologia 76 3323 3334 10.1007/s11756-021-00821-x Khan T. H. Jairajpuri M. S. Ahmad W. 1989 Two new species of Labronema (Nematoda: Dorylaimida) from India Indian Journal of Nematology 19 194 189 10.1163/002825989X00179 Khan Z. Araki M. 2002 Study of dorylaims (Nematoda) from Japan with descriptions of five new species International Journal of Nematology 12 1 12 Mohilal N. Dhanachand C. 2001 Investigations of soil nematodes from Manipur, species of the family Qudsianematidae Uttar Pradesh Journal of Zoology 21 41 45 Nunn G. B. 1992 Nematode molecular evolution: an investigation of evolutionary patterns among nematodes based upon DNA sequences Ph.D. dissertation, University of Nottingham Nottingham, UK Nylander J. A. A. 2004 MrModeltest v2. Program distributed by the author Uppsala University, Sweden Evolutionary Biology Centre Peña-Santiago R. Abolafia J. 2019 Morphological and molecular characterization of Labronema montanum sp. n. (Dorylaimida, Dorylaimidae) from Spain Journal of Nematology 51 e2019 29 10.21307/jofnem-2019-029 Peña-Santiago R. Cortés N. García-Ruiz M. Abolafia J. 2022 Morphological and molecular characterization of Talanema ibericum sp. n. (Dorylaimida, Qudsianematidae) from southern Iberian Peninsula Nematology [in press]. 10.1163/15685411-bja10214 Ronquist F. Huelsenbeck J. P. 2003 MRBAYES: Bayesian inference of phylogenetic trees Bioinformatics 19 1572 1574 10.1093/bioinformatics/17.8.754 12912839 Shaheen A. Ahmad W. 2005 Descriptions of four new species of Dorylaimida (Nematoda) Journal of Nematode Morphology and Systematics 8 1 9 29 Swofford D. L. 2003 PAUP*: phylogenetic analysis using parsimony (*and other methods). Version 4.0b10 Sunderland, MA Sinauer Associates Tamura K. Stecher G. Peterson D. Filipski A. Kumar S. 2013 MEGA6: molecular evolutionary genetics analysis. Version 6.0 Molecular Biology and Evolution 30 2725 2729 10.1093/molbev/mst197 24132122 Thorne G. 1939 A monograph of the superfamily Dorylaimoidea Capita Zoologica 8 1 261 Vinciguerra M. T. 1984 Description of two new species and remarks on some known species of nematodes from Sardinia Animalia 11 127 134 Vinciguerra M. T. Clausi M. 1994 Nematodes of Salina. Three new and one rare species of Qudsianematidae (Dorylaimida) Animalia 21 97 112

PMC010xxxxxx/PMC10241307.txt

==== Front J Nematol J Nematol jofnem jofnem Journal of Nematology 0022-300X 2640-396X Sciendo 37284000 jofnem-2023-0020 10.2478/jofnem-2023-0020 Research Paper Reaction of Commercial Cultivars of Kiwifruit to Infection by Root-knot Nematode and Its Biocontrol Using Endophytic Bacteria Banihashemian Seyedeh Najmeh samaneh.bn@yahoo.com Jamali Salar Golmohammadi Morteza Noorizadeh Sina Atighi Mohammad Reza Plant Protection Department, Faculty of Agricultural Sciences, University of Guilan, Rasht, Iran Horticultural Science Research Institute, Citrus and Subtropical Fruits Research Center, Agricultural Research Education and Extension Organization (AREEO), Ramsar, Iran Plant Protection Department, Agriculture Faculty, Tabriz University, Tabriz, Iran Department of Plant Pathology, Faculty of Agriculture, Tarbiat Modares University, Tehran, Iran This paper was edited by Sapinder Bali. 2 2023 5 6 2023 55 1 2023002013 2 2023 © 2023 Seyedeh Najmeh Banihashemian et al., published by Sciendo 2023 Seyedeh Najmeh Banihashemian et al., published by Sciendo https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. Abstract Root-knot nematodes (RKN) cause considerable economic losses to kiwifruit production annually. Screening of resistant cultivars has been one of the long-standing methods to manage root-knot nematodes. Here, the reaction of the four most common commercial cultivars of kiwifruit, namely, Actinidia chinensis var. deliciosa cv. Hayward, A. chinensis var. deliciosa cv. Abbott, A. chinensis var. deliciosa cv. Bruno, and A. chinensis var. chinensis cv. Haegeum (commonly known as ‘Golden’ kiwifruit) to infection by the RKN, Meloidogyne incognita, was evaluated. Among examined cultivars ‘Golden’ was the most susceptible, having on average 52.8 galls, 56.1 egg masses per gram of root, and 642 J2 population per 200 gram of soil. ‘Bruno’ showed the highest resistance, with 3.3 galls, 4.1 egg masses per gram of root, and 79 J2 in 200 g of soil. Then, two potential biological control agents, namely Priestia megaterium 31.en and Agrobacterium tumefaciens 19.en were used on ‘Hayward’ seedlings against M. incognita and showed a significant reduction in the number of galls and egg masses on roots, juvenile population in the soil, and increased the growth parameters of the plants compared to non-treated seedlings. We demonstrated that integrated management using resistant cultivars and biological control can provide a safe and economic method to control RKN, and these resistant cultivars can be used in breeding programs. Keywords cultivars endophytic bacteria kiwifruit root-knot nematode ==== Body pmcKiwifruit (Actinidia spp. ‘Lindl’) is one of the most important horticultural crops around the world. It is important not only for its nutritional properties, e.g., high in antioxidants and vitamin C content, but also for its application for the treatment of cancer, hepatitis, and cardiovascular disease (Li et al., 2016; Pan et al., 2020). The genus Actinidia includes more than 70 species (Garcia et al., 2011; Ma et al., 2021) and is distributed around the world, especially in China, Italy, New Zealand, Chile, Greece, and Iran (Ferguson, 2016). The Food and Agriculture Organization (FAO) reported that the production of kiwifruit in the world was about 4.4 million tons in 2021. Iran, with a production of 294,000 tons, was ranked fifth (FAOSTAT, 2021). Among the major commercial kiwifruits varieties like A. chinensis var. deliciosa cv. Hayward (green kiwifruit) and A. chinensis var. chinensis cv. Haegeum (‘Golden’ kiwifruit), Hayward is the most common commercial cultivar in Iran (Maghdouri et al., 2021). Soil-borne pathogens cause a significant yield loss on kiwifruit annually and among them, root-knot nematodes (RKN), Meloidogyne spp., are one of the most important pathogens. Four species of RKN, including M. incognita, M. javanica, M. arenaria, and M. hapla have been reported from Iran and M. incognita (Mi) is known as the major species infecting kiwifruit orchards of Mazandaran Province, the biggest kiwifruit production region in Iran (Tanha Maafi and Mahdavian, 1997; Banihashemian et al., 2022). Management of nematodes is difficult because of their wide host range and survival in diverse environmental conditions. Using chemical nematicides is the most common management method for nematodes. Due to the extensive use of pesticides and their deleterious effects on the environment and human health, their use has been banned or limited. Hence, using eco-friendly methods like resistant cultivars and biological control have attracted special attention recently (Vetrivelkalai, 2019; Eliwa and Hagag, 2021). The use of nematode-resistant cultivars along with biological control in integrated management programs provides a suitable strategy to control RKN (Mukhtar et al., 2014; Mukhtar, 2018; Eliwa and Hagag, 2021). Resistant cultivars have been successfully used against RKN in tomato (Lizardo et al., 2022), soybean (Izuogu et al., 2015), pepper (Bello et al., 2015) and even provided the complete-spectrum resistance in the case of Myrobalan plum (Claverie et al., 2011). Using biological control agents against plant pathogens, especially plant parasitic nematodes where resistant cultivars are not available or in combination with partially resistant cultivars and other control strategies provides effective and durable control against nematodes. Endophytic bacteria employ a diversity of mechanisms to promote plant growth or protect plants against pathogens. Directly, they provide more nutrients that are required for plants, especially in soils with poor content of nitrogen, iron, and phosphorus. Production of phytohormones like auxin, gibberellic acid, ethylene, cytokinin, and abscisic acid not only promote plant growth but also are involved in plant defense against pathogens (Gamalero and Glick, 2011; Rosier et al., 2018; Ali et al., 2021). They may indirectly protect and promote plant growth by nutrient competition, antibiotic production, and induction of resistance against pathogens (Ali et al., 2021). The Bacillus cereus strain D13 protects rice plants against Xanthomonas oryzae pv. oryzae by producing a range of volatile compounds like 3,5,5-trimethylhexanol and decyl alcohol (Xie et al., 2018). It has been shown that Priestia megaterium strain JR48 induces resistance against Xanthomonas campestris pv. campestris in cruciferous plants by induction of hydrogen peroxide accumulation, callose deposition, and elevated expression of defense-related genes, especially pathogenesis-related (PR) genes through the salicylic acid signaling pathway (Li et al., 2022). Application of Streptomyces sp. on banana provided a biocontrol efficiency of 70.7% against Meloidogyne javanica (Su et al., 2017). It has been shown that treatment of kiwifruit seedlings with Pantoea ananatis and Pseudomonas chlororaphis causes a significant reduction in the number of galls and egg masses. Moreover, they significantly increased growth parameters, including root-fresh and shoot-fresh weight compared to non-treated kiwifruit seedlings (Banihashemian et al., 2022). Here, we first investigated the resistance of four common commercial cultivars of kiwifruit, including A. chinensis var. deliciosa cv. Hayward, A. chinensis var. deliciosa cv. Abbott, A. chinensis var. deliciosa cv. Bruno, and A. chinensis var. chinensis cv. Haegeum (commonly known as ‘Golden’ kiwifruit) to Mi. Also, the antagonistic potential of two endophytic bacteria namely Priestia megaterium 31.en and Agrobacterium tumefaciens 19.en against Mi was evaluated. Materials and Methods Preparation of nematode inoculum The Mi population, which was previously isolated from infected roots of kiwifruit plants in the Mazandaran provinces, Iran, was used in this study (Banihashemian et al., 2022). Briefly, a single egg mass, isolated from infected kiwifruit roots, was propagated and maintained on roots of susceptible tomato plants (Early Urbana variety) under greenhouse conditions (temperature of 25°C ± 2 and RH of 70%). After three months, egg masses were hand-picked under a stereomicroscope (Nikon, SMZ800) and were left to hatch at 25°C for five to seven days using small trays according to the tray method (Whitehead and Hemming, 1965). The freshly hatched second-stage juveniles were then used for inoculation of plants. Plant material The four most common commercial cultivars of kiwifruit cultivated in the north of Iran, namely, A. chinensis var. deliciosa cv. Hayward, A. chinensis var. deliciosa cv. Abbott, A. chinensis var. deliciosa cv. Bruno, and A. chinensis var. chinensis cv. Haegeum (‘Golden’ kiwifruit) were selected to evaluate their resistance to Mi. Seeds of kiwifruit plants were grown in an equal volume of the sterilized mixture of perlite, sand, and cocopeat in pots containing 1 k of soil under controlled temperature conditions (25°C±2) and 70% relative humidity (RH) in the greenhouse at the Citrus and Subtropical Fruits Center, Agricultural Research Education and Extension Organization (AREEO), Ramsar, Iran. Infection assay and measurement of plant growth parameters The six-month-old seedlings of kiwifruit were inoculated with 2,000 second-stage juveniles of Mi through four holes in the soil around the roots of the seedlings. The control seedlings were inoculated with the same amount of water used for nematode inoculation. Plants were watered every 7 days, and 50 days after inoculation, seedlings were uprooted, and the roots were washed gently under running water to clean off soil debris. The clean roots were stained by boiling in 0.01% acid fuchsin for 15 min, then destained in acid glycerol (100 ul of HCL in 100 ml of glycerol) for three weeks with a gentle constant shaking (80 rpm) (Atighi et al., 2021). The numbers of galls and egg masses per gram of root were counted under a stereomicroscope (Nikon SMZ800). Counting was repeated three times for each root system to make an average. The nematode gall index of kiwifruit seedlings was rated using a 0–5 scale according to Taylor and Sasser (1978). Some modifications in the rating scale based on the number of galls were used to evaluate the degree of resistance/susceptibility of cultivars (Table 1). Also, to count the number of J2s in the soil, all the soil in the pot was mixed thoroughly, and then 200 g of soil was used to extract nematode using the tray method (Whitehead and Hemming, 1965). In addition, the kiwifruit growth parameters, such as fresh and dry weight of root and shoot, were measured while plants were harvested. These experiments were conducted in a randomized block design with five biological replicates per experiment and repeated three times independently. Table 1. Modified rating scale for the evaluation of the level of resistance/susceptibility of kiwifruit cultivars based on the number of galls according to (Taylor and Sasser, 1978). Scale Number of Galls per Plant Resistance Rating 0 0 Immune (I) 1 1–2 Resistant (R) 2 3–10 Moderately Resistant (MR) 3 11–30 Moderately Susceptible (MS) 4 31–100 Susceptible (S) 5 >100 Highly Susceptible (HS) Preparation and characterization of bacterial antagonists The endophytic isolates of Priestia megaterium strain 31.en (OK560186) and Agrobacterium tumefaciens strain 19.en (OK398382) had previously been isolated from seemingly healthy kiwifruit plants in orchards of Mazandaran and Guilan provinces, Iran (Banihashemian et al., 2022). Isolation of these two isolates was done according to the method of Wicaksono et al. (2018). The samples were sterilized with 96% ethanol for 10 seconds and 2% sodium hypochlorite solution for 3 min followed by washing three times (1 min each time) with sterile-distilled water in a laminar flow cabinet. The disinfected tissues were chopped into small pieces in sterilized water and kept for 30 to 40 min; then 30 μL of suspension was cultured on sucrose nutrient agar (NAS) medium after serial dilution. Also, 100 μl of the last wash was transferred to Luria-Bertani (LB; Tryptone 10 gr, NaCl 10 g, yeast extract 5 g, distilled water 950 ml) as a control (Taechowisan et al., 2003). The plates were incubated at 25°C for 7 days. The colonies were recultured on NAS plates until pure colonies were obtained. Single colonies were stored in 60% glycerol stock at −80°C for future studies. Inoculation of endophytic bacteria, P. megaterium strain 31.en and A. tumefaciens strain 19.en on plants and evaluation of plant growth parameters The cultivar Hayward is the main cultivar grown in the north of Iran. The endophytic bacteria were cultured in liquid Luria-Bertani (LB) medium at 25°C with constant shaking at 200 rpm for 48 h to obtain a bacterial suspension. The culture of bacteria was centrifuged at 6,000 rpm for 5 min and washed three times with sterile deionized water. The achieved precipitates resuspended in sterilized water to an optical density (OD) value of 0.5 at 600 nm. To evaluate the efficacy of endophytic bacteria against Mi, six-month-old seedlings were inoculated with 40 ml suspension of the endophytic bacteria P. megaterium strain 31.en and A. tumefaciens strain 19.en (107 CFU/ml) and after 2 days, 2,000 J2 of Mi were inoculated on roots of kiwifruit seedlings and kept at greenhouse conditions, as described above. The treatments were divided into four groups as follows: (i) treated with sterile water as control; (ii) inoculated only with the Mi; (iii) pretreated with P. megaterium strain 31.en and then inoculated with the Mi; and (iv) pre-treated with A. tumefaciens strain 19.en and then inoculated with the Mi. After 50 days, the number of galls and egg masses, fresh and dry weights of shoot and root, and J2 population of nematodes in the soil were measured. The experiment was carried out in a randomized block design with five replications and repeated three times independently. Statistical analysis Data were analyzed using SAS version 9.1 software with a one-way variance analysis (ANOVA) test. The mean ± standard deviation (X±SD) was calculated in all experimental data. The Bartlett's test was performed to check the equality of variances, so data were pooled to analyze them together. The significance of differences (P < 0.05) within treatments was determined using the least significant difference (LSD) test. For all treatments, there were five replicates, and the experiment was repeated three times independently. Results Response of kiwifruit cultivars to infection by Mi The response of the selected kiwifruit cultivars to infection by Mi is shown in Figures 1A–C. The greatest number of galls (52.8) and egg masses (56.1) per root system were observed in ‘Golden’ kiwifruit, while the minimum number of galls and egg masses were observed in Bruno and Hayward cultivars, respectively (Figs. 1A,B). The rating scale showed that the cultivar ‘Golden’ was susceptible to M. incognita with 31 to 100 galls per plant. Cultivar Abbot was moderately susceptible with 11 to 30 galls per plant. Bruno and Hayward were moderately resistant, with 3 to 10 galls per plant, and Bruno showed a significant resistance when compared to Hayward (Table 1 and Fig. 1A). Also, the number of J2 per 200 g of soil was significantly increased in ‘Golden’ kiwifruit inoculated with Mi, and the minimum number of J2 was seen in the Bruno cultivar (Fig. 1C). Figure 1: Reaction of commercial cultivars of kiwifruit to infection by M. incogntia. (A) Number of galls on roots; (B) Number of egg masses on roots; (C) Number of J2s in the soil. Different letters denote significant differences (p < 0.05). Error bars indicate STD. (n=15). Likewise, fresh and dry root and shoot weights in all the cultivars were different from each other. The maximum reduction in fresh root and shoot weight was observed in the ‘Golden’ cultivar, and the maximum fresh and dry root and shoot weight was seen in the Bruno and Hayward cultivars, respectively (Figs. 2A–D). Also, regression analysis displayed positive relationships between the number of galls and reductions in dry and fresh root and shoot weights (Figs. 3A,B). Figure 2: Growth parameters of plants in response to infection by M. incogntia. (A) Fresh root weight; (B) dry root weight; (C) fresh shoot weight; (D) dry shoot weight. Different letters denote significant differences (p < 0.05). Error bars indicate STD (n=15). Figure 3: Relationships between the number of galls and percentage of reductions in fresh and dry root and shoot weights. (A) (●) and (▲) represent reductions in fresh and dry root weights of kiwifruit plants, respectively. (…) and (―) represent trend lines showing reductions in fresh and dry root weights of kiwifruit plants, respectively; (B): (●) and (▲) represent reductions in fresh and dry shoot weights of kiwifruit plants, respectively. (…) and (―) represent trend lines showing reductions in fresh and dry shoot weights of kiwifruit plants, respectively. Antagonistic effects of the endophytic bacteria P. megaterium strain 31.en and A. tumefaciens strain 19.en against Mi in greenhouse studies The number of Mi galls and egg masses in the root system decreased in kiwifruit pretreated with P. megaterium strain 31.en and A. tumefaciens strain 19.en strains compared to the control plants (Fig. 4). Moreover, kiwifruit growth parameters including the fresh and dry weight of the shoot were increased (Fig. 5). Figure 4: Antagonistic effects of P. megaterium strain 31.en and A. tumefaciens strain 19.en against M. incognita in kiwifruit under greenhouse conditions. Different letters denote significant differences (p < 0.05). Data are the mean of five replications which were repeated three times independently (n=15). Figure 5: Effects of P. megaterium strain 31.en and A. tumefaciens strain 19.en on growth parameters of kiwifruit under greenhouse conditions. Different letters denote significant differences (p < 0.05). Data are the mean of five replications that were repeated three times independently (n=15). Discussion Infection of kiwifruits by Mi causes a significant reduction in its growth and production. Different measures are used to control nematodes and due to their difficult management, using integrated management methods, especially eco-friendly measures like the use of resistant cultivars and biocontrol agents, have attracted more attention recently and can provide benefits for both farmers and the environment. Using resistance cultivars against plant parasitic nematodes has been one of the long-lasting approaches to fend off their attack. To reach this end, screening resistance plants is the main step not only to be used in the infested areas and areas with high potential for infection by RKN, but also for breeding programs or as a rootstock for a highly productive scion (Saucet et al., 2016; Lesmes-Vesga et al., 2022; Mahoonaki et al., 2023). Also, the use of biological agents to control plant diseases has been the subject of much research due to the deleterious effect of chemical pesticides on human health and the environment. The integration of eco-friendly management approaches like resistant cultivars and biological agents provides an efficient management strategy to control diseases, especially for nematodes, due to difficulties in their management. Here, we provide what we believe is the first study evaluating the reaction of the four most common commercial cultivars of kiwifruit in Iran. Also, the biological control efficacy of two endophytic bacteria, namely, P. megaterium strain 31.en and A. tumefaciens strain 19.en against Mi was evaluated. Two out of four selected cultivars showed moderate resistance against infection by Mi according to scales provided by Taylor and Sasser (1978). The same result was observed previously where the resistance of three kiwifruit genotypes against several species of RKNs was evaluated, demonstrating that their genetic background influences their resistance against Mi. Notably, variation has been observed in the reproduction of Mi on kiwifruit cultivars in agreement with the previous study by Nicotra et al. (2003). This provides a new genetic resistant source against Mi that can be used in breeding programs or as rootstock for grafting highly productive scions in infested soils. It has been shown that in case of myrobalan plum, the woody plant host, the presence of the Ma gene, confers complete-spectrum resistance against Meloidogyne spp. by hypersensitive response during migration of second-stage juveniles in host tissue (Claverie et al., 2011; Saucet et al., 2016). The myrobalan plum has been used widely as both rootstock for grafting susceptible but highly productive plum trees or in breeding programs for the generation of new resistant plants (Salesses et al., 1998; Rubio-Cabetas et al., 1998; Lecouls et al., 2004). The reduction of growth parameters of the four kiwifruit cultivars compared to non-inoculated kiwifruits, especially in the susceptible cultivar, ‘Golden’, was consistent with a previous study on okra infected with Mi (Mukhtar et al., 2014). Here, the combination of both methods for cultivar Hayward resulted in a significant reduction of the severity of the disease as well as improved plant growth by increasing nutrient absorption and production of secondary metabolites (Moyes et al., 2016; Chen et al., 2017; Naylor et al., 2017; Liu et al., 2019). Many endophytic bacteria have been reported to inhibit pathogens and promote the growth and health of plants (Maheshwari and Annapurna, 2017; Su et al., 2017; Hu et al., 2018; Santos et al., 2018; Tran et al., 2019; Vetrivelkalai, 2019). It is shown that Priestia megaterium (previously known as Bacillus megaterium) is used as a biocontrol agent against RKN (Elshafie et al., 2012; Mohammadi et al., 2017). B. megaterium DS9 reduced the nematode population in the soil and increased plant growth parameters in pepper (Tran et al., 2019). Also, B. megaterium had nematicidal activity against Mi and increased the growth parameters in sugar beet (Youssef et al., 2017). Moreover, B. megaterium increased the accessibility of available phosphorus in the soil to uptake by plants, enhanced the synthesis of organic matter in soil, and increased the growth parameters. They also caused antibiosis potential against the nematode activity (Mostafa et al., 2018). In agreement with our results, the application of endophytic A. tumefaciens strains as plant-growth-promoting bacteria was observed in other studies where endophytic A. tumefaciens CCNWGS0286 promoted the growth of Robinia pseudoacacia L. significantly (Hao et al., 2012). A. tumefaciens CR22 showed an antagonistic ability against F. oxysporum (Hernández-Pacheco et al., 2021). The endophyte A. tumefaciens showed the potential to control soybean diseases (de Almeida Lopes et al., 2018). Here, we showed the response of kiwifruit cultivars to infection by Mi. The cultivars Bruno and Hayward were found to be moderately resistant, and Bruno showed a significant resistance when compared to Hayward. Moreover, we showed the antagonistic potential of two endophytic bacteria, P. megaterium strain 31.en and A. tumefaciens strain 19.en, against Mi. The integration of both resistant cultivars and biological control agents can provide a new eco-friendly strategy to manage RKN in kiwifruit orchards and reduces the negative effects of chemical pesticides on the environment and human health. Thus, this information is valuable for kiwifruit growers to use resistant plants in infested areas, especially in combination with endophytic bacteria. In our study, the number of galls, egg masses, and nematodes in Bruno and Hayward cultivars indicate the presence of moderate resistance, and its mechanisms remain to be elucidated. Also, which mechanisms by two endophytic bacteria are employed to decrease the negative effect on Mi infection needs to be explored. Acknowledgements Authors would like to thank the Agricultural Research Education and Extension Organization (AREEO), Ramsar, Iran, for providing facilities to perform experiments. ==== Refs Literature Cited Ali M. Ali Q. Sohail M. A. Ashraf M. F. Saleem M. H. Hussain S. Zhou L. 2021 Diversity and taxonomic distribution of endophytic bacterial community in the rice plant and its prospective International Journal of Molecular Sciences 22 10165 https://doi.org/10.3390/ijms221810165 34576331 Atighi M. R. Verstraeten B. De Meyer T. Kyndt T. 2021 Genome-wide shifts in histone modifications at early stage of rice infection with Meloidogyne graminicola Molecular Plant Pathology 22 440 455 https://doi.org/10.1111/mpp.13037 33580630 Banihashemian S. N. Jamali S. Golmohammadi M. Ghasemnezhad M. 2022 Isolation and identification of endophytic bacteria associated with kiwifruit and their biocontrol potential against Meloidogyne incognita Egyptian Journal of Biological Pest Control 32 1 12 https://doi.org/10.1186/s41938-022-00601-y Bello T. Fawole B. Claudius-Cole A. 2015 Susceptibility of seven varieties of pepper and tomato to root-knot nematodes (Meloidogyne spp.) in Ibadan Journal of Agriculture and Veterinary Science 8 79 82 10.9790/2380-081017982 Chen S. Zhang M. Wang J. Lv D. Ma Y. Zhou B. Wang B. 2017 Biocontrol effects of Brevibacillus laterosporus AMCC100017 on potato common scab and its impact on rhizosphere bacterial communities Biological Control 106 89 98 https://doi.org/10.1016/j.biocontrol.2017.01.005 Claverie M. Dirlewanger E. Bosselut N. Van Ghelder C. Voisin R. Kleinhentz M. Lafargue B. Abad P. Rosso M.-N. Chalhoub B. 2011 The Ma gene for complete-spectrum resistance to Meloidogyne species in Prunus is a TNL with a huge repeated C-terminal post-LRR region Plant Physiology 156 779 792 https://doi.org/10.1104/pp.111.176230 21482634 de Almeida Lopes K. B. Carpentieri-Pipolo V. Fira D. Balatti P. A. López S. M. Y. Oro T. H. Stefani Pagliosa E. Degrassi G. 2018 Screening of bacterial endophytes as potential biocontrol agents against soybean diseases Journal of Applied Microbiology 125 1466 1481 https://doi.org/10.1111/jam.14041 29978936 Eliwa G. I. Hagag E. S. 2021 Approach to new peach rootstocks resistant to root-knot nematodes (Meloidogyne species) selected from local Mit-Ghamer peach cultivar Scientia Horticulturae 284 110118 https://doi.org/10.1016/j.scienta.2021.110118 Elshafie H. S. Camele I. Racioppi R. Scrano L. Iacobellis N. S. Bufo S. A. 2012 In vitro antifungal activity of Burkholderia gladioli pv. agaricicola against some phytopathogenic fungi International Journal of Molecular Sciences 13 16291 16302 https://doi.org/10.3390/ijms131216291 23208371 FAOSTAT 2021 Citrus production in the world Ferguson A. R. 2016 World economic importance 37 42 Testolin R. Huang H. Ferguson A. R. The kiwifruit genome Cham Springer Gamalero E. Glick B. R. 2011 Mechanisms used by plant growth-promoting bacteria 17 46 Maheshwari D. Bacteria in agrobiology: plant nutrient management Berlin Springer https://doi.org/10.1007/978-3-642-21061-7_2 Garcia C. V. Quek S.-Y. Stevenson R. J. Winz R. A. 2011 Characterization of the bound volatile extract from baby kiwi (Actinidia arguta) Journal of Agricultural and Food Chemistry 59 8358 8365 https://doi.org/10.1021/jf201469c 21702480 Hao Z. Fayolle L. van Tuinen D. Chatagnier O. Li X. Gianinazzi S. Gianinazzi-Pearson V. 2012 Local and systemic mycorrhiza-induced protection against the ectoparasitic nematode Xiphinema index involves priming of defence gene responses in grapevine Journal of Experimental Botany 63 3657 3672 https://doi.org/10.1093/jxb/ers046 22407649 Hernández-Pacheco C. E. del Carmen Orozco-Mosqueda M. Flores A. Valencia-Cantero E. Santoyo G. 2021 Tissue-specific diversity of bacterial endophytes in Mexican husk tomato plants (Physalis ixocarpa Brot. ex Horm.), and screening for their multiple plant growth-promoting activities Current Research in Microbial Sciences 2 100028 https://doi.org/10.1016/j.crmicr.2021.100028 34841319 Hu H. Wang C. Li X. Tang Y. Wang Y. Chen S. Yan S. 2018 RNA-Seq identification of candidate defense genes targeted by endophytic Bacillus cereus-mediated induced systemic resistance against Meloidogyne incognita in tomato Pest Management Science 74 2793 2805 https://doi.org/10.1002/ps.5066 29737595 Izuogu N. Gbenle M. Yakubu I. Abolusoro S. 2015 Reaction of some selected soybean varieties (Glycine max (L) Merril) to root-knot nematode infection Ethiopian Journal of Environmental Studies and Management 8 541 547 https://doi.org/10.4314/ejesm.v8i5.7 Lecouls A.-C. Bergougnoux V. Rubio-Cabetas M.-J. Bosselut N. Voisin R. Poessel J.-L. Faurobert M. Bonnet A. Salesses G. Dirlewanger E. 2004 Marker-assisted selection for the wide-spectrum resistance to root-knot nematodes conferred by the Ma gene from Myrobalan plum (Prunus cerasifera) in interspecific Prunus material Molecular Breeding 13 113 124 https://doi.org/10.1023/B:MOLB.0000018758.56413.cf Lesmes-Vesga R. A. Cano L. M. Ritenour M. A. Sarkhosh A. Chaparro J. X. Rossi L. 2022 Rootstocks for commercial peach production in the southeastern United States: current research, challenges, and opportunities Horticulturae 8 602 https://doi.org/10.3390/horticulturae8070602 Li Q. Hou Z. Zhou D. Jia M. Lu S. Yu J. 2022 A plant growth-promoting bacteria Priestia megaterium JR48 induces plant resistance to the crucifer black rot via a salicylic acid-dependent signaling pathway Frontiers in Plant Science 13 4616 https://doi.org/10.3389/fpls.2022.1046181 Li Y. Zhong Y. Huang K. Cheng Z.-M. 2016 Genomewide analysis of NBS-encoding genes in kiwi fruit (Actinidia chinensis) Journal of Genetics 95 997 1001 https://doi.org/10.1007/s12041-016-0700-8 27994200 Liu Y. Yao S. Deng L. Ming J. Zeng K. 2019 Different mechanisms of action of isolated epiphytic yeasts against Penicillium digitatum and Penicillium italicum on citrus fruit Postharvest Biology and Technology 152 100 110 https://doi.org/10.1016/j.postharvbio.2019.03.002 Lizardo R. C. M. Pinili M. S. Diaz M. G. Q. Cumagun C. J. R. 2022 Screening for resistance in selected tomato varieties against the root-knot nematode Meloidogyne incognita in the Philippines using a molecular marker and biochemical analysis Plants 11 1354 https://doi.org/10.3390/plants11101354 35631779 Ma J.-T. Li D.-W. Liu J.-K. He J. 2021 Advances in research on chemical constituents and their biological activities of the genus Actinidia Natural Products and Bioprospecting 11 573 609 https://doi.org/10.1007%2Fs13659-021-00319-8 34595735 Maghdouri M. Ghasemnezhad M. Rabiei B. Golmohammadi M. Atak A. 2021 Optimizing seed germination and seedling growth in different kiwifruit genotypes Horticulturae 7 314 https://doi.org/10.3390/horticulturae7090314 Maheshwari D. K. Annapurna K. 2017 Endophytes: crop productivity and protection Sustainable Development and Biodiversity 16 61 72 https://doi.org/10.1007/978-3-319-66544-3 Mahoonaki F. S. Moghadam E. M. Zakiaghl M. Pedram M. 2023 Penetration and development of Meloidogyne javanica on four pistachio rootstocks and their defense responses Journal of Nematology 54 20220056 https://doi.org/10.2478%2Fjofnem-2022-0056 36879951 Mohammadi P. Elif T. Recep K. Şenol K. M. 2017 Potential of some bacteria for biological control of postharvest citrus green mould caused by Penicillium digitatum Plant Protection Science 53 134 143 https://doi.org/10.17221/55/2016-PPS Mostafa F. A. Khalil A. E. Nour El-Deen A. H. Ibrahim D. S. 2018 The role of Bacillus megaterium and other bio-agents in controlling root-knot nematodes infecting sugar beet under field conditions Egyptian Journal of Biological Pest Control 28 1 6 https://doi.org/10.1186/s41938-018-0068-6 Moyes A. B. Kueppers L. M. Pett-Ridge J. Carper D. L. Vandehey N. O’Neil J. Frank A. C. 2016 Evidence for foliar endophytic nitrogen fixation in a widely distributed subalpine conifer New Phytologist 210 657 668 https://doi.org/10.1111/nph.13850 27000956 Mukhtar T. 2018 Management of root-knot nematode, Meloidogyne incognita, in tomato with two Trichoderma species Pakistan Journal of Zoology 50 1589 1592 http://dx.doi.org/10.17582/journal.pjz/2018.50.4.sc15 Mukhtar T. Hussain M. A. Kayani M. Z. Aslam M. N. 2014 Evaluation of resistance to root-knot nematode (Meloidogyne incognita) in okra cultivars Crop Protection 56 25 30 https://doi.org/10.1016/j.cropro.2013.10.019 Naylor D. DeGraaf S. Purdom E. Coleman-Derr D. 2017 Drought and host selection influence bacterial community dynamics in the grass root microbiome The ISME journal 11 2691 2704 https://doi.org/10.1038/ismej.2017.118 28753209 Nicotra A. Simeone A. De Vito M. 2003 Research on kiwifruit source of genetic resistance to root-knot and lesion nematodes 449 453 Proceedings of the fifth international symposium on kiwifruit Leuven, Belgium International Society for Horticultural Science Pan L. Zhao X. Chen M. Fu Y. Xiang M. Chen J. 2020 Effect of exogenous methyl jasmonate treatment on disease resistance of postharvest kiwifruit Food Chemistry 305 125483 https://doi.org/10.1016/j.foodchem.2019.125483 31610420 Rosier A. Medeiros F. H. Bais H. P. 2018 Defining plant growth promoting rhizobacteria molecular and biochemical networks in beneficial plant-microbe interactions Plant and Soil 428 35 55 https://doi.org/10.1007/s11104-018-3679-5 Rubio-Cabetas M. Lecouls A. Salesses G. Bonnet A. Minot J. Voisin R. Esmenjaud D. 1998 Evidence of a new gene for high resistance to Meloidogyne spp. in Myrobalan plum, Prunus cerasifera Plant Breeding 117 567 571 https://doi.org/10.1111/j.1439-0523.1998.tb02209.x Salesses G. Dirlewanger E. Esmenjaud D. Lecouls A. 1998 Root knot nematode resistance in Myrobalan plum: inheritance and rootstock breeding perspectives using marker-assisted selection 45 52 Grzyb Z. S. Zmarlicki K. VI international symposium on plum and prune genetics, breeding, pomology Leuven, Belgium International Society for Horticultural Science Santos M. L. d. Berlitz D. L. Wiest S. L. F. Schünemann R. Knaak N. Fiuza L. M. 2018 Benefits associated with the interaction of endophytic bacteria and plants Brazilian Archives of Biology and Technology 61 https://doi.org/10.1590/1678-4324-2018160431. Saucet S. B. Van Ghelder C. Abad P. Duval H. Esmenjaud D. 2016 Resistance to root-knot nematodes Meloidogyne spp. in woody plants New Phytologist 211 41 56 https://doi.org/10.1111/nph.13933 27128375 Su L. Shen Z. Ruan Y. Tao C. Chao Y. Li R. Shen Q. 2017 Isolation of antagonistic endophytes from banana roots against Meloidogyne javanica and their effects on soil nematode community Frontiers in Microbiology 8 2070 https://doi.org/10.3389/fmicb.2017.02070 29123509 Taechowisan T. Peberdy J. F. Lumyong S. 2003 Isolation of endophytic actinomycetes from selected plants and their antifungal activity World Journal of Microbiology & Biotechnology 19 381 385 https://doi.org/10.1023/A:1023901107182 Tanha Maafi Z. Mahdavian S. 1997 Species and physiological races of root knot nematodes (Meloidogyne spp.) on kiwifruit and the effect of M. incognita on kiwifruit seedlings Applied Entomology and Phytopathology 65 1 3 Taylor A. L. Sasser J. N. 1978 Biology, identification and control of root-knot nematodes (Meloidogyne species) Taylor A. Sasser J. N. Biology, identification and control of root-knot nematodes (Meloidogyne species) Raleigh, NC Dept. of Plant Pathology, North Carolina State University Tran T. P. H. Wang S.-L. Nguyen V. B. Tran D. M. Nguyen D. S. Nguyen A. D. 2019 Study of novel endophytic bacteria for biocontrol of black pepper root-knot nematodes in the central highlands of Vietnam Agronomy 9 714 https://doi.org/10.3390/agronomy9110714 Vetrivelkalai P. 2019 Evaluation of endophytic bacterial isolates against root knot nematode, Meloidogyne incognita in tomato under glasshouse condition International Journal of Current Microbiology Applied Science 8 2584 2589 https://doi.org/10.20546/ijcmas.2019.801.271 Whitehead A. Hemming J. 1965 A comparison of some quantitative methods of extracting small vermiform nematodes from soil Annals of Applied Biology 55 25 38 10.1111/j.1744-7348.1965.tb07864.x Wicaksono W. A. Jones E. E. Casonato S. Monk J. Ridgway H. J. 2018 Biological control of Pseudomonas syringae pv. actinidiae (Psa), the causal agent of bacterial canker of kiwifruit, using endophytic bacteria recovered from a medicinal plant Biological Control 116 103 112 https://doi.org/10.1016/j.biocontrol.2017.03.003 Xie S. Zang H. Wu H. Uddin Rajer F. Gao X. 2018 Antibacterial effects of volatiles produced by Bacillus strain D13 against Xanthomonas oryzae pv. oryzae Molecular Plant Pathology 19 49 58 https://doi.org/10.1111/mpp.12494 27682316 Youssef M. Abd Abd-El-Khair H. El-Nagdi W. M. 2017 Management of root knot nematode, Meloidogyne incognita infecting sugar beet as affected by certain bacterial and fungal suspensions Agricultural Engineering International: CIGR Journal Special issue 293 301

PMC010xxxxxx/PMC10241308.txt

==== Front J Nematol J Nematol jofnem jofnem Journal of Nematology 0022-300X 2640-396X Sciendo 37283998 jofnem-2023-0016 10.2478/jofnem-2023-0016 Research Paper First Report of Direct Damage Caused by the Stubby-Root Nematode, Nanidorus minor, to Strawberry (Fragaria x ananassa), in Florida Oliveira Clemen J. clemenjdo@gmail.com Inserra Renato N. Desaeger Johan A. Department of Entomology and Nematology, Gulf Coast Research and Education Center, University of Florida, Wimauma, FL, 33598, USA. Florida Department of Agriculture and Consumer Services, DPI, Nematology Section, P.O. Box 147100, Gainesville, FL 32614-7100, USA. This paper was edited by Guiping Yan. 2 2023 5 6 2023 55 1 2023001621 11 2022 © 2023 Clemen J. Oliveira et al., published by Sciendo 2023 Clemen J. Oliveira et al., published by Sciendo https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. Abstract In 2019–2022, declining symptoms were observed in two commercial strawberry farms in Hillsborough County, Florida. The fields in the two farms consisted of raised beds covered by plastic mulch. Both were fumigated with a mixture of 1,3-dichloropropene (40%) + chloropicrin (60%) before planting. Samples collected from large patches with declining plants were infested with stubby-root nematodes. No sting and root-knot nematode species were detected. The results of morphological and molecular analyses indicated that the stubby-root nematode populations were representative of the species Nanidorus minor. The two cultivars ‘Florida Brilliance’ and ‘Florida Sensation’ in the two fields included plants with stubby root symptoms showing a reduction in the size of the root system and arrested growth and elongation of the feeder roots on the first strawberry crop. The nematode population densities in the two fields increased at the end of strawberry season and averaged 66 and 96 specimens in 200 cm3 soil. In one of the fields, a second strawberry crop was established as in the previous year using the same practices (fumigation and raised beds covered with plastic). However, in this field the population of N. minor declined and did not reach damaging levels at the end of the season on the second strawberry crop. The factors causing the decline of the nematode population were not elucidated. This is the first report of a direct damaging effect of N. minor to strawberry. Keywords Fumigated soil host-parasitic relationship nematode parasite of strawberry nematode population densities root symptoms ==== Body pmcIn Florida, strawberry production is an important component of the vegetable industry with an annual value of $282 million (USDA-NASS, 2017). Many factors negatively affect strawberry production in the state, such as diseases and pests, including plant-parasitic nematodes (PPN). The most damaging PPN species to strawberry in order of economic importance are sting nematodes (Belonolaimus longicaudatus Rau, 1958) and root-knot nematode species in the genus Meloidogyne (Göldi, 1887) Chitwood, 1949. Among these species, M. hapla Chitwood, 1949 is the most common because it was introduced into Florida fields with nematode-infested propagative runners (Desaeger, 2019). Although association of stubby-root nematodes, such as Nanidorus minor (Colbran, 1956) Siddiqi, 1980 with strawberry (Fragaria x ananassa Duchesne) has been reported by Rhode and Jenkins (1957), there are no records of infestations of stubby-root nematode species (Trichodoridae (Thorne, 1955) Clark, 1961) causing damage to strawberry in the literature. In 2020, symptoms of decline, consisting of stunted plants showing reduced root systems with a low number of short feeder roots, were observed in a field where preliminary analyses of a few samples indicated that stubby-root nematode species were the predominant PPN. Infestations of stubby-root nematodes are common in Florida fumigated fields (especially tomato and strawberry fields) because these nematodes are more tolerant to currently used fumigants, or because they escape the effects of the fumigants by migrating into the deep layers of the soil not reached by the available fumigants after the loss of the methyl bromide (Weingartner and Shumaker, 1990; Weingartner et al., 1993). A field investigation was conducted with the objectives to determine: i) the involvement of stubby-root nematode species in the crop decline, ii) the nematode species inducing the decline, and iii) the symptoms induced on strawberry by the nematode infestation. Two fumigated farms located in the same strawberry producing areas of Central Florida were selected for this study and indicated as Farm 1 and 2 for convenience. In Farm 1, where the first declining symptoms were noticed, certified runners of the strawberry cultivar ‘Florida Brilliance,’ imported from a nursery in Quebec, Canada, were transplanted in October 2019, in raised beds covered by a totally impermeable field plastic (TIF Total Blockade, Berry Plastics Corporation, Evansville, Indiana) after fumigation with a mixture of 1,3-dichloropropene (40 %) + chloropicrin (60 %) (PicClor60 @ 336 kg/ha). In this farm, observations were conducted on two strawberry crops, the first in 2020 and the second in 2021. In Farm 2, certified runners of the strawberry cultivar ‘Florida Sensation,’ imported from a nursery in California, were transplanted in October 2021 in raised beds covered by a totally impermeable field plastic after fumigation with the same fumigant. In this farm, the observations were conducted only on one strawberry crop in 2022. Soil and root samples were collected from the two farms at different time intervals as specified below. Plant-parasitic nematodes were extracted from 200 cm3 soil using a modified Baermann funnel technique (Rodríguez-Kábana and Pope, 1981) and identified using a compound microscope. Specimens were hand-picked in tap water, immobilized by gentle heating, and mounted in water agar on a slide for measurements and photographs using a modified technique described by Esser (1986). Measurements of specimens were made using a Nikon (Optiphot) ocular micrometer. Additional specimens were sent to Dr. Sergei Subbotin for sequencing and phylogenetic analysis to confirm our morphological identification. Root symptoms were examined, and photographs were taken using a digital microscope Keyence (VHX-7000 series, KEYENCE CORPORATION, Itasca, Illinois). In Farm 1, the declining symptoms were observed five months after transplanting (March 16, 2020), at the end of the season of the first strawberry crop, when soil and root samples, for a total of 17, were collected. In 2021, after soil fumigation and planting of the runners of second strawberry crop, another three and three samples of soil mixed with roots were collected at midseason (January 2021) and the end of the season (March 2021) of the second strawberry crop, respectively. The aim of this sampling was to verify the persistence of the nematodes in the field and the reoccurrence of symptoms they caused on the second strawberry crop. In Farm 2, a total of two and five samples were collected as mentioned above from one strawberry crop only, at midseason (January 2022), and the end of the season (March 2022), respectively. Nematode populations extracted from the samples collected in the two farms consisted mainly of stubby-root nematodes. No root-knot and sting nematodes were found associated with the stubby-root nematodes. The morphology of the stubby-root nematodes fitted that of Nanidorus minor Siddiqi 1980. The morphological characters of eight females of this population included: Body length (L) = 700.9 ± 78.8 μm (619.1 − 769); Stylet length (ST) = 34.1 ± 1.1 μm (32.2–35.6); Body length/greatest body diameter (a) = 17 ± 1.3 μm (15.5–19.8); Body length/distance from anterior end to posterior end of median pharyngeal bulb (b) = 5.1 ± 0.3 (4.8–5.6); Distance of anterior body end from the vulva/body length % (V) = 51.7 ± 1 (50–53.5). A phylogenetic analysis using the D2-D3 expansion fragments of 28S rRNA gene with 58 valid and putative species of stubby root nematode revealed that the strawberry population collected in Hillsborough County was 99.18 to 100% identical to other populations of N. minor (Subbotin et al. 2020). The results of the morphological analysis were confirmed by those of the phylogenetic analysis that were published by Subbotin et al. (2020). In Farm 1, the average number of stubby-root nematodes collected at the end of the season of the first strawberry crop, in 2020 was 96/200 cm3 soil. These population densities were associated with stunted plants showing reduced root systems with a low number of short-feeder roots, consistent with stubby-root symptoms like those described for N. minor by Christie and Perry (1951) on other crops (Figure 1). Stunted plants were localized in large areas (8.1 meter wide and 13.7 meter long) in the field. The populations densities detected for the second strawberry crop at the mid- and end of season in 2021 were lower than those in the previous crop in 2020, and averaged 7 and 4/200 cm3 soil, respectively. No reoccurrence of the stunting symptoms was noticed on the second strawberry crop at these nematode densities. These findings indicated the populations of N. minor did not increase to damaging level during the second strawberry cycle. Figure 1: Strawberry damage caused by N. minor. (A) Plant stunting and canopy reduction observed in a commercial strawberry field; (B) Stubby root in strawberry; (C) Digital microscopy (magnification 200x) of a section of a parasitized strawberry root tip; (D) Digital microscopy of a parasitized strawberry root system (magnification 20x). In Farm 2, an average of 10 and 66 specimens/200cm3 soil was found in January 2022 and March 2022, respectively. In this field, the population density increased at the end of the 2021–2022 season, resulting in visual field damage localized in large areas (13.8 meter wide and 17.4 meter long) as observed in Farm 1. In both farms, strawberry root damage occurred at the end of the season when the highest population densities were recorded. The results of this field study provide evidence that N. minor can induce serious damage to strawberry comparable to that reported in fields infested with sting nematodes. However, in contrast to observations on sting nematodes, the populations of this species did not increase consistently on the second strawberry crop. Data in our study are not sufficient to explain the decline of N. minor populations on the second strawberry crop cycle. We can hypothesize that N. minor populations localized in the top layers of the soil during the second strawberry cycle were more vulnerable to the adverse effect of the second application of the fumigant, which prevented nematode reproduction. ==== Refs Literature Cited Christie J. R. Perry V. G. 1951 A root disease of plants caused by a nematode of the genus Trichodorus Science 113 491 493 Desaeger J. 2019 Meloidogyne hapla, the Northern Root-knot Nematode, in Florida Strawberries and Vegetables. ENY070 Gainesville University of Florida Institute of Food and Agricultural Sciences 2019 https://edis.ifas.ufl.edu/in1224 Esser R. P. 1986 A water agar in face technique Proceedings of the Helminthological Society of Washington 53 254 255 Rodríguez-Kábana R. Pope M. H. 1981 A simple incubation method for the extraction of nematodes from soil Nematropica 11 175 186 Rohde R. A. Jenkins W. R. 1957 Host range of a species of Trichodorus and its host-parasite relationships on tomato Phytopathology 47 295 298 Subbotin S. A. Cid del Prado Vera I. Inserra R. N. Chizhov V. N. Decraemer W. 2020 Molecular characterization of some stubby root nematodes (Nematoda: Trichodoridae) from the USA and other countries Nematology 22 39 57 USDA-NASS 2017 United States Department of Agriculture, National Agricultural Statistics Service https://www.nass.usda.gov Weingartner D. Shumaker J. 1990 Effects of soil fumigants and aldicarb on corky ringspot disease and Trichodorid nematodes in potato Journal of Nematology 22 665 71 19287777 Weingartner D. McSorley R. Goth R. 1993 Management strategies in potato for nematodes and soil-borne diseases in subtropical Florida Nematropica 23 233 45

PMC010xxxxxx/PMC10241948.txt

==== Front Sci Rep Sci Rep Scientific Reports 2045-2322 Nature Publishing Group UK London 37277428 35637 10.1038/s41598-023-35637-z Article Universal mtDNA fragment for Cervidae barcoding species identification using phylogeny and preliminary analysis of machine learning approach http://orcid.org/0000-0003-2313-8398 Filip Ewa ewa.filip@usz.edu.pl 12 http://orcid.org/0000-0002-7761-1630 Strzała Tomasz 3 http://orcid.org/0000-0002-5638-7676 Stępień Edyta 4 http://orcid.org/0000-0002-1503-0064 Cembrowska-Lech Danuta 15 1 grid.79757.3b 0000 0000 8780 7659 Institute of Biology, University of Szczecin, Wąska 13, 71-415 Szczecin, Poland 2 grid.79757.3b 0000 0000 8780 7659 The Centre for Molecular Biology and Biotechnology, University of Szczecin, Szczecin, Poland 3 grid.411200.6 0000 0001 0694 6014 Department of Genetics, Faculty of Biology and Animal Science, Wrocław University of Environmental and Life Sciences, Wrocław, Poland 4 grid.79757.3b 0000 0000 8780 7659 Institute of Marine and Environmental Sciences, University of Szczecin, Adama Mickiewicza 16, 70-383 Szczecin, Poland 5 Sanprobi Sp. z o. o. Sp. k., Kurza Stopka 5C, 70-535 Szczecin, Poland 5 6 2023 5 6 2023 2023 13 91331 2 2023 21 5 2023 © The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The aim of the study was to use total DNA obtained from bone material to identify species of free-living animals based on the analysis of mtDNA fragments by molecular methods using accurate bioinformatics tools Bayesian approach and the machine learning approach. In our research, we present a case study of successful species identification based on degraded samples of bone, with the use of short mtDNA fragments. For better barcoding, we used molecular and bioinformatics methods. We obtained a partial sequence of the mitochondrial cytochrome b (Cytb) gene for Capreolus capreolus, Dama dama, and Cervus elaphus, that can be used for species affiliation. The new sequences have been deposited in GenBank, enriching the existing Cervidae mtDNA base. We have also analysed the effect of barcodes on species identification from the perspective of the machine learning approach. Machine learning approaches of BLOG and WEKA were compared with distance-based (TaxonDNA) and tree-based (NJ tree) methods based on the discrimination accuracy of the single barcodes. The results indicated that BLOG and WEKAs SMO classifier and NJ tree performed better than TaxonDNA in discriminating Cervidae species, with BLOG and WEKAs SMO classifier performing the best. Subject terms Molecular biology Environmental sciences issue-copyright-statement© Springer Nature Limited 2023 ==== Body pmcIntroduction The current state of knowledge of molecular biologists has led to the widespread use of mitochondrial DNA (mtDNA) as a marker for species-specific identification in animals1–5. For intraspecific detection of unrelated individuals, sequences with high variability are recommended, e.g. certain nuclear genes6. For species identification within Cervidae, we choose conservative sequences shared among the animals with species-specific variables, because doing so brings the best effect7–9. Mitochondrial DNA is known to be an effective molecular marker in phylogenetic analyses10,11. This is due to the high polymorphism of the control region, as well as a lack of recombination, and very good isolation efficiency, even from small amounts of biological tissue, as well as the resistance of mtDNA to degradation processes. Different gene regions such as mtDNA, have been used for the DNA barcoding approach, but cytochrome c oxidase COI) is a universal DNA barcode for animals12, such Cervidae13,14. Ward et al.15 analysed mtDNA COI sequences in animals and indicated that the success of barcoding depended upon recent speciation, incorrect morphological taxonomy and species hybridisation, where barcoding could not differentiate interspecies. There are many drawbacks to the use of a barcoding tool for species identification, so, the scientific community must be cautious in accepting the above factors and use additional genes for further clarification. Generally, biological phenomena, such as the hybridisation process of close species, natural introgression process, error in specimen identification using classical taxonomy and recent speciation process, are known to strongly interfere with the DNA barcoding process, and these phenomena are known to occur at different degrees depending on the animal groups and datasets16–19. So, it is authoritative to have not only more databases on individual species COI gene sequences from different geographical locations and correct identification of species through traditional taxonomy. It should also be noted that taxonomic decisions based on a single molecular marker that is maternally inherited might not resolve all species identification and should be supported by a second molecular marker, such as cytochrome b (Cytb). The analysis of species-specific variation using the homologous cytochrome b (Cytb) is characterized by high reproducibility and sensitivity of results11,20–24. To distinguish closely related species, selected mtDNA fragments with very high specificity are needed. Often conservative gene sequences encoding proteins are used in studies on interspecies diversity25, while the control region is used to provide a reliable source of knowledge about intraspecific variability7,10,25. In some cases, cytochrome b provides excellent phylogenetic information on the taxonomic position of various vertebrates; and thus, it can be used in the analysis of live specimens or for forensic identification purposes26–29, with the same success rate as COI30. In addition, this gene is often considered when determining the origins of samples from difficult biological materials, i.e. hair, feathers, tooth fragments or other bones, which mainly utilize mitochondrial DNA polymorphisms26. Irwin et al.31 have determined the rate of evolutionary changes for the genera of some species in different components of cytochrome b amino acid sequences based on fossil DNA analyses. Several recent studies show that when the DNA template is derived from bone material, a 300–500 bp Cytb fragment, is suitable for mammalian species identification12,27,32–34. Machine learning (ML) is a branch of artificial intelligence (AI) where machines are trained to solve self-designed problems by learning new rules through repeated trials and feedback. ML enables inferring of models or relationships by learning from data. With the development of AI, machine learning has been rapidly developed and applied in DNA barcoding. Over the years, several different analytical methods were devised for the assessment of the species discrimination ability, such as TaxonDNA, NJ tree and machine learning approaches (BLOG and WEKA). Most machine learning approaches (MLA) bear its origin from statistical methods of regression analysis. Machine learning approaches are computer tools, which can be successfully applied in species identification35. BLOG (Barcoding with LOGic) and WEKA (Waikato Environment for Knowledge Analysis) are methods of ML, which can recognize unknown species (query set) present in the reference dataset composed of DNA barcode sequence (training set) of known species36,37. Based on the literature, we have found the cytochrome b mitochondrial gene to be useful in identifying species of wild animals using bone material (mandible, frontal bone) and we tested its usefulness on the real life example. The specific aim of our research was to develop a short universal fragment from mtDNA, which could be used in the species identification of various deer populations. In addition, we proposed the use of machine learning approach methods to classify species with DNA barcode sequences. Results DNA isolation Table 1 show a spectrophotometer readings on DNA isolates, giving OD 260/280 ratios ranging from 1.8 to 2.33, for the 18 different bone fragments. Among the studied samples, values exceeding 2.0 were obtained for several samples: two fragments of red deer bone marked KBMICSZ3 (2.01), the fallow deer bone fragment KBMICSZ13 (2.21), and a fragment of roe deer frontal bone marked KBMICSZ25 (2.33).Table 1 List of the study materials and results of DNA isolations. No. Sample ID Species Type of bone Weight [g] DNA concentration [ng/μL)] A260/A280 ratio 1 KBMICSZ1 Cervus elaphus Mandible 0.28 216.8 1.95 2 KBMICSZ2 Cervus elaphus Mandible 0.26 386.0 1.85 3 KBMICSZ3 Cervus elaphus Mandible 0.26 91.1 2.01 4 KBMICSZ9 Cervus elaphus Mandible 0.35 142.8 1.85 5 KBMICSZ20 Cervus elaphus Mandible 0.24 259.8 1.83 6 KBMICSZ4 Cervus elaphus Mandible 0.38 180.2 1.87 7 KBMICSZ5 Capreolus capreolus Frontal bone 0.36 1264.2 1.82 8 KBMICSZ6 Capreolus capreolus Frontal bone 0.34 661.8 1.83 9 KBMICSZ8 Capreolus capreolus Frontal bone 0.28 187.8 1.85 10 KBMICSZ14 Capreolus capreolus Frontal bone 0.35 474.0 1.85 11 KBMICSZ15 Capreolus capreolus Frontal bone 0.36 242.3 1.88 12 KBMICSZ16 Capreolus capreolus Frontal bone 0.39 289.4 1.86 13 KBMICSZ21 Capreolus capreolus Frontal bone 0.30 782.9 1.81 14 KBMICSZ22 Capreolus capreolus Frontal bone 0.30 123.2 1.87 15 KBMICSZ24 Capreolus capreolus Frontal bone 0.30 119.1 1.97 16 KBMICSZ25 Capreolus capreolus Frontal bone 0.30 12.9 2.33 17 KBMICSZ7 Dama dama Mandible 0.33 353.7 1.83 18 KBMICSZ13 Dama dama Mandible 0.28 12.5 2.21 The place of collection samples: No. 1–6 and 17–18 Plecemin: 53°16′31.931″ N 16°48′30.246″ E; No. 7–8 Drawsko Pomorskie 53°31.8336′N 15°48.5802′E; No. 9–12 Stara Korytnica 53°18′2.902″ N 16°2′15.199″ E; No. 13–14 Karwowo 53°41′26.002" N 15°33′2.002″ E; No. 15–16 Dorowo 53°43′20.775″ N 15°27′26.097″ E. Species identification of analysed DNA sample As a result of performing PCR and DNA sequencing on the collected deer samples, 18 sequences of the Cytb gene were obtained, which helped in the identification of each species belonging to the Cervidae family. This analysis involved 18 nucleotide sequences with a total of 207 positions in the final dataset. The average GC content was 50%. The Cytb region was characterized by a high level of monomorphism with a small number of 163 sites and polymorphic sites a number of 44 and a number of parsimoniously informative sites number (PIC) of 2. Based on the whole length of the Cytb gene sequenced, a total of 5 haplotypes were detected with a Hd (Haplotype diversity) equal to 0.771. The most frequent haplotype was Hap_3, which were found among 7 individuals. It should be mentioned that the type of genetic frequency of these haplotypes in North-western Poland Cervidae haplotypes Hap_1: and Hap_2: for Cervus elaphus, Hap_3: and Hap_4: for Capreolus capreolus, and Hap_5 was for Dama dama (Table S3). Figure 1 show the obtained phylogenetic tree, which was resolved into three distinct clades that consisted of representatives of the three analysed species. Samples were grouped together with each species representative showing a high probability (100%) of assignment, indicating clear species identification. Within the clades, we found substantial polytomy, which is a result of a lack of sequence informativity within the species level.Figure 1 Bayesian phylogenetic tree showing species identification of analysed DNA samples (samples are indicated with a star). Sequences of Antidorcas marsupialis and Beatragus hunteri were used for rooting. Numbers along nodes are the posterior probability values of nodes. Tree was generated with MrBayes 3.2.638. Another phylogenetic analysis inferred from the Cytb sequences was constructed to illustrate the phylogenetic relationship of Cervidae species based on Cytb sequences from GenBank and our studies (Fig. 2). A phylogenetic ML tree was constructed using a pre-trained neural network. The tree important Cervidae species were formed monophyletic clades and found well-supported with bootstrap values (> 80%).Figure 2 Phylogenetic tree inferred from the Cytb sequences. Results from the ML and the MP analyses were mapped onto the NJ tree. Tree was generated with FigTree 1.4.439. Machine learning approaches of BLOG and WEKA were compared with distance-based (TaxonDNA) and tree-based (NJ tree) methods based on discrimination accuracy and cost-effectiveness of barcode gene (Table 2, Fig. 2). The results indicated that BLOG and WEKAs SMO classifier and NJ tree performed better than TaxonDNA in discriminating Cervidae species. Specifically, the single barcode Cytb exhibited the highest species resolution (100%) for identifying 3 Cervidae species when BLOG or WEKAs SMO classifiers were used. This study showed that machine learning approaches provided higher discrimination accuracy and cost-effectiveness over other analytical methods in DNA barcoding of Cervidae species.Table 2 Species resolution success rates for the Cervidae based on different analytical methods. Rates, % BLOG WEKA TaxonDNA NJ tree Naïve Bayes SMO Jrip J48 Correctly identify 100.00 60.00 100.00 94.00 94.00 90.00 100.00 Misidentify 0.00 40.00 0.00 6.00 6.00 10.00 0.00 Not identify 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Moreover, cytb region also provided the highest accuracy of species discrimination (100%) when using WEKAs SMO classifier (Fig. 3).Figure 3 Confusion matrix of WEKAs SMO classifier generated by cytb showing classification results of Cervidae species, obtained with R environment40. Discussion We present that DNA barcoding is an effective molecular tool for Cervidae species identification and phylogenetic inferences as a result of our research, we have obtained the new sequences that have been deposited in a section of the Genbank belonging to the Cervidae family. Identification of species must be effective and precise even from degraded environmental material. In addition, well-resolved molecular phylogenies derived from these DNA barcode sequences have the potential to improve investigations of the mechanisms underlying community assembly and functional trait evolution. Our proposed methodology can be used in the future as a routine marker in cases when degraded samples will be used. It also contributes to the development of the subject of species identification in different environments. With the development of DNA barcoding, several analytical methods were developed for the assessment of species discrimination ability. There are no criteria for evaluating the quality of the analytical methods for species discrimination. We have analyzed the effect of barcodes on species identification from the perspective of the machine learning approach (MLA). We tested a reinforcement-learning algorithm to solve the challenge of reconstructing phylogenetic trees, which are used to describe the relationships among a set of sequences. Current tools for phylogenetic- tree reconstruction integrate heuristic approaches to evaluate only a subset of all potential trees, thus they suffer from the known trade-off between accuracy and running time. In our study, we tested the methodology for predicting the maximum likelihood tree. Our preliminary results, based on a machine learning approach algorithm demonstrate that the trained algorithm can accurately and efficiently reconstruct maximum-likelihood trees. This development technique ML could provide rapid, simple, and reliable tools for species confirmation and can be applied to the modelling of species distribution. In the present study, we proposed and tested the use of a relatively short mitochondrial DNA sequence for the species identification of members of the genus Cervidae. Our proposed methodology, based on machine learning, confirmed the identification results of the classical phylogeny-based approach, which will enable its wider use in future routine studies of this type. Nowadays, DNA analysis of biological samples has become the standard practice in animal identification at various taxonomic levels. Different types of tissues, such as bones, blood, hair (fur), feathers, skin, meat (muscle sample), faecal, and others are often the subject of many studies in various DNA analysis laboratories22,41. A universal fragment of genetic information is constantly being sought to use in many areas, e.g. poaching26,42,43, illicit trafficking of endangered species26,42–44, protection of endangered animal species27, or determination of meat origin (for identification purposes)44,45. Anna Ramón-Laca et al.22, showed that differentiation of species can be achieved by using a species-specific primer that amplifies dissimilar length fragments. There are differences of opinion among the researchers, regarding which of the markers, COI or Cytb, provides more reliable and reproducible results for DNA barcoding analysis. In 2010, a group of researchers led by Tobe et al.1 carried out an assessment of genetic intraspecific variability based on COI and Cytb sequences from 217 mammalian species. The results showed that the discriminatory power was higher for the Cytb gene, i.e. there was a higher probability that two random individuals from a given population would have sequence differences at the marker locus than for the COI sequence. Research carried out by Wilson-Wilde et al.46 demonstrated that identification based on the COI gene sequence is suitable for genetically distant species, while in the case of closely related species, it is no longer unambiguous and requires additional tests. However, COI, Cytb, and the mt-CR control region are still used for this purpose1,20–24,47–50. To find better molecular tools the compilation of known DNA markers led to the construction of the genetic map of Cervus elaphus51–53. This genetic map comprises 621 sites (length of 2532 cM, with average intervals of 5.7 cM), and it integrates modern technologies and research methods, including comparative genomics and orthologous alleles of DNA markers derived from ruminants and other mammals (i.e. Pere David’s deer, Elaphurus davidianus and red deer, C. elaphus)54. The genetic map of deer was used as an annotation for further research, such as the origin and evolution of ruminant genomes52, QTL scanning53, SNP analyses of the whole genome55,56 and whole genome sequencing as well as the annotation and assembly of pseudochromosomes54. In the results presented in this study, the total length of DNA fragments of all analysed individuals was 207 bp due to the removal of the last nucleotides in the sequences. The reason for obtaining different lengths was probably due to the inhibition of sequencing reactions by individual matrices. Similar results were obtained by Gupta et al.33, who worked on stool samples and also obtained short Cytb sequence fragments of 366, 374 and 503 bp22,41,57. We show that when using bone tissue, the primers used in this work for the Cytb gene fragment amplifications work better because firstly, they differentiate closely related species well and have the additional advantage that they can be used for many other mammalian species as well. Our research is confirmed by many studies, not only for the family Cervidae but also other works on the identification of other wild mammalian species1,5,32,41,48,58. Our phylogenetic analysis grouped the analysed sequences within individual species with 100% probability (Fig. 1). The Cytb fragment, analysed in this study, allows correct species identification, however, the lack of intraspecific polymorphism results in the inability to use it in population studies. This is clearly shown in the phylogenetic tree obtained (Fig. 1) where most monospecific nodes are polytomous. The lack of the node’s solution (polytomy) is in this case is the result of a lack of genetic information from the analysed DNA sequences (soft polytomy). Our results suggest that the intraspecific genetic polymorphism is low for all mammalian species. Similar results were obtained in earlier studies1,59. MLAs extract the distinct features from the DNA sequences by training the reference dataset and then used for identifying the query sequences. Cytb region provided the highest species resolution when using the BLOG method, and WEKAs SMO classifier. In the comparison of TaxonDNA and NJ tree, the BLOG and SMO methods produced a relatively low level of misidentification. Moreover, though the above methods have to achieved > 90% species identification success rate, still there is a need for further improvement in success rate (Fig. 2, Table 2). The discrimination ability of combined different barcodes in the species of the Cervidae genus is still fully unknown. We present that DNA barcoding is an effective molecular tool for Cervidae species identification and phylogenetic inferences a result of our research. Identification of species must be effective and precise even from degraded environmental material. In addition, well-resolved molecular phylogenies derived from these DNA barcode sequences have the potential to improve investigations of the mechanisms underlying community assembly and functional trait evolution. Our proposed methodology can be used in the future as a routine marker in cases when degraded samples will be used. It also contributes to the development of the subject of species identification in different environments. With the development of DNA barcoding, several analytical methods were developed for the assessment of species discrimination ability. There are no criteria for evaluating the quality of the analytical methods for species discrimination. We have analyzed the effect of barcodes on species identification from the perspective of the machine learning approach (MLA). We tested a reinforcement-learning algorithm to solve the challenge of reconstructing phylogenetic trees, which are used to describe the relationships among a set of sequences. Current tools for phylogenetic- tree reconstruction integrate heuristic approaches to evaluate only a subset of all potential trees, thus they suffer from the known trade-off between accuracy and running time. In our study, we tested the methodology for predicting the maximum likelihood tree. Our preliminary results, based on a machine learning approach algorithm demonstrate that the trained algorithm can accurately and efficiently reconstruct maximum-likelihood trees. This development technique ML could provide rapid, simple, and reliable tools for species confirmation and can be applied to the modelling of species distribution. Materials and methods Sampling DNA In total 18 skull samples were obtained from wild-living specimens of three ungulate species in 2016-2018. DNA isolations were performed using the column-based method and the GeneMatrix Bond DNA Purification Kit (Eurx). The purity and concentration of DNA from the bone material were determined using a NanoDrop 2000c spectrophotometer (Thermo Scientific) (Table 1). Mitochondrial DNA analysis The following primer pair was used for PCR amplification32: Mcb_KPF398: TACCATGAGGACAAATATCATTCTG, Mcb_KPR869:CCTCCTAGTTTGTTAGGGATTGATCG. PCR reactions were performed in a total volume of 20 μL consisting of 20 ng of DNA, 1× DreamTaq Buffer with MgCl2, 0.2 mM dNTP, 0.2 μM of each primer, and 1 U DreamTaq DNA Polymerase (Thermo Scientific). The thermal reaction profile used to amplify the Cytb regions was as follows: initial denaturation at 95 °C for 2 min followed by 35 cycles of denaturation at 95 °C for 30 s, annealing at 58 °C for 30 s, extension of the primer at 72 °C for 30 s, and a final extension of 72 °C for 7 min. PCR products were checked by electrophoresis in a 1.5% agarose gel containing ethidium bromide and a TBE buffer (pH 8.0); the gels were visualized under UV and archived using the GeneSys V.1.3.5.0 software (Syngene). The sequences reported in this paper have been deposited in the GenBank nucleotide sequence database with the accession numbers marked '*' in Tables S1, S2. Sequence analysis At first, the forward and reverse sequences were aligned, and consensus sequences were obtained using Basic Local Alignment Tool software. ClustalW and Mega7.1 software were used to perform multiple sequence alignments60. Substitution patterns and rates were estimated under the Kimura 2-parameter model61. The genetic variability of haplotypes was characterized by the total alignment length (bp), the number of monomorphic sites, the number of polymorphic sites, the number of parsimony informative sites (PIC), the number of haplotypes, and the average G + C content in each region using DnaSP6.10.0162. Species identification To reveal the species of each sample analysed, we performed phylogeny reconstruction using the Bayesian approach. Seven Cytb sequences for Cervus elaphus, two for Dama dama and 9 for Capreolus capreolus were grouped together along with 131 Cytb sequences (Tables S1, S2) of the three species from Genbank, as well as two outgroup sequences (Antidorcas marsupialis, Beatragus hunteri) for comparison. Next, all sequences were aligned with the Muscle algorithm61 and cut to obtain the proper alignment set in Seaview63. The best-fit substitution model was chosen using jModelTest 2.1064. Finally, the tree was constructed with MrBayes 3.2.638 using two, randomly started and independent runs, carried out for 20,000,000 generations of Markov chain steps. A consensus tree was constructed based on the set of trees collected after both runs converged—i.e. when the standard deviation of both runs was much below 0.01. ML tree was also searched using DeepNNPhylogeny, pre-trained neural networks to predict the best models of sequence evolution and the best tree topologies65. All neural networks have been trained with a large number of alignments simulated with the software PolyMoSim, designed to test phylogenetic tree reconstruction and to train machine learning models for phylogenetic reconstruction. We used also ModelTeller, a machine-learning based algorithm, which is based on the Random Forest, for the prediction of the optimal phylogenetic model for branch-length estimation66. The MLAs BLOG and WEKA (machine learning approach) BLOG 2.0 (Barcoding with LOGic; Institute of Systems Analysis and Computer Science, National Research Council, Rome, Italy) and WEKA (The University of Waikato, Hamilton, New Zealand) were applied. BLOG provides a supervised MLA, which selects suitable nucleotide positions and computes the logic formulae for species identification36. The WEKA workbench is used for classification, clustering and selection problems37,67. The four classifiers: Naïve Bayes68, support vector machines (SMO)69, the decision tree C4.5 (J48)70 and the rule-based RIPPER (Jrip)71 were implemented to analyse the DNA sequences. Distance-based analysis (TaxonDNA) The Kimura72 2-parameter model (K2P)distances between all sequence pairs were calculated with TaxonDNA 1.9 (National University of Singapore, Singapore) and applied to compute the mean and the range of the intra- and interspecific distances for the barcode. The relative distribution of the pairwise intra- and inter-specific distances were estimated with the “best match” and “best close match” functions in the TaxonDNA under the Kimura72 2-parameter (K2P) distance model. Tree-based analysis (neighbour-joining) The phylogenetic analysis was carried out in the MEGA11 (Center for Evolutionary Medicine and Informatics, The Biodesign Institute, Tempe, AZ, USA) based on the K2P model with 1000 bootstrap replications and pairwise deletions73. Species discrimination was considered successful only when all the conspecific individuals formed a monophyletic clade. Ethics approval and consent to participate Under Poland law, Institutional Animal Ethics Committee approval was not required for the study of Cervidae. Statement The consent of the bioethical commission is not required for this type of research in Poland. This is due to the fact that the study material was not taken from live animals, but from bones. In addition, the animals were not caught, euthanized or killed. Samples were taken from fallen animals by North West Forest Districts and sent for this study. Conclusions Despite the challenging biological material of bone tissue, the Cytb gene was successfully used to identify individuals of closely related ungulate bone DNA species using PCR analysis, Sanger DNA sequencing and accurate bioinformatics tools such as the Bayesian approach and the machine learning approach. This research will be extended to analyse more sequences of many barcodes for Cervidae species identification using DNA barcode sequences through machine learning approaches. The data obtained will serve used for comparisons with gene bank records. The proposed methodology will be helpful as a routine identification procedure for a variety of tissue sources, even in cases where the samples are degraded. Limitations The efficiency of the applied DNA isolation method varied. The resulting DNA concentration values demonstrated over a 100-fold difference between the lowest and highest concentration. The study revealed that one of the most important moments during the DNA extraction process was the preliminary preparation of the bone material. Identification of bone samples depended on the quality and quantity of DNA present in the sample. The efficiency of the applied DNA isolation method varied. The resulting DNA concentration values demonstrated over a 100-fold difference between the lowest and highest concentration. The study revealed that one of the most important moments during the DNA extraction process was the preliminary preparation of the bone material. Identification of bone samples depended on the quality and quantity of DNA present in the sample. Supplementary Information Supplementary Table S1. Supplementary Table S2. Supplementary Table S3. Supplementary Information The online version contains supplementary material available at 10.1038/s41598-023-35637-z. Author contributions E.F. was involved in conceptualization, and performed the experiments. T.S., D.C.L. data curation. E.F., D.C.L. writing—original draft and writing—review and editing. E.S. review and editing. All authors contributed to the article and approved the submitted version. Funding This research was funded by the University of Szczecin. Street Papieża Jana Pawła II 22A; 70-453 Szczecin; tax identification number: PL851-020-80-05. Data availability All data generated or analyzed during this study are included in this published article. The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable. All the sequences have been deposited in NCBI GenBank and can be found under accession numbers No: MK575604-MK575603. The Cytb sequences are deposited on the corresponding author’s data. Competing interests The authors declare no competing interests. Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Tobe SS Kitchener AC Linacre AMT Reconstructing mammalian phylogenies: A detailed comparison of the cytochrome b and cytochrome oxidase subunit I mitochondrial genes PLoS ONE 2010 5 e14156 10.1371/journal.pone.0014156 21152400 2. Nakaki S Study of animal species (human, dog and cat) identification using a multiplex single-base primer extension reaction in the cytochrome b gene Forensic Sci. Int. 2007 173 97 102 10.1016/j.forsciint.2007.02.010 17382503 3. Wada K Nishibori M Yokohama M The complete nucleotide sequence of mitochondrial genome in the Japanese Sika deer (Cervus nippon), and a phylogenetic analysis between Cervidae and Bovidae Small Rumin. Res. 2007 69 46 54 10.1016/j.smallrumres.2005.12.002 4. Kuwayama R Ozawa T Phylogenetic relationships among European red deer, wapiti, and sika deer inferred from mitochondrial DNA sequences Mol. Phylogenet. Evol. 2000 15 115 123 10.1006/mpev.1999.0731 10764539 5. Tobe SS Linacre AMT DNA typing in wildlife crime: Recent developments in species identification Forensic Sci. Med. Pathol. 2010 6 195 206 10.1007/s12024-010-9168-7 20526699 6. Xavier MA Paternity identification in sugarcane polycrosses by using microsatellite markers Genet. Mol. Res. 2014 13 2268 2277 10.4238/2014.March.31.7 24737475 7. Ramos A Frequency and pattern of heteroplasmy in the complete human mitochondrial genome PLoS ONE 2013 8 e74636 10.1371/journal.pone.0074636 24098342 8. Knapik K Jȩdrzejczak M Dybus A Mitochondrialny gen cytochromu b (MTCYB) Med. Weter. 2006 62 1229 1232 9. Han S Identification of species and sex of Korean Roe Deer (Capreolus pygargus tianschanicus) using SRY and CYTB genes Integr. Biosci. 2007 11 165 168 10.1080/17386357.2007.9647331 10. Holland MM Parsons TJ Mitochondrial DNA sequence analysis—validation and use for forensic casework Forensic Sci. Rev. 1999 11 21 50 26255820 11. Lahiff S Species-specific PCR for the identification of ovine, porcine and chicken species in meat and bone meal (MBM) Mol. Cell Probes 2001 15 27 35 10.1006/mcpr.2000.0336 11284433 12. Ajmal Ali M The changing epitome of species identification—DNA barcoding Saudi J. Biol. Sci 2014 21 204 231 10.1016/j.sjbs.2014.03.003 24955007 13. Cai Y Identification of deer species (Cervidae, Cetartiodactyla) in China using mitochondrial cytochrome c oxidase subunit I (mtDNA COI) Mitochondrial DNA Part A 2016 27 4240 4243 10.3109/19401736.2014.1003919 14. Hoffmann GS Johannesen J Griebeler EM Species cross-amplification, identification and genetic variation of 17 species of deer (Cervidae) with microsatellite and mitochondrial DNA from antlers Mol. Biol. Rep. 2015 42 1059 1067 10.1007/s11033-014-3845-7 25424838 15. Ward RD Hanner R Hebert PDN The campaign to DNA barcode all fishes FISH-BOL. J. Fish. Biol. 2009 74 329 356 10.1111/j.1095-8649.2008.02080.x 20735564 16. Hebert PDN Ratnasingham S de Waard JR Barcoding animal life: Cytochrome c oxidase subunit 1 divergences among closely related species Proc. R. Soc. Lond. B Biol. Sci. 2003 270 S96 S99 10.1098/rsbl.2003.0025 17. Hebert PDN Penton EH Burns JM Janzen DH Hallwachs WT Ten species in one: DNA barcoding reveals cryptic species in the neotropical skipper butterfly Astraptes fulgerator Proc. Natl. Acad. Sci. USA 2004 101 14812 14817 10.1073/pnas.0406166101 15465915 18. Mitchell A DNA barcoding demystified J. Entomol. 2008 47 169 173 19. Rubinoff D Essays: Utility of mitochondrial DNA barcodes in species conservation Conserv. Biol. 2006 20 1026 1033 10.1111/j.1523-1739.2006.00372.x 16922219 20. Gilbert C Ropiquet A Hassanin A Mitochondrial and nuclear phylogenies of Cervidae (Mammalia, Ruminantia): Systematics, morphology, and biogeography Mol. Phylogenet. Evol. 2006 40 101 117 10.1016/j.ympev.2006.02.017 16584894 21. González S Identification of the endangered small red brocket deer (Mazama bororo) using noninvasive genetic techniques (Mammalia; Cervidae) Mol. Ecol. Resour. 2009 9 754 758 10.1111/j.1755-0998.2008.02390.x 21564736 22. Ramón-Laca A Reliable discrimination of 10 ungulate species using high resolution melting analysis of faecal DNA PLoS ONE 2014 9 e92043 10.1371/journal.pone.0092043 24637802 23. Escobedo-Morales LA Mandujano S Eguiarte LE Rodríguez-Rodríguez MA Maldonado JE First phylogenetic analysis of Mesoamerican brocket deer Mazama pandora and Mazama temama (Cetartiodactyla: Cervidae) based on mitochondrial sequences: Implications for Neotropical deer evolution Mamm. Biol. 2016 81 303 313 10.1016/j.mambio.2016.02.003 24. Gutiérrez EE A gene-tree test of the traditional taxonomy of american deer: The importance of voucher specimens, geographic data, and dense sampling Zookeys 2017 697 87 131 10.3897/zookeys.697.15124 25. Bruford MW Bradley DG Luikart G DNA markers reveal the complexity of livestock domestication Nat. Rev. Genet. 2003 4 900 910 10.1038/nrg1203 14634637 26. Parson W Pegoraro K Niederstätter H Föger M Steinlechner M Species identification by means of the cytochrome b gene Int. J. Legal Med. 2000 114 23 28 10.1007/s004140000134 11197623 27. Hsieh H-M Cytochrome b gene for species identification of the conservation animals Forensic Sci. Int. 2001 122 7 18 10.1016/S0379-0738(01)00403-0 11587860 28. Xiao C-T Zhang M-H Fu Y Koh H-S Mitochondrial DNA distinction of northeastern China roe deer, Siberian roe deer, and European roe deer, to clarify the taxonomic status of northeastern China roe deer Biochem. Genet. 2007 45 93 102 10.1007/s10528-006-9052-z 17219081 29. Guha S Goyal SP Kashyap VK Molecular phylogeny of musk deer: A genomic view with mitochondrial 16S rRNA and cytochrome b gene Mol. Phylogenet. Evol. 2007 42 585 597 10.1016/j.ympev.2006.06.020 17158073 30. Yan D Forensic DNA barcoding and bio-response studies of animal horn products used in traditional medicine PLoS ONE 2013 8 e55854 10.1371/journal.pone.0055854 23409064 31. Irwin DM Kocher TD Wilson AC Evolution of the cytochromeb gene of mammals J. Mol. Evol. 1991 32 128 144 10.1007/BF02515385 1901092 32. Verma SK Singh L Novel universal primers establish identity of an enormous number of animal species for forensic application Mol. Ecol. Notes 2002 3 28 31 10.1046/j.1471-8286.2003.00340.x 33. Gupta SK Kumar A Hussain SA Novel primers for sequencing of the complete mitochondrial cytochrome b gene of ungulates using non-invasive and degraded biological samples Conserv. Genet. Resour. 2014 6 499 501 10.1007/s12686-014-0143-6 34. Candan K First assessment on the molecular phylogeny of Anatololacerta (Squamata, Lacertidae) distributed in Southern Anatolia: Insights from mtDNA and nDNA markers Mitochondrial DNA 2014 27 1 8 10.3109/19401736.2014.987238 24438248 35. Bertolazzi P Felici G Weitschek E Learning to classify species with barcodes BMC Bioinform. 2009 10 S7 10.1186/1471-2105-10-S14-S7 36. Weitschek E Van Velzen R Felici G Bertolazzi P BLOG 2.0: A software system for character-based species classification with DNA barcode sequences. What it does, how to use it Mol. Ecol. Resour. 2013 13 5 5 37. Hall M The WEKA data mining software: An update ACM SIGKDD Explor. Newsl 2009 11 10 18 10.1145/1656274.1656278 38. Ronquist F MrBayes 3.2: Efficient Bayesian phylogenetic inference and model choice across a large model space Syst. Biol. 2012 61 539 542 10.1093/sysbio/sys029 22357727 39. Rambaut, A. FigTree v. 1.4.4. http://tree.bio.ed.ac.uk/software/figtree/ (2014). 40. R Development Core Team. R: A Language and Environment for Statistical Computing https://www.r-project.org/ (2018). 41. Panday R Jha DK Thapa N Pokharel BR Aryal NK Forensic wildlife parts and their product identification and individualization using DNA barcoding Open Forensic Sci. J. 2014 7 6 13 10.2174/1874402801407010006 42. Lopez-Oceja A Gamarra D Borragan S Jiménez-Moreno S de Pancorbo MM New cyt b gene universal primer set for forensic analysis Forensic Sci. Int. Genet. 2016 23 159 165 10.1016/j.fsigen.2016.05.001 27206224 43. Jun J Wildlife forensics using mitochondrial DNA sequences: Species identification based on hairs collected in the field and confiscated tanned Felidae leathers Genes Genomics 2011 33 721 726 10.1007/s13258-011-0080-7 44. Tobe SS Linacre AMT A multiplex assay to identify 18 European mammal species from mixtures using the mitochondrial cytochrome b gene Electrophoresis 2008 29 340 347 10.1002/elps.200700706 18080254 45. Spychaj A Mozdziak PE Pospiech E PCR methods in meat species identification as a tool for the verification of regional and traditional meat products Acta Sci. Pol. Technol. Aliment. 2009 8 5 20 46. Wilson-Wilde L Norman J Robertson J Sarre S Georges A Current issues in species identification for forensic science and the validity of using the cytochrome oxidase I (COI) gene Forensic Sci. Med. Pathol. 2010 6 233 241 10.1007/s12024-010-9172-y 20563888 47. Andrejevic M Identification of a broad spectrum of mammalian and avian species using the short fragment of the mitochondrially encoded cytochrome b gene Forensic Sci. Med. Pathol. 2019 15 169 177 10.1007/s12024-019-00096-4 30806911 48. Kang M Genetic analysis of ancient bones of Cervidae animals from archaeological site in Jeju, Korea Integr. Biosci. 2007 11 147 153 10.1080/17386357.2007.9647328 49. Borowski Z Purifying selection, density blocking and unnoticed mitochondrial DNA diversity in the red deer, Cervus elaphus PLoS ONE 2016 11 e0163191 10.1371/journal.pone.0163191 27649313 50. Muangkram Y Wajjwalku W Amano A Sukmak M The novel primers for mammal species identification-based mitochondrial cytochrome b sequence: Implication for reserved wild animals in Thailand and endangered mammal species in Southeast Asia Mitochondrial DNA Part A 2018 29 62 72 10.1080/24701394.2016.1238902 51. Tate ML A new gene mapping resource: interspecies hybrids between Père David’s deer (Elaphurus davidianus) and red deer (Cervus elaphus) Genetics 1995 139 1383 1391 10.1093/genetics/139.3.1383 7768446 52. Slate J A deer (subfamily Cervinae) genetic linkage map and the evolution of ruminant genomes Genetics 2002 160 1587 1597 10.1093/genetics/160.4.1587 11973312 53. Slate J A genome scan for quantitative trait loci in a wild population of red deer (Cervus elaphus) Genetics 2002 162 1863 1873 10.1093/genetics/162.4.1863 12524355 54. Bana NÁ The red deer Cervus elaphus genome CerEla1.0: Sequencing, annotating, genes, and chromosomes Mol. Genet. Genomics 2018 293 665 684 10.1007/s00438-017-1412-3 29294181 55. Brauning, R. et al. Utilization of high throughput genome sequencing technology for large scale single nucleotide polymorphism discovery in red deer and Canadian elk. bioRxiv 027318, (2015). 56. Johnston SE Huisman J Ellis PA Pemberton JM A High-density linkage map reveals sexual dimorphism in recombination landscapes in red deer (Cervus elaphus) G3 Genes Genomes Genet. 2017 7 2859 2870 10.1534/g3.117.044198 57. Baker KH Strong population structure in a species manipulated by humans since the Neolithic: The European fallow deer (Dama dama dama) Heredity 2017 119 16 26 10.1038/hdy.2017.11 28353685 58. Polziehn RO Strobeck C A phylogenetic comparison of red deer and wapiti using mitochondrial DNA Mol. Phylogenet. Evol. 2002 22 342 356 10.1006/mpev.2001.1065 11884159 59. Lorenzini R Garofalo L Qin X Voloshina I Lovari S Global phylogeography of the genus Capreolus (Artiodactyla: Cervidae), a Palaearctic meso-mammal Zool. J. Linn. Soc. 2014 170 209 221 10.1111/zoj.12091 60. Kumar S Stecher G Tamura K MEGA7: Molecular evolutionary genetics analysis version 7.0 for bigger datasets Mol. Biol. Evol. 2016 33 1870 1874 10.1093/molbev/msw054 27004904 61. Edgar RC MUSCLE: Multiple sequence alignment with high accuracy and high throughput Nucleic Acids Res. 2004 32 1792 1797 10.1093/nar/gkh340 15034147 62. Librado P Rozas J DnaSP v5: A software for comprehensive analysis of DNA polymorphism data Bioinformatics 2009 25 1451 1452 10.1093/bioinformatics/btp187 19346325 63. Gouy M Guindon S Gascuel O SeaView Version 4: A multiplatform graphical user interface for sequence alignment and phylogenetic tree building Mol. Biol. Evol. 2010 27 221 224 10.1093/molbev/msp259 19854763 64. Darriba D Taboada GL Doallo R Posada D jModelTest 2: More models, new heuristics and parallel computing Nat. Methods 2012 9 772 772 10.1038/nmeth.2109 22847109 65. Kulikov N Mayer C 2023 Machine learning can be as good as maximum likelihood when reconstructing phylogenetic trees and determining the best evolutionary model on four taxon alignments Dryad Dataset 10.5061/dryad.ksn02v783 66. Abadi S Avram O Rosset S Pupko T Mayrose I ModelTeller: Model selection for optimal phylogenetic reconstruction using machine learning Mol. Biol. Evol. 2020 37 3338 3352 10.1093/molbev/msaa154 32585030 67. Smith TC Frank E Introducing machine learning concepts with WEKA Methods Mol. Biol. 2016 1418 353 378 10.1007/978-1-4939-3578-9_17 27008023 68. Lewis, D. D. Naive (Bayes) at forty: The independence assumption in information retrieval. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) vol. 1398, 4–15 (1998) 69. Suykens JAK Vandewalle J Least squares support vector machine classifiers Neural Process Lett. 1999 9 293 300 10.1023/A:1018628609742 70. Quinlan JR Improved use of continuous attributes in C4.5 J. Artif. Intell. Res. 1996 4 77 90 10.1613/jair.279 71. Waseem S Salman A Muhammad AK Feature subset selection using association rule mining and JRip classifier Int. J. Phys. Sci. 2013 8 885 896 10.5897/IJPS2013.3842 72. Kimura M A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide sequences J. Mol. Evol. 1980 16 2 111 120 10.1007/BF01731581 7463489 73. Tamura K Stecher G Kumar S MEGA11: Molecular evolutionary genetics analysis version 11 Mol. Biol. Evol. 2021 38 3022 3027 10.1093/molbev/msab120 33892491

PMC010xxxxxx/PMC10243494.txt

==== Front J Nematol J Nematol jofnem jofnem Journal of Nematology 0022-300X 2640-396X Sciendo 37288386 jofnem-2023-0024 10.2478/jofnem-2023-0024 Research Paper Rotation of Cotton (Gossypium hirsutum) Cultivars and Fallow on Yield and Rotylenchulus reniformis Soto-Ramos Casiani Wheeler Terry A. ta-wheeler@tamu.edu Shockey Jonathan Monclova-Santana Cecilia Texas A&M AgriLife Extension Service, Lubbock, TX 79403 Texas A&M AgriLife Research, Lubbock, TX 79403 Texas A&M AgriLife Extension Service, Lubbock, TX 79403, currently at United States Department of Agriculture, 1400 Independence Ave SW, Washington, DC 22314 This paper was edited by Horacio Lopez-Nicora. 2 2023 6 6 2023 55 1 2023002429 1 2023 © 2023 Casiani Soto-Ramos et al., published by Sciendo 2023 Casiani Soto-Ramos et al., published by Sciendo https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. Abstract A three-year rotation of cotton (Gossypium hirsutum) cultivars either resistant (R) or susceptible (S) to Rotylenchulus reniformis and fallow (F) was examined for effect on cotton yield and nematode density. In year 1, 2, and 3, the resistant cultivar (DP 2143NR B3XF) yielded 78, 77, and 113% higher than the susceptible cultivar (DP 2044 B3XF). Fallow in year 1 followed by S in year 2 (F1S2) improved yield in year 2 by 24% compared with S1S2, but not as much as R1S2 (41% yield increase over S1S2). One year of fallow followed by R (F1R2) had lower yield in year 2 (11% reduction) than R1R2. The highest yield after three years of these rotations occurred with R1R2R3, followed by R1S2R3 (17% less yield) and F1F2S3 (35% less yield). Rotylenchulus reniformis density in soil averaged 57, 65, and 70% lower (year 1, 2, 3, respectively) in R1R2R3 compared with S1S2S3. In years 1 and 2, LOG10 transformed nematode density (LREN) was lower in F1, and F1F2, than for all other combinations. In year 3, the lowest LREN were associated with R1R2R3, F1S2F3, and F1F2S3. The highest LREN were associated with F1R2S3, F1S2S3, S1S2S3, R1R2S3, and R1S2S3. The combination of higher yield and lower nematode density will be a strong incentive for producers to use the R. reniformis resistant cultivars continuously. Keywords cotton Gossypium hirsutum management reniform nematode resistance Rotylenchulus reniformis ==== Body pmcThe reniform nematode Rotylenchulus reniformis causes substantial (>40%) yield losses in cotton (Robinson, 2007; Dyer et al, 2020). Effective management for many years was limited to crop rotation with nonhosts like sorghum or corn (Davis et al., 2003; Robinson, 2007; Stetina et al., 2007). There are several nematicide options available in cotton that have demonstrated effectiveness against R. reniformis, including fumigation with 1,3-dichloropropene (Koenning et al., 2007), aldicarb (Lawrence et al., 1990), aldicarb plus oxamyl (Lawrence and McLean, 2000), and fluopyram (Dyer et al., 2020). However, for various reasons (lack of effectiveness under dry conditions, cost, difficulty in application, health and environmental concerns), nematicides have not been a satisfactory management option for all cotton producers. Historically, there were no commercial cotton cultivars with resistance to R. reniformis (Robinson, 2007). Recently, R. reniformis resistance has been transferred to commercially available cotton cultivars. In 2020, Phytogen released the first R. reniformis and Meloidogyne incognita resistant cotton varieties (PHY 332 W3FE, plant variety protection [PVP] #202000220; and PHY 443 W3FE, PVP #202000221). In 2021, Deltapine released their first R. reniformis and M. incognita resistant cotton varieties (DP 2141NR B3XF and DP 2143NR B3XF). Rotation of cotton with fallow has not been traditionally recommended for R. reniformis control compared with rotation with nonhost crops (Robinson, 2007). Prior to the availability of reniform nematode resistant cotton varieties, crop rotation for one or two years in a nonhost like sorghum or corn, followed by cotton was the recommended practice. However, the decision of using deficit-limited irrigation on a grain crop (sorghum or corn) versus leaving land fallow is now more difficult in the Southern High Plains of Texas, due to diminished irrigation availability. In deficit-irrigated cotton production as in the Southern High Plains of Texas, the ability to irrigate entire center pivots has declined with the reduced availability of irrigation water (Mitchell-McCallister et al., 2021). Future scenarios project utilizing only one fourth of the area of a center pivot for irrigated cotton, with transition to increased dryland (rainfed) cotton production. However, for fields infested with R. reniformis, it is more likely that part of the land under a center pivot will be left fallow, rather than growing dryland cotton and maintaining higher densities of the R. reniformis. There has been a need previously to incorporate cotton rotation with a nonhost to reduce R. reniformis density, thus allowing maximum cotton yields in the subsequent crop. However, rotation with nonhost plants has not always shown consistent yield improvements in R. reniformis fields. Field fumigation trials were conducted in two R. reniformis infested fields, with one planted continuously to cotton, and the other in a cotton-sorghum rotation (Thames and Heald, 1974). In one year, there was a significant increase in cotton yield with fumigation, where the best fumigation treatment yielded 57% higher than the untreated check in continuous cotton, but no significant yield differences in the field with cotton following sorghum. However, the next year, the reverse was true, with significant yield differences for all chemical treatments when cotton followed sorghum (maximum of 40% yield increase), compared to no significant differences between treatments in the continuous cotton field. Davis et al. (2003) found that in a R. reniformis field, continuous cotton yielded consistently less than cotton following corn or following R. reniformis resistant soybean. Cotton yield only increased in a R. reniformis test after two consecutive years of corn, followed by cotton compared to continuous cotton (Stetina et al., 2007). A corn-cotton-corn-cotton rotation did not significantly increase lint yields. The reduction of R. reniformis density is not always enhanced with planting nonhosts for R. reniformis compared with fallow. Rotylenchulus reniformis density under fallow maintained similar levels as the nonhost plants (Sunn hemp, Rhodes grass, and Pangola grass) when monitored for 342 days (Caswell et al., 1991). Rotylenchulus reniformis density only declined by 52% after 342 days in fallow field plots. Part of the concern over using fallow as opposed to nonhost plants, is that a fallow field would not be irrigated, and R. reniformis can survive dry conditions better than many other plant parasitic nematode species. The second, third, and fourth molts occur quickly under moisture stress for R. reniformis, and the cuticles of the previous stages form superimposed sheaths which are helpful in reducing water loss (Gaur and Perry, 1991). This nematode may also have coiling of their body, which would reduce the surface area exposed, and thus reduce water loss. Womersley and Ching (1989) found that slow drying like what would be found in agricultural soils, particularly in deeper soil, allowed coiled R. reniformis juveniles to survive in greater numbers. Rapid drying and drying down to lower relative humidity (<60%), resulted in higher mortality of R. reniformis. Rotylenchulus reniformis has a wide host range, including many weeds (Lawrence et al., 2008; Molin and Stetina, 2016). Therefore, management of weeds (weed-free fallow), would also be important when utilizing fallow as a management tool to reduce the nematode density for a future cotton crop. Management of R. reniformis with either fallow conditions and/or resistant cotton cultivars are important strategies that should be tested. The objective of this study was to test three-year rotations that included various combinations of fallow, reniform nematode resistant and susceptible cultivars, with regards to late summer nematode density and cotton yield. Materials and Methods A R. reniformis infested field located at the Texas A&M AgriLife Research and Extension Center, Lubbock, Texas, United States, that had no prior history of use of reniform nematode resistant cultivars was used for the experiment. The soil in the field is an Acuff sandy clay loam (sand = 45%, silt = 26%, clay = 29%), pH = 7.7, organic matter = 0.6%, and CEC=13.2. The plots were 48.8 m long, 4-rows wide, on 1 m centers. The R. reniformis resistant cultivar (R) was DP 2143NR B3XF and the susceptible (S) cultivar was DP 2044 B3XF. A third treatment was weed-free fallow (F). The nine different combinations of rotation treatments tested were: R1R2R3 (first letter was for 2020 variety type followed by 1, second letter for 2021 variety type followed by 2, and third letter for 2022 variety type followed by 3), S1S2S3, F1F2S3, F1S2F3, F1R2S3, F1S2S3, R1R2S3, R1S2R3, and R1S2S3. The nine treatments were arranged in a semi-randomized complete block design, with four replications per combination. Each replication was laid across 36 rows (east/west direction, nine treatments x four row plots), and then the four replications were stacked (1.5 m alleys between replications) down the rows (north/south). In 2020 (first year of the trial), the fallow treatments were all placed into rows 1–4, 13–16, 25–28, and rows 33–36. The specific fallow rotation treatment was randomized across the four replications, within those 16 rows. This was done so that the fallow rows were not irrigated in 2020. The cotton was furrow irrigated, typically a week before planting, and then approximately once a month during June, July, and August, unless rainfall negated the need for irrigation. After 2020, fallow treatments (F1F2S3 in 2021 and F1S2F3 in 2022) were irrigated since there were cotton treatments also in the same rows. Cotton was planted on 20 May, 5 June, and 23 May in 2020, 2021, and 2022, respectively. Cotton was harvested on 19, 18, and 16 November in 2020, 2021, and 2022, respectively. Weed control was primarily managed with herbicides. Preplant herbicide treatment was with trifluralin (840 g ai/ha) incorporated, followed by fluometuron applied after planting (preemergence) at 1.12 kg ai/ha. Herbicide applications with glyphosate (630 kg ai.ha) combined with glufosinate (594 g ai/ha) were made to the cotton varieties and fallow areas as needed. S-metolachlor (1.07 kg ai/ha) was also applied prior to the 5th leaf stage to the field in-season (after cotton emergence) to prolong soil herbicide protection. Hand hoeing was also done to handle weed escapes. Composite soil samples were taken for nematode assays in each plot on 25 April and 17 August in 2020, 14 September in 2021, and 8 September in 2022. Sampling was always conducted either after a significant rainfall event or approximately one week after furrow irrigation, when soil moisture was adequate for sampling. Samples consisted of ten subsamples per plot collected with a narrow-bladed shovel (40 cm depth, 15 cm width at top and 8 cm width at the bottom) to a depth of 20 cm, close to the taproot. The top 6 cm of soil was discarded and then soil from 6–20 cm depth, including some roots was removed. The soil was mixed in a bucket and then a subsample of 750 cm3 soil was placed in a plastic bag. The soil samples were refrigerated for <2 weeks before being assayed for plant parasitic nematodes. A pie-pan assay with 200 cm3 soil + root fragments was used to extract R. reniformis over 48 hours (Thistlethwayte, 1970). The circular pie-pans are made of glass with 18 cm diameter at the base, 22 cm at the top, and 3 cm tall. Three washers were placed in the base of the pie-pan and wire mesh (0.64 cm diameter) laid on top. Two pieces of Kleenex paper tissues (2-ply) were laid on top of the mesh and then the soil sample was placed on the paper tissues. Tap water (250 ml) was gently added to the pie-pan without disturbing the soil, and then the wet paper tissues were arranged around the soil to keep the soil from floating into the water. A cover was placed over the pie-pan to eliminate evaporation. The extracted nematodes were assessed by concentrating the pie-pan water to 100 ml and then counting a 5 ml aliquot. This assay is effective only on live and mobile nematodes. Cotton was harvested with a modified four-row cotton stripper (John Deere 484), which has a fabricated cage within the body of the harvester that is set on load cells to weigh the harvested cotton. Cotton strippers remove lint, seed, burrs, and other plant debris. Samples were taken from the harvested cotton and ginned to obtain the percentage of the harvest weight that was lint, and lint yield (kg/ha) was calculated. The lint from ginned samples was sent to the Fiber and BioPolymer Research Institute at Texas Tech University, for High Volume Instrument (HVI) testing. The properties measured included micronaire (a relative measure of fiber mass per unit length determined by air permeability), fiber length, strength (force require to break a fiber sample), elongation (the amount that a fiber sample will stretch prior to breakage), uniformity (the ratio of mean length to upper half mean length), Rd (degree of reflectance, indicates how light or dark the fiber sample is), +b (yellowness, the measure of color pigmentation), leaf index (visual estimation of the amount of cotton plant leaf material is on the lint), and color grade (a function of Rd and +b of the fiber sample, based on the Nickerson-Hunter cotton colorimeter diagram) (Cotton Division, 1993). These fiber property values can then be used to calculate loan value (National Cotton Council, 2022) each year for the harvested cotton. Rotylenchulus reniformis densities, which were taken in August or September (Ren), were analyzed using a LOG10(Ren+1) = LRen transformation. A mixed model analysis was performed on the year 1 and 2 treatments, since there was an unequal number of plots associated with each unique combination, using PROC GLIMMIX (SAS version 9.4, Cary, NC, U.S.A). The least square means were compared at P = 0.05, using the DIFF options to compare all pairwise least mean square differences. Since the 2022 data set had four replications for all nine treatments, that data set was analyzed using analysis of variance (PROC GLM, SAS version 9.4), and mean comparisons were made with the Waller-Duncan k-ratio t-test at P = 0.05. Similar analyses were done with lint yield in each of the three years. In year 1 there were three treatments (S1 [four plots], R1 [16 plots], and F1 [16 plots]). In year 2 there were six treatment combinations (S1S2 [4 plots], R1R2 [8 plots], F1F2 [4 plots], F1S2 [8 plots], F1R2 [4 plots], and R1S2 [8 plots]). In year 3 there were nine unique combinations (presented previously), each with four replications. Results The initial plot soil samples taken in April of 2020, before planting, indicated that R. reniformis density averaged 926/100 cm3 soil and was not different between treatments. Rain from the previous year (September 2019 through April of 2020) was high (30.6 cm) for this area, and probably contributed to the excellent nematode recovery. Following year 1 (2020), the resistant cotton cultivar (DP 2143NR B3XF) yielded significantly higher (622 kg lint/ha) than the susceptible cotton cultivar (DP 2044 B3XF, 348 kg lint/ha). The loan values did not differ (P = 0.834) between the two cultivars. Rotylenchulus reniformis density in August was higher in the susceptible cotton plots (1,025 R. reniformis/100 cm3 soil, Lren = 2.95) than the resistant cotton (439 R. reniformis/100 cm3 soil, Lren = 2.63), and both cotton cultivars had significantly higher Lren counts than the fallow treatment (288 R. reniformis/100 cm3 soil, Lren = 2.40). In year 2 (2021), lint yield was highest with two consecutive years of resistant cotton (916 kg lint/ha), and significantly higher than resistant cotton following fallow (815 kg lint/ha), or any of the year 2-susceptible cotton treatment combinations (Table 1). The lowest lint yield was associated with two years of susceptible cotton (516 kg lint/ha). The loan value did not differ between the two varieties (P = 0.864), or between any treatments (P = 0.847). Table 1. Effect of cultivar and fallow rotation treatments in year 2 and year 3 on cotton lint yield and Rotylenchulus reniformis (REN) density. Treatment Lint yield (kg/ha) RENb LREN 2021 R1R2 916 a 295 2.42 c S1S2 516 d 835 2.89 a F1F2 -------- 60 1.73 d R1S2 730 b 745 2.78 ab F1R2 815 b 385 2.52 bc F1S2 641 c 750 2.77 ab 2022 R1R2R3 1,271 a 150 2.17 d S1S2S3 597 d 505 2.64 ab F1F2S3 823 c 295 2.33 cd R1S2R3 1,051 b 265 2.36 bcd R1S2S3 599 d 385 2.53 abc F1R2S3 736 cd 655 2.78 a F1S2F3 ------- 150 2.12 d F1S2S3 628 cd 570 2.75 a R1R2S3 724 cd 395 2.54 abc a R indicates the R. reniformis resistant cotton cultivar DP 2143NR B3XF was grown, S indicates the susceptible cotton cultivar DP 2044 B3XF was grown, and F indicates weed-free fallow. The crop in the first year is the first letter followed by 1, the second year is the second letter followed by 2, and third year is the third letter followed by 3. b REN is R. reniformis soil population per 100 cm3 soil taken in September. c LREN is the transformation of the R. reniformis density (LOG10(REN+1)). Rotylenchulus reniformis densities in September of 2021 were highest for all the plots where susceptible cotton was grown in 2021, and lowest for plots with two years of fallow (Table 1). Transformed R. reniformis density was significantly higher in year 2 for the S1S2, R1S2, and F1S2 treatments compared with the R1R2, and F1F2 treatments. The F1R2 treatment was intermediate and not different from R1S2 and F1S2. It was clear from year 2, that regardless of history from year 1 (weed-free fallow, resistant or susceptible cotton cultivar), growing a susceptible cultivar resulted in substantial R. reniformis buildup by September. However, using a weed-free fallow system prior to growing a susceptible or resistant cotton cultivar did cause a yield penalty compared to using a resistant cultivar in year 1 (R1R2 yielded more than F1R2; and R1S2 yielded more than F1S2). The weed-free fallow in year 1 did not receive irrigation and there was only 30.5 cm of rain from May of 2020 through April of 2021. In year 3 (2022), lint yield was higher in the three-year resistant cotton combination (R1R2R3) than all other combinations (Table 1). Lint yield was lower in all combinations that included a year 3 susceptible cotton cultivar than in the two combinations that had the resistant cotton cultivar in year 3 (R1R2R3 and R1S2R3). Loan value did not differ between the two varieties (P = 0.316) or between any treatments (P = 0.395). Rotylenchulus reniformis transformed densities in September 2022 were significantly (P < 0.05) lower in both the resistant (2.27) and susceptible (2.59) cultivars than in 2020 (2.63 and 2.95, respectively) or 2021(2.45 and 2.80, respectively). This may be due to the extremely dry soil going into planting time (5.6 cm rain from September 2021 through April 2022), and the inability to maintain adequate soil moisture, even with irrigation during the hotter than normal summer months of 2022. The lowest Lren were associated with the R1R2R3, F1S2F3, F1F2S3, and R1S2R3. The highest Lren were associated with F1R2S3, F1S2S3, S1S2S3, R1S2S3, and R1R2S3. Those plots planted with susceptible cotton in year 3, in most cases, had the highest transformed R. reniformis densities. The exception was that after two years of fallow, the buildup on a susceptible cultivar was less than that found with most other year 3 susceptible cultivar cotton plots. Discussion Currently, there are two companies that have R. reniformis resistant cultivars available, Phytogen (four commercially available cultivars as of 2022) and Deltapine (two commercially available cultivars as of 2022). The source of resistance in both company's cultivars are likely to be the same (Gossypium barbadense GB713 [Robinson et al., 2004]), though the only confirmation is in the PVP certificates available for PHY 332 W3FE and PHY 443 W3FE. The use of this excellent source of R. reniformis resistance (McCarty et al., 2013) was facilitated by the development of SSR markers associated with the resistance on chromosome 21 (Gutierrez et al., 2011). Cotton breeding lines with reniform nematode resistance that had been developed from crosses with GB713 yielded higher than susceptible commercial cotton varieties in R. reniformis infested field trials (Koebernick et al., 2021). The commercially available cultivars have not only obtained resistance to R. reniformis, but also to M. incognita (2-gene resistance), and in addition contain six transgenic traits (three Bt genes and three herbicide tolerant genes), making a total of at least nine genes that must be selected in addition to acceptable yield and fiber quality traits. Identification of markers for nematode resistance is essential for the development of commercial cultivars, particularly in a crop where transgenic traits are common. The two M. incognita resistance genes also have good SSR markers that can be utilized in commercial variety development (Gutiérrez et al., 2010; Jenkins et al., 2012). The results in this study were as anticipated: fallow soil and the use of R. reniformis resistant cultivar DP 2143NR B3XF reduced R. reniformis density relative to a susceptible cultivar (DP 2044 B3XF). The yield advantage of using a resistant cultivar continuously was 78%, 77%, and 113% higher than the susceptible cultivar in year 1, 2, and 3, respectively. The advantage of using consecutive years of resistant cultivars was seen both in the reduction of R. reniformis density and improvement in lint yield. The economic value of using the reniform nematode resistant cultivar is close to the value of the yield difference, since the loan values for both cultivars were similar, and all other management costs would also have been similar. The cost of the seed in 2023 in this region was estimated at $1.47 per 1000 seed for DP 2044 B3XF and $1.87 per 1000 seed for DP 2143NR B3XF (Plains Cotton Growers Inc., 2023). For a planting density of 96,971 seed/ha, it would cost $38.79 more per hectare for the reniform nematode resistant variety than a susceptible variety, which is a minor expense considering the yield benefits. One surprise in year 2 was that one year of fallow, followed by a resistant or susceptible cultivar, resulted in less yield than using a resistant cultivar in year 1, followed by a resistant or susceptible cultivar in year 2. That would suggest that something about the fallow treatment caused lower than expected yield in the following cotton crop. It is possible that there was higher than recovered R. reniformis populations in the fallow treatment, either because they were deeper in the soil profile than where sampling occurred, or because the existing nematode population was still immobilized after the 48-hour water extraction due to being in an anhydrobiotic state from dry soil (Womersley and Ching, 1989). The summer in 2020 was relatively dry (16 cm of rain from May – August). Rotylenchulus reniformis density in continuous cotton was found to be higher in the upper profile of field soil, compared to that of a corn-cotton rotation, where the nematode density peaked at the lower soil profile (Lee et al., 2015). It is also possible, though less likely, that the soil was drier going into the 2021 growing season in the rows that had been fallow in 2020, compared to the rows that had cotton in 2020 and were irrigated in-season. However, May 2021 was wet, which was why the plots were planted on 5 June in 2021, and so it is unlikely that initial differences in underlying moisture explained the yield differences in 2021. The use of fallow ground rather than a nonhost rotation crop like corn or sorghum is only desirable under the scenario where water is limited and insufficient to grow an alternative crop to cotton. Corn or sorghum allow the outcome of farm income as well as the opportunity to reduce R. reniformis density (Davis et al., 2003; Stetina et al., 2007). However, in the semi-arid region where this work was conducted, producers already have insufficient irrigation capacity for their center pivot systems and may need to reduce the irrigated crop area to ¼ or ½ of the circle (Mitchell-McCallister et al., 2021). In this scenario, it would be more beneficial to use fallow as a tactic to reduce R. reniformis density rather than dryland cotton production, which could maintain nematode populations at damaging levels. The results of this project support using continuous production of R. reniformis resistant cotton cultivars in infested fields. The profitability would be much higher than either using susceptible cultivars or using a weed-free fallow rotation with cotton. While the reduction in R. reniformis density was significant with the resistant cultivar, there can be sufficient population remaining to be damaging to susceptible cotton in the following year. It is likely that producers will want to repeatedly use the R. reniformis resistant cultivars. However, there is concern that repeated planting of a single source of resistance (presumably chromosome 21 Renbarb1 and/or Renbarb2 resistance QTL (Wubben et al. 2017; Gaudin and Wubber, 2021)) will ultimately result in the breakdown of that resistant source. The rotation of resistant and susceptible cultivars, or use of fallow or nonhosts of R. reniformis may delay that possibility. Acknowledgements We appreciate Bayer CropSciences for providing seed for the experiments, and the support from Texas A&M AgriLife Extension Service. ==== Refs Literature Cited Caswell E. P. DeFrank K. J. Apt W. J. Tang C. S. 1991 Influence of nonhost plants on population decline of Rotylenchulus reniformis Journal of Nematology 23 91 98 19283098 Cotton Division 1993 The classification of cotton Agricultural Marketing Service, USDA Agricultural Handbook 566 32 Davis R. F. Koenning S. R. Kemerait R. C. Cummings T. D. Shurley W. D. 2003 Rotylenchulus reniformis management in cotton with crop rotation Journal of Nematology 35 58 64 19265975 Dyer D. R. Groover W. Lawrence K. S. 2020 Yield loss of cotton cultivars due to Rotylenchulus reniformis and the added benefit of a nematicide Plant Health Progress 21 113 118 Gaudin A. G. Wubben M. J. 2021 Genotypic and phenotypic evaluation of wild cotton accessions previously identified as resistant to root-knot (Meloidogyne incognita) or reniform nematode (Rotylenchulus reniformis) Euphytica 217 207 Gaur H. S. Perry R. N. 1991 The role of the moulted cuticles in the desiccation survival of adults of Rotylenchulus reniformis Revue de Nematologie 14 491 496 Gutiérrez O. A. Jenkins J. N. McCarty J. C. Wubben M. J. Hayes R. W. Callahan F. E. 2010 SSR markers closely associated with genes for resistance to root-knot nematode on chromosomes 11 and 14 of Upland cotton Theoretical Applied Genetics 121 1323 1337 20607210 Gutiérrez O. A. Robinson A. F. Jenkins J. N. McCarty J. C. Wubben M. J. Callahan F. E. Nichols R. L. 2011 Identification of QTL regions and SSR markers associated with resistance to reniform nematode in Gossypium barbadense L. accessions GB713 Theoretical and Applied Genetics 122 271 280 20845024 Jenkins J. N. McCarty J. C. Wubben M. J. Hayes R. Gutiérrez O. A. Callahan F. Deng D. 2012 SSR markers for marker assisted selection of root-knot nematode (Meloidogyne incognita) resistant plants in cotton (Gossypium hirsutum L.) Euphytica 183 49 54 Koebernick J. Kaplan G. Lawrence K. Patel J. Brown S. Sikkens R. 2021 Response to nematicide by cotton genotypes varying in reniform nematode resistance Crop Science 61 929 935 Koenning S. R. Morrison D. E. Edmisten K. L. 2007 Relative efficacy of selected nematicides for management of Rotylenchulus reniformis in cotton Nematropica 37 227 235 Lawrence G. W. McLean K. S. 2000 Effect of foliar applications of oxamyl with aldicarb for the management of Rotylenchulus reniformis in cotton Supplement to the Journal of Nematology 32 542 549 Lawrence G. W. McLean K. S. Batson W. E. Miller D. Borbon J. C. 1990 Response of Rotylenchulus reniformis to nematicide applications on cotton Supplement to the Journal of Nematology 22 707 711 Lawrence K. S. Price A. J. Lawrence G. W. Jones J. R. Akridge J. R. 2008 Weed hosts for Rotylenchulus reniformis in cotton fields rotated with corn in the southeast of the United States Nematropica 38 13 22 Lee H. K. Lawrence G. W. DuBien J. L. Lawrence K. S. 2015 Seasonal variation and cotton-corn rotation in the spatial distribution of Rotylenchulus reniformis in Mississippi cotton soils Nematropica 45 72 81 McCarty J. C. Jenkins J. N. Wubben M. J. Gutierrez O. A. Hayes R. W. Callahan F. E. Deng D. 2013 Registration of the germplasm lines of cotton derived from Gossypium barbadense L. accession GB713 with resistance to the reniform nematode Journal of Plant Registrations 7 220 223 Mitchell-McCallister D. McCullough R. Johnson P. Williams R. B. 2021 An economic analysis on the transition to dryland production in deficit-irrigated cropping systems of the Texas High Plains Frontiers in Sustainable Food Systems 5 531601 Molin W. T. Stetina S. R. 2016 Weed hosts and relative weed and cover crop susceptibility to Rotylenchulus reniformis in the Mississippi delta Nematropica 46 121 131 National Cotton Council 2023 CCC loan premium and discounted schedule: Upland cotton http://cotton.org/econ/govprograms/ccloan/ccc-upland-discounts.cfm (checked on March 27, 2023). Plains Cotton Growers Inc. 2023 Upland cotton seed cost comparison worksheet – 2023 http://www.plainscotton.org (checked on March 27, 2023). Robinson A. F. 2007 Reniform in U.S. cotton: when, where, why, and some remedies Annual Review of Phytopathology 45 263 288 Robinson A. F. Bridges A. C. Percival A. E. Edward A. 2004 New sources of resistance to the reniform (Rotylenchulus reniformis) and root-knot (Meloidogyne incognita) nematode in upland (Gossypium hirsutum L.) and Sea Island (G. barbadense L.) cotton Journal of Cotton Science 8 191 197 Stetina S. R. Young L. D. Pettigrew W. T. Bruns H. A. 2007 Effect of corn-cotton rotations on reniform nematode populations and crop yield Nematropica 37 237 248 Thames W. H. Heald C. M. 1974 Chemical and cultural control of Rotylenchulus reniformis on cotton Plant Disease Reporter 58 337 341 Thistlethwayte B. 1970 Reproduction of Pratylenchus penetrans (Nematoda:Tylenchida) Journal of Nematology 2 101 105 19322280 Womersley C. Ching C. 1989 Natural dehydration regimes as a prerequisite for the successful induction of anhydrobiosis in the nematode Rotylenchulus reniformis Journal of Experimental Biology 143 359 372 2732663 Wubben M. J. McCarty J. C. Jr. Jenkins J. N. Callahan F. E. Deng D. 2017 Individual and combined contributions of the Renbarb1, Renbarb2 and Renbarb3 quantitative trait loci to reniform nematode (Rotylenchulus reniformis Linford & Oliveira) resistance in upland cotton (Gossypium hirsutum L.) Euphytica 213 1 9

PMC010xxxxxx/PMC10258564.txt

==== Front J Nematol J Nematol jofnem jofnem Journal of Nematology 0022-300X 2640-396X Sciendo 37313350 jofnem-2023-0027 10.2478/jofnem-2023-0027 Research Paper First Reports and Morphological and Molecular Characterization of Pratylenchus delattrei and Quinisulcius capitatus Associated with Chickpea in Ethiopia Kefelegn Habtamu habtamukefelegn@gmail.com Meressa Beira Hailu Yon Sunheng Couvreur Marjolein Wesemael Wim M. L. Teklu Misghina G. Bert Wim Wim.Bert@ugent.be Nematology Research Unit, Department of Biology, Ghent University, Campus Ledeganck, Ledeganckstraat 35, B-9000 Ghent, Belgium College of Agriculture and Veterinary Medicine, Jimma University, P.O. Box, 307, Jimma, Ethiopia Flanders Research Institute for Agriculture, Fisheries and Food (ILVO), Burg Van Gansberghelaan 96, B-9820 Merelbeke, Belgium Laboratory for Agrozoology, Department of Plants and Crops, Ghent University, Coupure links 653, B-9000 Ghent, Belgium Plant Research, Plant Sciences Group, Wageningen University and Research Centre, P.O. Box 16, 6700 AA Wageningen, The Netherlands. This paper was edited by Guiping Yan. 2 2023 11 6 2023 55 1 2023002727 1 2023 © 2023 Habtamu Kefelegn et al., published by Sciendo 2023 Habtamu Kefelegn et al., published by Sciendo https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. Abstract Chickpea (Cicer arietinum L.) is classed among the most important leguminous crops of high economic value in Ethiopia. Two plant-parasitic nematode species, Pratylenchus delattrei and Quinisulcius capitatus, were recovered from chickpea-growing areas in Ethiopia and characterized using molecular and morphological data, including the first scanning electron microscopy data for P. delattrei. New sequences of D2-D3 of 28S, ITS rDNA and mtDNA COI genes have been obtained from these species, providing the first COI sequences for P. delattrei and Q. capitatus, with both species being found for the first time on chickpea in Ethiopia. Furthermore, Pratylenchus delattrei was recovered in Ethiopia for the first time. The information obtained about these nematodes will be crucial to developing effective nematode management plans for future chickpea production. Keywords D2-D3 of 28S ITS COI Cicer arietinum Ethiopia Lesion nematode Morphometrics Morphology Phylogeny Plant-parasitic nematodes SEM Stunt nematode ==== Body pmcChickpea (Cicer arietinum L.) is classed second among the most important leguminous grain crops after the common bean, and is grown throughout tropical, subtropical, and temperate regions (Singh et al., 2008; FAOSTAT, 2020). Ethiopia is the largest producer of chickpeas in Africa, contributing 60% of the continent's total production and ranking sixth internationally (Shiferaw et al., 2007; FAOSTAT, 2020; Fikre et al., 2020). Chickpea is grown in Ethiopia for both domestic consumption and export purposes. It is also used to restore soil fertility as part of a crop rotation with wheat and teff (Dadi et al., 2005; Shiferaw et al., 2007; Fikre et al., 2020). In Ethiopia, growers of chickpea experience different diseases and insect pests in their fields for which management methods are being implemented, and although plant-parasitic nematodes also represent an important chickpea pest, their importance is usually neglected due to local inabilities to recognize relevant symptoms and/or in identifying the associated species (Castillo et al., 2008; Abebe et al., 2015; Sikora et al., 2018). The root-lesion nematodes (RLN), Pratylenchus spp., are ranked as the third most damaging group of plant-parasitic nematodes in terms of economic loss to agricultural production after root-knot and cyst nematodes (Castillo and Vovlas, 2007; Jones et al., 2013). Pratylenchus is the most important genus that infects chickpea roots globally and reduces crop yields (Di Vito et al., 1992; Thompson et al., 2010; Reen et al., 2019; Behmand et al., 2022; Rostad et al., 2022), and various Pratylenchus species from chickpea roots and rhizospheres have been reported from countries in Asia, Africa, Europe, North America, South America, and Australia (Castillo et al., 2008; Sikora et al., 2018; Zwart et al., 2019). According to studies by Hollaway et al. (2000) and Behmand et al. (2018), in different parts of Turkey where chickpea is grown, chickpea crops are generally considered as being more susceptible to P. neglectus, P. penetrans, and P. thornei attack than field pea, fava bean, and lupin bean crops, but less vulnerable than wheat crops. In Australia, P. thornei and P. neglectus are known to cause substantial damage to chickpea production (Riley and Kelly, 2002; Hollaway et al., 2008; Thompson et al., 2010). Likewise, P. thornei has been reported to cause severe crop losses in Syria, Morocco, Tunisia, Algeria, India, and Spain (Di Vito et al., 1992; Di Vito et al., 1994; Castillo et al., 1996; Ali and Sharma, 2003). The stunt nematode, Quinisulcius capitatus (Allen, 1955) Siddiqi, 1971 (= Tylenchorhynchus capitatus Allen, 1955) is a polyphagous ectoparasite with a wide host range, common in leguminous crops (Greco et al., 1992), field peas in the USA (Upadhaya et al., 2018), and is commonly found parasitizing chickpea fields in Tunisia, Morocco, and Turkey (Di Vito et al., 1994; Ali and Sharma, 2003; Catillo et al., 2008). Quinisulcius species are also widely distributed throughout tomato, pepper, cabbage, and potato crops in many countries worldwide (Bafokuzara, 1996; Baimey et al., 2009; Geraert, 2011; Hussain et al., 2019). The correct identification of nematodes using the link between DNA sequences and morphological characters is crucial in avoiding species misidentification (Janssen et al., 2017a, 2017b), and therefore for the implementation of effective pest management strategies and control measures (Munawar et al., 2021). Nevertheless, in sub-Saharan Africa (SSA), where facilities for morphological and molecular characterizations are scarce, nematode identification has hitherto been limited to genus level (Powers et al., 2011; Coye et al., 2018). For example, in Ethiopia, despite the number of described species of Pratylenchus (Janssen et al., 2017b; Singh et al., 2018; Nguyen et al., 2019; Handoo et al., 2021) and Quinisulcius (Geraert, 2011; Hussain et al., 2019), only P. goodeyi from enset (Peregrine and Bridge, 1992) and P. zeae, P. brachyurus, and P. coffeae from maize have been identified to date (Abebe et al., 2015). This current study reports for the first time the presence of P. delattrei in Ethiopia, and in addition, it provides the first report of P. delattrei and Q. capitatus associated with chickpea. This study also characterizes these two species based on morphological features obtained from light microscope (LM) and scanning electron microscope (SEM), molecular information of ITS, 28S of rDNA and COI of mtDNA. Overall, the study provides a better understanding of nematodes as a potential concern in chickpea production in the country. Materials and Methods Sample collection and nematode extraction Soil and root samples were collected from chickpea growing areas in Minjar, Adea’a, and Mesekan districts during the 2021 main growing season, located in central and southern parts of Ethiopia. Details regarding sample locality, altitude, GPS coordinators, and GenBank accession numbers are summarized in Table 1. From each sampling locality, 20 soil cores were taken in a zig-zag pattern from within the top 30 cm using a 3 cm diameter tube from the chickpea rhizosphere, mixed to obtain a 500 g soil sample. For each sample, 80 chickpea roots were collected and put in labelled plastic bags. Subsequently, both soil and root samples were taken to the Plant Disease Diagnostics laboratory at Jimma University and stored at 4°C until nematode extraction (Barker et al., 1969). The nematodes were extracted from aliquots of 100 ml of soil and 10 g of roots by the modified Baermann tray method described by Hooper et al. (2005). Table 1. Pratylenchus delattrei and Quinisulcius capitatus recovered from chickpea roots and soil, districts, sampling locality, laboratory codes, altitude, GPS coordinators and GenBank accession numbers. Districts Sampling locality Codes Species Altitude (m) Longitude(°) Latitude(°) GenBank accession number 28S–rDNA ITS–rDNA COI gene Minjar Kitecha Mki-5 P. delattrei 1600-1800 8°52′58.16″N 39°29′46.01″E OP646170 - - Mki-8 OP646169 - - Mki-12 - - OP730534 Adea’a Gollodhertu AG-1 P. delattrei 1800-1900 8°38′55.54″N 38°55′0.94″E - OP646171 - AG-2 OP646168 OP646172 - AG-3 OP646167 - OP730535 Mesekan Jolle-2 JO2-3 Q. capitatus 1900-1950 8°11′49.79″N 38°27′51.24″E OP626319 OP646173 OP627909 JO2-5 OP626320 OP646174 OP627910 JO2-7 OP626321 OP646175 OP627911 Morphological characterization Morphological and morphometric data were recorded from both temporary and permanent slides. In order to link molecular data with morphological vouchers of individual nematodes, live nematodes were heat relaxed by quickly passing them over a flame and examined, photographed, and measured using an Olympus BX51 DIC Microscope (Olympus Optical, Tokyo, Japan), equipped with an HD Ultra camera. Subsequently, each specimen was recovered from the temporary slide for genomic DNA extraction. For permanent slides, the nematode suspensions were concentrated in a drop of water in an embryo glass dish, with a few drops of fixative (4% formalin, 1% glycerol (in water) in it. The nematodes were immediately heated in a microwave (700 watts) for about 4 sec and left at room temperature for 1 h at 4°C for 24 h. This was followed by gradually transferring to anhydrous glycerin, ready to be mounted on glass slides as described by Seinhorst (1959). Specimens for scanning electron microscopy (SEM) were fixed in Trump's fixative, washed in 0.1 M-phosphate buffer (pH = 7.5), dehydrated in a graded series of ethanol solutions, critical point dried with liquid CO2 and mounted on stubs with carbon tabs (double conductive tapes), coated with 25 nm gold, and photographed with a JSM-840 EM (JEOL) at 12 kV (Singh et al., 2021). Molecular characterization After making morphological vouchers, nematodes were recovered from temporary slides, washed with distilled water, cut into 2–3 pieces, and transferred to a PCR tube containing 20 μL of worm lysis buffer (WLB) (50 mM KCl;10 mM Tris pH 8.3; 2.5 mM MgCl2; 0.45% NP-40 (Tergitol Sigma); 0.45% Tween-20). Then, the samples were incubated at −20°C for 10 min, followed by adding 1 μL proteinase K (1.2 mg/ml) and incubation for 1 h at 65°C and 10 min at 95°C and centrifugation for 1 min at 14,000 rpm. Finally, the samples were stored at −20°C until used for the PCR, as previously described by Singh et al. (2019) and Nguyen et al. (2019). A DNA template of 3 μL was transferred to an Eppendorf tube containing 23.5 μL master mix containing 10 μL of PCR water, 12.5 μL Dream tag, and 0.5 μL of each primer (Derycke et al., 2010) and PCR amplification was performed using a Bio-Rad T100™ thermocycler. PCR amplifications of the D2-D3 region of 28S-rDNA were performed using the forward primer D2A (5′-ACA AGT ACC GTG AGG GAA AGT TG-3′), and reverse primer D3B (5′-TCG GAA GGA ACC AGC TAC TA-3′) (Subbotin et al., 2006). For ITS rDNA, the forward primer Vrain2F (5′-CTT TGT ACA CAC CGC CCG TCG CT-3′), and reverse primer Vrain2R (5′-TTT CAC TCG CCG TTA CTA AGG GAA TC-3′), were used following the protocol of Vrain et al. (1992) with the touch-down thermal profiles described by Singh et al. (2019). For the amplification of the cytochrome oxidase subunit 1 (COI) gene of mitochondrial DNA, the primer JB3 (5′-TTT TTT GGG CAT CCT GAA GTC TAT-3′) and JB4.5 (5′-CCT ATT CTT AAA ACA TAA TGA AAA TG-3′) and the primer JB3Prat (5′-TTT TTT GGG CAT CCT GAA GTC TAT-3′) and JB4Prat (5′-CCT ATT CTT AAA ACA TAA TGA AAA TG-3′) were used following the protocol of Bowles et al. (1992) with the thermal profile described in the study of Singh et al. (2019). All the PCR products were checked by gel electrophoresis stained with GelRed (Biotium) and visualized under UV light illumination. The successful PCR reactions were purified and sent to Macrogen (https://dna.macrogen.com, Europe) for sequencing. Consensus sequences were assembled in forward and reverse directions using Geneious 2022.1 (Biomatters; http://www.geneious.com) and deposited in the NCBI GenBank (Table 1). Phylogenetic analysis Resulting sequences were compared with other relevant sequences available in the GenBank. Multiple alignments of the different DNA sequences were made using MUSCLE with default parameters, followed by manual trimming of the poorly aligned ends using Geneious 2022.1. Phylogenetic trees were created by using MrBayes 3.2.6, adding Geneious with the GTR + I + G model. The Markov chains for generating phylogenetic trees were set at 1 × 106 generations, four runs, 20% burn-in and sub-sampling frequency of 500 generations (Huelsenbeck and Ronquist, 2001). Results Pratylenchus delattrei Luc, 1958 (Fig. 1). Figure 1: Light microscopy and scanning electron microscopy images of Pratylenchus delattrei. A–C, E–G: Anterior part of the body showing lip and neck region; D: En face view; H: Whole female's body; I–L: Vulva region (L, ventral view); M,N: Lateral fields at mid-body; O–U: Tail region. Measurements See Table 2. Table 2. Comparison of morphometrics of the Ethiopian Pratylenchus delattrei from chickpea in Ethiopia with the original description from Madagascar (Luc, 1958), and two other P. delattrei populations from Cape Verde and Iran. All measurements are in μm and in the form: mean ± s.d. (range). Character P. delattrei from chickpea in Ethiopia (present study) P. delattrei from cotton in Madagascar (Luc, 1958) P. delattrei from tomato in Cape Verde (Flis et al., 2018) P. delattrei from vegetables in Iran (Majd Taheri et al., 2013) Hormozgan 1 Hormozgan 2 n 10 13 20 7 12 L 475 ± 47.1 (410 – 560) 390 – 470 532 ± 33 (498 – 586) 543 ± 55 (467 – 616) 508 ± 49.2 (434 – 576) a 25.2 ± 2.2 (22.8 – 29) 20.4 – 25.8 26.6 ± 2.2 (22.1 – 31.3) 23.8 ± 2.1 (21.2 – 26.9) 22.6 ± 1.0 (21.1 – 25) b 5.6 ± 0.6 (4.8 – 6.6) 3.7 – 4.8 6.6 ± 0.5 (6.1 – 7.7) 6.1 ± 0.7 (5.0 – 7.2) 5.9 ± 0.6 (5.2 – 6.9) b’ 4.7 ± 0.5 (4.1 – 5.6) – 4.5 ± 0.4 (4.0 – 5.3) 4.3 ± 0.2 (4.0 – 4.6) 4.1 ± 0.4 (3.6 – 4.9) c 14.6 ± 1.9 (13.7 – 20.3) 18 – 22.3 21.9 ± 2.1 (18.5 – 25.1) 20 ± 2 (18.1 – 23.1) 19.7 ± 2.6 (16.7 – 24.1) c’ 2.3 ± 0.2 (2.0 – 2.8) – 2.2 ± 0.2 (1.9 – 2.8) 2.2 ± 0.3 (1.9 – 2.6) 1.9 ± 0.2 (1.6 – 2.2) V 72.2 ± 7.6 (62.4 – 86.9) 73 – 81 76 ± 1 (75 – 78) 75.1 ± 1.9 (71.4 – 77.1) 75.9 ± 1.3 (74 – 78.7) Stylet length 16.7 ± 0.7 (15.8 – 17.8) 16.5 – 18.0 16.4 ± 0.4 (15.4 – 16.9) 16.3 ± 0.8 (15 – 17) 16.0 ± 0.6 (15 – 17) Dorsal gland opening from stylet base 3.0 ± 0.5 (2.3 – 3.5) – 2.9 ± 0.3 (2.4 – 3.1) – – O 17.7 ± 2.9 (13.6 – 22.2) – 17.4 ± 1.7 (14.3 – 21.3) – – Pharynx length 85 ± 1.0 (82.5 – 85.9) – 80.1 ± 3.2 (73.1 – 84.2) 89 ± 10.4 (76 – 105) 87 ± 4 (82 – 94) Pharyngeal overlap – – 39.0 ± 7.2 (29.8 – 49.0) – Anterior end to end of pharyngeal gland lobe 101 ± 1.4 (98 – 102) – – 125 ± 7.3 (116 – 135) 123 ± 8.3 (113 – 138) Maximal body diameter 22.7 ± 0.5 (21.6 – 23.1) – 20.2 ± 2.2 (16.8 – 23.7) 23 ± 3.3 (20 – 29) 22.5 ± 2.1 (19 – 26) Anal body diameter 13.0 ± 1.2 (10.9 – 14.1) – 11.2 ± 1.2 (8.9 – 13.2) 12.4 ± 1.5 (11 – 14) 13 ± 0.9 (12 – 14) Tail length 29.0 ± 0.9 (27.5 – 30.1) – 24.8 ± 2.3 (21.0 – 27.1) 26.6 ± 2.3 (23 – 29) 24.8 ± 2.3 (23 – 29) Tail annuli 20 ± 2 (18 – 23) – 19 ± 2 (16 – 24) 20 ± 2.1 (18 – 23) 19 ± 1.5 (17 – 21) Phasmid to terminus 14.1 ± 0.4 (13.5 – 14.5) – 10.6 ± 2.7 (6.0 – 14.9) – – Description Females Vermiform and slightly curved ventrally after heat-killing and fixation. Labial region continuous from the rest of the body and lip region with three annuli. Under SEM (Figs. 1 C,D), en face view showing an oval oral aperture surrounded by six inner labial sensilla, submedian segments fused to oral disc, corresponding to head pattern group 2 according to Corbett and Clark (1983). Stylet was well developed (16–18 μm long) with anteriorly directed rounded knobs. The areolation was only visible at tail level and lateral field with four incisures, with the outer two being entirely crenate, and the inner lines being finely striated. Rounded to oval-shaped metacorpus with short isthmus, pharyngeal gland overlapped ventrally. Excretory pore ws located slightly above pharyngo-intestinal junction. There was a vulva, a transverse slit in ventral view, and well developed post-vulval uterine sac. The tail had (27–30) annuli, subcylindrical, and with rounded to conical, smooth terminus. Voucher material Vouchers (two females) are available in the UGent Nematode Collection (slide UGnem-314) of the Nematology Research Unit, Department of Biology, Ghent University, Ghent, Belgium. Molecular characterization Pratylenchus delattrei Four sequences of the D2-D3 28S rDNA (OP646167-OP646170; 571-619 bp; 1–3 bp differences), two ITS rDNA sequences (OP646172-OP646171; 731bp; 17 bp differences) and two COI sequences (OP730535-OP7330534; 421 bp; with 100% identical) were generated for P. delattrei from Minjar and Adea’a districts (Table 1; Figs. 3 A–C). Based on the D2–D3 sequences, isolates formed the highest supported clade with P. delattrei sequences from Cape Verde (KY677820) and two sequences from Iran (JX261949 and JX261948), which are 99.6–100% identical. For ITS, the P. delattrei sequences formed a maximally supported clade with three unidentified Pratylenchus sp. sequences from India (MN100134, MN100135 and MH375058) with 94–97% similarity (Fig. 3B). Finally, two identical COI sequences have been generated for the first time for P. delattrei, and these sequences were in a poorly supported sister relationship (0.68 PP) with P. parazeae (Fig. 3C). Remarks Male nematodes were not found. As first reported on chickpea and from Ethiopia, this species was recovered from the Minjar and Adea’a districts in the central parts of Ethiopia, both in the rhizosphere and the roots (Table 1). It has been also reported in other African countries, including Madagascar (cotton), Sudan (sugarcane), and Cape Verde (tomato), and from several Asian countries: South Korea, Pakistan, Oman, Iran (on tomato and eggplant, date palm, pigeon pea and peanut, and medicinal plants) (Luc, 1958; Sharma et al., 1992; Jothi et al., 2004; Mani et al., 2005; Castillo & Vovlas, 2007; MajdTaheri et al., 2013; Flis et al., 2018). The studied female morphology and morphometrics are in agreement with the original description (Luc, 1958), and other descriptions of P. delattrei from Iran and Cape Verde (MajdTaheri et al., 2013; Flis et al., 2018). The characteristics also agree without variation with the matrix code for the tabular key of Castillo and Vovlas (2007): A2 (three labial annuli), B1 (male absent), C3 (stylet length 16–17 μm), D1 (shape of spermatheca absent or reduced), E2 (V ratio = 75–79.9%), F3 (PUS: 20–24.9 μm), G3 (conoid tail shape), H1 (smooth tail tip), I1 (<30 μm pharyngeal overlapping length), J1 (four lateral field lines), K1 (smooth bands of lateral field structures), and a subcylindrical tail shape with a conical to rounded tail tip. The Ethiopian P. delattrei has a slightly longer tail compared to populations from Iran and Cape Verde (27.5–30.1 vs. 21–29 μm); however, the tail length was not included in the original description. In the phylogenetic tree of the ITS region, the Ethiopian P. delattrei sequences formed a maximally supported clade with three unidentified Pratylenchus species from India; these may therefore also represent P. delattrei based on the relatively limited molecular variability (26–49 bp difference). This study links for the first time ITS sequences to P. delattrei. Quinisulcius capitatus (Allen, 1955) Siddiqi, 1971 (Fig. 2) Figure 2: Light microscopy images of Quinisulcius capitatus. A–E: Anterior part of the body showing lip and neck regions; F,G: Vulval regions (lateral view); H: Whole female body; J,K: Lateral field showing five distinct incisures; I, L–P: Tail region lateral view. Measurements See Table 3. Table 3. Comparison of important morphological character and morphometrics of the Ethiopian Quinisulcius capitatus, found from chickpea in Ethiopia, with original description from USA (Allen, 1955), and two other Q. capitatus from Ethiopia and Canada. All measurements are in μm and in the form: mean ± s.d. (range). Character Q. capitatus from chickpea in Ethiopia (present study) Q. capitatus from pear in USA (Allen, 1955) Q. capitatus from coffee in Ethiopia (Mekete et al., 2008) “Q. capitatus”* from grass in Canada (Munawar et al., 2021) n 10 13 10 20 L 699 ± 11.6 (667 – 707) 630 – 850 630 – 790 810.3 ± 44.6 (744.0 – 911.0) a 31.2 ± 1.3 (29.7 – 34.4) 30 – 38 30.9 – 38.6 41.4 ± 1.8 (38.6 – 43.7) b 5.4 ± 0.2 (5.1 – 5.7) 5.0 – 5.8 – 5.5 ± 0.3 (5.0 – 6.3) c 14.7 ± 0.7 (14.1 – 16.3) 12 – 17 15.3 – 17.6 22.4 ± 1.1 (19.9 – 23.8) c’ 1.9 ± 0.2 (1.5 – 2.1) – – 2.6 ± 0.2 (2.2 – 3.2) V 56.6 ± 2.2 (52.7 – 59.0) 51 – 58 54.7 – 63.6 57.4 ± 1.5 (53.4 – 59.8) Stylet length 18.8 ± 0.6 (17.8 – 19.7) 16 – 18 15 – 18 18.3 ± 1.0 (15.5 – 20.4) Lip height 4.0 ± 0.5 (3.2 – 4.5) – – 4.0 ± 0.2 (3.7 – 4.4) Lip width 7.6 ± 0.5 (7.0 – 8.3) – – 7.6 ± 0.4 (6.9 – 8.3) Median bulb length 16.1 ± 0.3 (15.7 – 16.6) – – 14.0 ± 1.6 (11.3 – 16.9) Median bulb width 13.1 ± 0.6 (12.2 – 13.9) – – 10.4 ± 1.4 8.4 – 14.2 Pharyngeal length 129 ± 5.1 (123 – 138) – – 147.8 ± 5.8 (140.2 – 159.0) SE pore from anterior end 121 ± 3.3 (113 – 123) – 121 128.6 ± 5.3 (121.0 – 139.0) Midbody diameter 22.4 ± 1.0 (20.5 – 23.5) 21 – 27 – – Tail length 47.8 ± 3.0 (41.0 – 50.0) – – 35.8 ± 2.4 (31.3 – 40.4) Anal body diameter 25.3 ± 3.2 (23.0 – 34.2) – – – Phasmid position Middle of tail Middle of tail – Middle of tail * Q. capitatus Canada is likely misidentified, see remarks. Description Females The body of females spiral, or become C shaped after heat relaxation. The lip region hemispherical, set off with necks, with having five to six annulations, strong stylet (17.8–19.7 μm), long, rounded basal knobs, lateral field with five incisures. Rounded median bulb with strongly developed central valves, slender isthmus surrounded by nerve ring and conspicuous rounded cardia. Deirids absent and excretory pores at level between anterior margin and the middle of the basal pharyngeal bulb. Protruding vulva lips and poorly developed round spermatheca. The tail terminus conoid, distinctly annulated, tail cylindrical, with distinct phasmid at the middle of the tail. Male Not found Molecular characterization Quinisulcius capitatus Three identical D2–D3 of 28S (OP62631–OP626321; 490–693 bp), three identical ITS rDNA (OP646173–OP646175; 882–915 bp), and three identical COI gene (OP627909–OP627911; 350 bp) sequences were generated (Table 1; Figs. 4A–C). The D2–D3 sequences formed a maximum supported clade with nine 99–100% identical Q. capitatus sequences from Pakistan (MT703017–MT703025) (Fig. 4A). Our ITS rDNA sequences also formed a maximally supported clade with seven identical Q. capitatus ITS sequences from Pakistan (MT703005–MT703011) (Fig. 4B). However, two Q. capitatus sequences from Canada (MW027537–MW027538) are only 83% similar and were in an unresolved position with Tylenchorhynchus and Q. curvus sequences (Fig. 4B). The three identical COI sequences are the first sequences for Q. capitatus, and these sequences showed a weakly supported sister relationship with Amplimerlinius icarus and Tylenchorhynchus (0.63 vs. 0.58 PP) (Fig. 4C). Figure 3: Bayesian 50% majority rule consensus phylogeny of Pratylenchus delattrei from Ethiopia and related species based on 28S (A) and (B) ITS of rDNA genes and (C) COI of mtDNA using a GTR model. Branch support is indicated with PP. The sequences from this study were marked by blue color and bold font. Remarks The studied specimens are morphologically and morphometrically similar to the original description (Allen, 1955) and subsequent descriptions by Mekete et al. (2008), Munawar et al. (2021), and Iqbal et al. (2021), except for the slightly longer stylet compared to populations from coffee in Ethiopia (17.8–19.7 vs. 15–18 μm) and longer tail compared to the Canadian population (41.0–50.0 vs. 31.3–40.4 μm) (Table 3). All of the specimens have five incisures in the lateral field and also conoid, enlarged and striated terminus shape (Fig. 2), agreeing with genus Quinisulcius, which sets it apart from Tylenchorhynchus (5 vs. 3–4) according to the key of Hunt et al. (2012). As in the current study, males have rarely been found (Hopper, 1959; Siddiqi, 1971; Knobloch and Laughlin, 1973; Maqbool, 1982; Mekete et al., 2008; Geraert, 2011). However, Iqbal et al. (2021), described Q. capitatus male specimens from apple, tomato, maize, potato, cabbage, and onion in Pakistan. Quinisulcius capitatus is known to parasitize over 27 plants across all continents (North America, Central and South America, temperate parts of Europe, Africa, Asia, Australia, and New Zealand) (Munawar et al., 2021). In Africa, this species has been reported in Ethiopia from coffee (Mekete et al., 2008), soybean in South Africa (Mbatyoti et al., 2020), and tomato and carrot in Benin (Baimey et al., 2009). The Ethiopian Q. capitatus specimens formed a well-supported clade with the Pakistan populations in our D2–D3 of 28S and ITS rDNA, however, the tree topology to the Canadian Q. capitatus population was not resolved for both gene regions (Figs. 4A,B). This suggests that the Canadian populations may have been mislabelled. Figure 4: Bayesian 50% majority rule consensus phylogeny of Quinisulcius capitatus from Ethiopia and related species on 28S (A) and (B) ITS of rDNA genes and (C) COI of mtDNA using a GTR model. Branch support is indicated with PP. The sequences from this study were marked by blue color and bold font. Discussion Using morphological and molecular data, P. delattrei was detected for the first time in chickpea, and for the first time in Ethiopia. Other RLN species, i.e., P. zeae, P. alleni, P. alkan, P. erzurumensis P. mulchandi, P. coffeae, P. thornei, P. neglectus, P. mediterraneus, P. penetrans, P. brachyurus, and P. minyus, have previously been reported from the root and rhizosphere of chickpea, and their associated damage to crops has been widely studied in different countries (Di Vito et al., 1992; Di Vito et al., 1994, Castillo et al., 1996; Ali and Sharma, 2003; Castillo et al., 2008; Hollaway et al., 2008; Thompson et al., 2010; Sikora et al., 2018; Zwart et al., 2019; Behmand et al., 2022; Rostad et al., 2022). Accurate identification of RLN species is important in applying appropriate pest management strategies, so it is remarkable that despite the status of chickpea as an important leguminous crop, neither the presence nor the damage potential of Pratylenchus spp. have been studied in Ethiopia. Furthermore, although Pratylenchus contains over 100 species (Janssen et al., 2017b; Singh et al., 2018; Nguyen et al., 2019; Handoo et al., 2021), only four (P. zeae, P. brachyurus, P. coffeae and P. goodeyi) have so far been reported in Ethiopia (Peregrine and Bridge, 1992; Abebe et al., 2015). Similarly, the genus Quinisulcius contains over 17 species and can multiply in several host plants (Geraert, 2011; Hussain et al., 2019; Iqbal et al., 2021; Munawar et al., 2021), including chickpea (Di Vito et al., 1994; Ali and Sharma, 2003; Catillo et al., 2008), yet none of the Quinisulcius species have ever been reported from chickpea in Ethiopia. It is therefore striking that this current study has generated not only the first COI sequences of Q. capitatus and P. delattrei, but also the very first sequences of the genus Quinisulcius. Although mitochondrial genes, and especially COI, appear to be very informative for nematode diagnostics (Singh et al., 2021; Nguyen et al., 2022), nematodes remain one of the animal taxa with the lowest representation in the COI barcode database as compared to rDNA gene markers, according to a GenBank search conducted by Thomas et al. (2017). Instead, 18S and 28S ribosomal sequences have been traditionally the focus for nematode barcoding (Blaxter et al., 1998; Powers et al., 2021), although the mitochondrial COI gene is the designated marker for many animals since it is present in multiple copies per cell (Powers et al., 2021). The first reports of both species and their morphological and molecular characterizations presented in the current study form a solid basis for future research on their economic impact, their interactions with other pathogens, and the development of nematode management strategies for Ethiopian chickpea growers. It is clear that the distribution and effect of these two species in other leguminous crops in Ethiopia, as well as in other SSA countries, represent important subjects for future investigation. Acknowledgments This work was supported by a Ghent University Special Research Fund (UGent 01W04120) and Jimma University. ==== Refs Literature Cited Abebe E. Tesfamariam M. Awol S. Meressa B. H. Mesfin W. Temesgen A. Gezahegn G. Birhan A. 2015 Research on plant-parasitic and entomopathogenic nematodes in Ethiopia: a review of current state and future direction Nematology 17 741 759 Ali S. S. Sharma S. B. 2003 Nematode survey of chickpea production areas in Rajasthan, India Nematologia Mediterranea 31 147 149 Allen M. W. 1955 A review of the genus Tylenchorhynchus University of California Publications in Zoology 61 129 166 Bafokuzara N. D. 1996 Incidence of different nematodes on vegetable and fruit crops and preliminary assessment of yield loss due to Meloidogyne species in Uganda Nematology Brasil 20 32 43 Baimey H. Coyne D. Dagnenonbakin G. James B. 2009 Plant-parasitic nematodes associated with vegetable crops in Benin: Relationship with soil physico-chemical properties Nematologia Mediterranean 37 227 236 Barker K. R. Nusbaum C. J. Nelson L. A. 1969 Effects of storage temperature and extraction procedure on recovery of plant-parasitic nematodes from field soils Journal of Nematology 1 240 247 19325684 Behmand T. Berger J. Elekcioglu I. H. Aydogan A. Kahraman A. 2022 Population fluctuations and reaction of chickpea genotypes to root lesion nematodes (Pratylenchus thornei and P. neglectus) in Turkey Indian Phytopathology 75 437 447 Behmand T. Elekcioğlu N. Z. Berger J. Can C. Elekcioğlu H. 2018 Determination of plant parasitic nematodes associated with chickpea in Turkey Turkish Journals of Entomology 43 357 366 Blaxter M. L. De Ley P. Garey J. R. Liu L. X. Scheldeman P. Vierstraete A. Vanfleteren J. R. Mackey L. Y. Dorris M. Frisse L. M. Vida J. T. Thomas W. K. 1998 A molecular evolutionary framework for the phylum Nematoda Nature 392 71 75 9510248 Bowles J. Blair D. McManus D. P. 1992 Genetic variants within the genus Echinococcus identified by mitochondrial DNA sequencing Molecular and Biochemical Parasitology 54 165 73 1435857 Castillo P. Gomez-Barcina A. Jiménez-Díaz R. M. 1996 Plant-parasitic nematodes associated with chickpea in southern Spain and effect of soil temperature on reproduction of Pratylenchus thornei Nematologica 42 211 219 Castillo P. Vovlas N. 2007 Pratylenchus (Nematoda: Pratylenchidae): Diagnosis, Biology, Pathogenicity and Management Leiden, the Netherlands Brill Academic Publishers Castillo P. Navas-Cortés J. A. Landa B. B. Jiménez-Díaz R. M. Vovlas N. 2008 Plant-parasitic nematodes attacking chickpea and their inplanta interactions with rhizobia and phytopathogenic fungi Plant Disease 92 840 853 30769725 Corbett D. C. M. Clark S. A. 1983 Surface features in the taxonomy of Pratylenchus species Review de Nematology 6 85 98 Coye D. L. Cortada L. Dalzell J. J. Claudius-Cole A. O. Solveig H. Luambano N. Talwana H. 2018 Plant-parasitic nematodes and food security in sub-Saharan Africa Annual Review of Phytopathology 56 381 403 Dadi L. Regassa S. Fikre A. Mitiku D. 2005 Adoption of chickpea varieties in the central highlands of Ethiopia Ethiopian Agricultural Research Organization 62 Derycke S. Vanaverbeke J. Rigaux A. Backeljau T. Moens T. 2010 Exploring the use of cytochrome oxidase c subunit 1 (COI) for DNA barcoding of free-living marine nematodes PLoS ONE 5 e13716 21060838 Di Vito M. Greco N. Saxena M. C. 1992 Pathogenicity of Pratylenchus thornei on chickpea in Syria Nematologica Mediterranea 20 71 73 Di Vito M. Greco N. Halila H. M. Mabsoute L. Labdi M. Beniwal S. P. S. Saxena M. Singh K. B. Solh M. B. 1994 Nematodes of cool season food legumes in North Africa Nematologica Mediterreanea 22 3 10 FAOSTAT (Food and Agriculture Organization of United Nations) 2020 Dry Bean Statistical Database Accessed on 22 June 2021. http://faostat.fao.org/site/567/default.aspx#ancor Fikre A. Desmae H. Ahmed S. 2020 Tapping the economic potential of chickpea in sub-Saharan Africa Agronomy 10 1707 Flis L. Dobosz R. Rybarczyk-Mydłowska K. Wasilewska-Nascimento B. Kubicz M. Winiszewska G. 2018 First report of the lesion nematodes: Pratylenchus brachyurus and Pratylenchus delattrei on tomato (Solanum lycopersicum L.) plants in Cape Verde Helminthologia 55 88 94 31662632 Geraert E. 2011 The Dolichodoridae of the world Identification of the family Dolichodoridae (Nematoda) (520 pp). Gent, Belgium Academia Press Greco N. Di Vito M. Saxena M. C. 1992 Plant parasitic nematodes of cool season food legumes in Syria Nematologia Mediterranea 20 37 46 Handoo Z. A. Yan G. Kantor M. R. Huang D. Chowdhury I. A. Plaisance A. Bauchan G. R. Mowery J. D. 2021 Morphological and Molecular Characterization of Pratylenchus dakotaensis n. sp. (Nematoda: Pratylenchidae), a New Root-Lesion Nematode Species on Soybean in North Dakota, USA Plants 10 168 33477266 Hollaway G. J. Vanstone V. A. Nobbs J. Smith J. G. Brown J. S. 2008 Pathogenic nematodes of cereal crops in south-west Victoria, Australia Australia Plant Pathology 37 505 510 Hooper D. Hallmann J. Subbotin S. 2005 Extraction, processing and detection of plant and soil nematodes 53 86 Luc M. Sikora R. A. Bridge J. Plant parasitic nematodes in subtropical and tropical agriculture Wallingford CABI Publishing Hopper B. E. 1959 Three new species of the genus Tylenchorhynchus (Nematoda: Tylenchida) Nematology 4 23 30 Huelsenbeck J. P. Ronquist F. 2001 MRBAYES: Bayesian inference of phylogenetic trees Bioinformatics 17 754 5 11524383 Hunt D. J. Bert W. Siddiqi M. R. 2012 Tylenchidae and Dolichodoridae 209 248 Manzanilla-López R. H. Marbán-Mendoza N. Practical Plant Nematology Montecillo, Mexico Bibliotheca Basica de Agricultura Hussain S. Erum Y. I. Shahina F. 2019 Compendium of the genus Quinisulcius and observations on occurrence of Q. capitatus Pakistan Journal of Zoology 51 1599 1606 Iqbal E. Shokoohi E. Hussain S. Mashela P. W. Kazi N. 2021 Morphological and molecular characterization of plant parasitic nematode Quinisulcius capitatus (Allen, 1955) Siddiqi, 1971 from Gilgit Province, Pakistan European Journal of Plant Pathology 159 569 582 Janssen T. Karssen G. Orlando V. Subbotin S.A. Bert W. 2017a Molecular characterization and species delimiting of plant-parasitic nematodes of the genus Pratylenchus from the penetrans group (Nematoda: Pratylenchidae) Molecular Phylogenetics and Evolution 117 30 48 28778818 Janssen T. Karssen G. Couvreur M. Waeyenberge L. Bert W. 2017b The pitfalls of molecular species identification: a case study within the genus Pratylenchus (Nematoda: Pratylenchidae) Nematology 19 1 17 Jones J. T. Haegeman A. Danchin E. G. J. Gaur H. S. Helder J. Jones M. G. K. Kikuchi T. Manzanilla-Lopez R. Palomares-Rius J. Wesemael W. M. L. Roland N. P. 2013 Top 10 plant-parasitic nematodes in molecular plant pathology Molecular Plant Pathology 14 946 961 23809086 Jothi G. Babu R. S. Ramakrishnan S. Rajendran G. 2004 Management of root lesion nematode, Pratylenchus delattrei in cros-sandra using oil cakes Bioresource Technology 93 257 259 15062820 Knobloch N. A. Laughin C. W. 1973 A collection of plant parasitic nematodes (Nematoda) from Mexico with description of three new species Nematology 19 205 217 MajdTaheri Z. Tanha Maafi Z. Subbotin S. A. Pourjam E. Eskandari A. 2013 Molecular and phylogenetic studies on Pratylenchidae from Iran with additional data on Pratylenchus delattrei, Pratylenchoides alkani and two unknown species of Hirschmanniella and Pratylenchus Nematology 15 633 651 Mani A. Handoo Z. A. Livingstone S. 2005 Plant parasitic nematodes associated with date palm trees (Phoenix dactylifera) in the Sultane in Oman Nematropica 35 135 143 Maqbool M. A. 1982 Description of Quinisulcius solani n. sp., (Nematoda: Tylenchorhynchidae) with a key to the species and data on Scutylenchus koreanus from Pakistan Journal of Nematology 14 221 225 19295701 Mbatyoti A. Daneel M. S. Swart A. Marais M. De Waele D. Fourie H. 2020 Plant-parasitic nematode assemblages associated with glyphosate tolerant and conventional soybean cultivars in South Africa African Zoology 55 93 107 Mekete T. Sikora R. A. Kiewnick S. Hallmann J. 2008 Description of plant parasitic nematodes associated with coffee in Ethiopia Nematologica Mediterranea 36 69 77 Munawar M. Yevtushenko D. P. Castillo P. 2021 Integrative taxonomy, distribution, and host associations of Geocenamus brevidens and Quinisulcius capitatus from southern Alberta, Canada Journal of Nematology 53 1 15 34179819 Nguyen H. T. Nguyen T. D. Le T. M. L. Trinh Q. P. Bert W. 2022 Remarks on phylogeny and molecular variations of criconematid species (Nematoda: Criconematidae) with case studies from Vietnam Scientific Report 12 14832 Nguyen H. T. Trinh Q. P. Couvreur M. Singh R. P. Decraemer W. Bert W. 2019 Molecular and morphological characterisation of a new root-lesion nematode, Pratylenchus horti n. sp. (Tylenchomorpha: Pratylenchidae), from Ghent University Botanical Garden Nematology 21 739 752 Peregrine W. T. H. Bridge J. 1992 The lesion nematode, Pratylenchus goodeyi, an important pest of enset in Ethiopia Tropical Pest Management 38 325 326 Powers T. Harris T. Higgins R. Mullin P. Sutton L. Powers K. 2011 MOTUs, morphology, and biodiversity estimation: a case study using nematodes of the suborder Criconematina and a conserved 18S DNA barcode Journal of Nematology 43 35 48 22791913 Powers T. O. Harris T. S. Higgins R. S. Mullin P. G. Powers K. S. 2021 Nematode biodiversity assessments need vouchered databases: A BOLD reference library for plant-parasitic nematodes in the superfamily Criconematoidea Genome 64 232 241 32526150 Reen R. A. Mumford M. H. Thompson J. P. 2019 Novel sources of resistance to root-lesion nematode (Pratylenchus thornei) in a new collection of wild Cicer species (C. reticulatum and C. echinospermum) to improve resistance in cultivated chickpea Phytopathology 109 1270 1279 30895867 Riley I. T. Kelly S. J. 2002 Endo parasitic nematodes in cropping soils of Western Australia Australian Journal of Experimental Agriculture 42 49 56 Rostad E. Reen A. Humford H. Zwart S. Thompson P. 2022 Resistance to root-lesion nematode Pratylenchus neglectus identified in a new collection of two wild chickpea species (Cicer reticulatum and C. echinospermum) from Turkey Plant Pathology 71 1205 1219 Seinhorst J. W. 1959 A rapid method for the transfer of nematodes from fixative to anhydrous glycerin Nematologica 4 67 69 Sharma S. B. Smith D. H. McDonald D. 1992 Nematode constraints of chickpea and pigeon pea production in the semiarid tropics Plant Disease 76 868 874 Shiferaw B. Jones R. Silim S. Tekelewold H. Gwata E. 2007 Analysis of production costs, market opportunities and competitiveness of Desi and Kabuli chickpeas in Ethiopia Nairobi, Kenya IPMS (Improving Productivity and Market Success) of Ethiopian Farmers Project Working Paper 3. ILRI (International Livestock Research Institute) Siddiqi M. R. 1971 Structure of the oesophagus in the classification of the superfamily Tylenchoidea (Nematoda) Indian Journal of Nematology 1 25 43 Sikora R. A. Claudius-Cole B. Sikora E. J. 2018 Nematode parasites of food legumes 290 345 Sikora R. A. Coyne D. Hallman J. Timper P. Plant parasitic nematodes in subtropical and tropical agriculture 3rd ed. Wallingford, UK CABI Publishing Singh P. R. Couvreur M. Decraemer W. Bert W. 2019 Survey of slug-parasitic nematodes in East and West Flanders, Belgium, and description of Angiostoma gandavensis n. sp. (Nematoda: Angiostomidae) from arionid slugs Journal of Helminthology 94 1 11 Singh P. R. Karssen G. Couvreur M. Subbotin S. A. Bert W. 2021 Integrative taxonomy and molecular phylogeny of the plant-parasitic nematode genus paratylenchus (Nematoda: Paratylenchinae): linking species with molecular barcodes Plants 10 408 33671787 Singh P. R. Nyiragatare A. Janssen T. Couvreur M. Decraemer W. Bert W. 2018 Morphological and molecular characterisation of Pratylenchus rwandae n. sp. (Tylenchida: Pratylenchidae) associated with maize in Rwanda Nematology 20 781 794 Singh R. Sharma P. Varshney R. K. Sharma S. K. Singh N. K. 2008 Chickpea improvement: Role of wild species and genetic markers Biotechnology and Genetic Engineering Review 25 267 314 Subbotin S. A. Sturhan D. Chizhov V. N. Vovlas N. Baldwin J. G. 2006 Phylogenetic analysis of Tylenchida Thorne, 1949 as inferred from D2 and D3 expansion fragments of the 28S rRNA gene sequences Nematology 8 455 74 Thomas P. Steiner F. M. Schlick-Steiner B. C. Arthofer W. 2017 Are we ready to detect nematode diversity by next generation sequencing? Ecology and Evolution 7 4147 4151 28649327 Thompson J. P. Clewett T. G. Sheedy J. G. 2010 Occurrence of root-lesion nematodes (Pratylenchus thornei and P. neglectus) and stunt nematode (Merlinius brevidens) in the northern grain region of Australia Australasian Plant Pathology 39 254 264 Upadhaya A. Yan G. Pasche J. Kalil A. 2018 Occurrence and distribution of vermiform plant–parasitic nematodes and the relationship with soil factors in field pea (Pisum sativum) in North Dakota, USA Nematology 1 1 13 Vrain T. C. Wakarchuk D. A. Levesque A. C. Hamilton R. I. 1992 Intraspecific rDNA restriction fragment length polymorphism in the Xiphinema americanum group Fundamental and Applied Nematology 15 563 573 Zwart R. S. Thudi M. Channale S. Manchikatla P. K. Varshney R. K. Thompson J. P. 2019 Resistance to plant-parasitic nematodes in chickpea: current status and future perspectives Frontiers in Plant Science 10 966 31428112

PMC010xxxxxx/PMC10268362.txt

==== Front J Nanobiotechnology J Nanobiotechnology Journal of Nanobiotechnology 1477-3155 BioMed Central London 37322478 1944 10.1186/s12951-023-01944-w Research Young Sca-1+ bone marrow stem cell-derived exosomes preserve visual function via the miR-150-5p/MEKK3/JNK/c-Jun pathway to reduce M1 microglial polarization Wang Yuan 12 Qin Wan-yun 12 Wang Qi 123 Liu Xin-na 123 Li Xiang-hui 12 Ye Xin-qi 12 Bai Ying 12 Zhang Yan 12 Liu Pan 1 Wang Xin-lin 12 Zhou Yu-hang 12 Yuan Hui-ping yuanhp2013@126.com 1 Shao Zheng-bo shaozhengbohmu@126.com 12 1 grid.412463.6 0000 0004 1762 6325 Department of Ophthalmology, The Second Affiliated Hospital of Harbin Medical University, Harbin, China 2 grid.412463.6 0000 0004 1762 6325 Future Medical Laboratory, The Second Affiliated Hospital of Harbin Medical University, Harbin, China 3 grid.410736.7 0000 0001 2204 9268 The Key Laboratory of Myocardial Ischemia, Harbin Medical University, Ministry Education, Harbin, China 15 6 2023 15 6 2023 2023 21 19424 3 2023 29 5 2023 © The Author(s) 2023, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Background Polarization of microglia, the resident retinal immune cells, plays important roles in mediating both injury and repair responses post-retinal ischemia–reperfusion (I/R) injury, which is one of the main pathological mechanisms behind ganglion cell apoptosis. Aging could perturb microglial balances, resulting in lowered post-I/R retinal repair. Young bone marrow (BM) stem cell antigen 1-positive (Sca-1+) cells have been demonstrated to have higher reparative capabilities post-I/R retinal injury when transplanted into old mice, where they were able to home and differentiate into retinal microglia. Methods Exosomes were enriched from young Sca-1+ or Sca-1− cells, and injected into the vitreous humor of old mice post-retinal I/R. Bioinformatics analyses, including miRNA sequencing, was used to analyze exosome contents, which was confirmed by RT-qPCR. Western blot was then performed to examine expression levels of inflammatory factors and underlying signaling pathway proteins, while immunofluorescence staining was used to examine the extent of pro-inflammatory M1 microglial polarization. Fluoro-Gold labelling was then utilized to identify viable ganglion cells, while H&E staining was used to examine retinal morphology post-I/R and exosome treatment. Results Sca-1+ exosome-injected mice yielded better visual functional preservation and lowered inflammatory factors, compared to Sca-1−, at days 1, 3, and 7 days post-I/R. miRNA sequencing found that Sca-1+ exosomes had higher miR-150-5p levels, compared to Sca-1− exosomes, which was confirmed by RT-qPCR. Mechanistic analysis found that miR-150-5p from Sca-1+ exosomes repressed the mitogen-activated protein kinase kinase kinase 3 (MEKK3)/JNK/c-Jun axis, leading to IL-6 and TNF-α downregulation, and subsequently reduced microglial polarization, all of which contributes to reduced ganglion cell apoptosis and preservation of proper retinal morphology. Conclusion This study elucidates a potential new therapeutic approach for neuroprotection against I/R injury, via delivering miR-150-5p-enriched Sca-1+ exosomes, which targets the miR-150-5p/MEKK3/JNK/c-Jun axis, thereby serving as a cell-free remedy for treating retinal I/R injury and preserving visual functioning. Supplementary Information The online version contains supplementary material available at 10.1186/s12951-023-01944-w. Keywords Exosomes Bone marrow Sca-1+ cell Ischemia/reperfusion injury miR-150-5p MEKK3 the National Natural Science Foundation of China82070956 Yuan Hui-ping the Applied Technology Research and Development Program of Heilongjiang Provincial Science and Technology DepartmentGA20C008 Yuan Hui-ping the National Natural Science Foundation of China81970799 Shao Zheng-bo the Heilongjiang Postdoctoral Scientific Research Developmental FundLBHQ18082 Shao Zheng-bo issue-copyright-statement© BioMed Central Ltd., part of Springer Nature 2023 ==== Body pmcIntroduction Ischemia and reperfusion (I/R)-elicited tissue injuries contributes to morbidity and mortality for a wide variety of diseases. In the case of the retina, I/R injury, which could stem from high intraocular pressure (IOP), results in neural apoptosis [1], and subsequently vision loss, possibly to the point of blindness [2]. One key factor leading to neuron death in I/R is neuroinflammation induced by microglial activation [3, 4]. Indeed, microglia have been observed to be involved in cellular reactions associated with high IOP-caused optic neuropathy. In particular, M1 microglia have been defined as the pro-inflammatory type, producing inflammatory cytokines, such as interleukin (IL)-6, tumor necrosis factor (TNF)-α, and IL-1β in the retina, all of which may contribute to neuronal apoptosis and eventual neuro-destruction [5, 6]. Therefore, new effective treatments for I/R-induced inflammation are required to counteract against these pathological effects. I/R injury increases retinal microglia polarization towards the M1 type, and aging aggravates this polarization tendency, along with being associated with decreased microglial functioning [7, 8]. There, targeting M1 microglia could serve as a therapeutic approach for alleviating retinal I/R-induced neurotoxicity in aged populations. In our previous study [9], we discovered that when young bone marrow (BM) stem cell antigen-1 positive (Sca-1+) cells were transplanted into older mice, these stem cells were able to find their way to the retina, where they differentiated into microglia. Furthermore, these cells were able to reduce cellular apoptosis and death, after acute I/R injury, by activating the fibroblast growth factor 2/Akt signaling pathway. However, the necessity of cell transplantation for current stem cell therapies has several limitations [10], such as concerns over safety and low cell survival efficiency, due to the transplantation process itself and immunological rejection, which have become significant concerns during clinical trials [11]. Therefore, the development of an effective cell-free treatment, to provide anti-inflammatory and micro-environmental protection, with no immune rejection risk, is of great importance. Due to the paracrine activities of stem cells [12], BM cell-derived products, such as exosomes, are capable of acting as a cell-free anti-inflammatory therapy, with low immunogenicity [13]. Exosomes are virus-size membranous vesicles, originating from the endocytic compartment of cells, and range from 30 to 150 nm in diameter. A growing body of evidence indicates that they play a major role in intercellular communication in both physiological and pathological conditions [14]. Exosomes are enriched in mRNAs, miRNAs, other non-coding (nc) RNAs, proteins, and biological factors [15], allowing them to exert paracrine effects on other cells. All of these molecules within exosomes have resulted in these entities becoming a potential delivery method for treating retinal diseases, such as optic nerve crush, glaucoma, laser injury, diabetic retinopathy, etc. [16]. In particular, Xue et al. found that exosome-mediated delivery of an anti-angiogenic peptide, KV11, was more effective in counteracting against pathological angiogenesis, compared to injecting KV11 alone [17]. Additionally, Mead et al. determined that miRNAs within mesenchymal stem cell-derived exosomes were able to maintain retinal ganglion cell survival in a rat optic nerve crush model [18]. Similar observations, using extracellular matrix-localized nanovesicles, were found in a rat model, with retinal ischemia induced by severe intraocular pressure [19]. However, possible therapeutic applications of exosomes for treating I/R-related retinal injuries has not been fully examined. Exosome-associated miRNA has been reported to be more stable and resistant to RNase enzymatic activity, compared to non-exosomal miRNAs [20, 21]. miRNAs, in turn, regulate expression of multiple target genes by binding to mRNAs and inhibiting their translation, or inducing mRNA degradation [22]. In this study, we enriched exosomes from young BM Sca-1+ and Sca-1− cells and injected them into the retinas of old mice post-I/R injury. We found that Sca-1+-derived exosomes were able to reduce ganglion cell apoptosis, as well as preserve visual function, owing to them being enriched for the miRNA miR-150-5p, which has previously been found to have altered expression levels during neurodegeneration [23]. This exosomal miR-150-5p was able to reduce microglial polarization via suppressing the mitogen-activated protein kinase kinase kinase 3 (MEKK3)/JNK/c-Jun axis, leading to downregulation of pro-inflammatory cytokines, thus demonstrating neuroprotective and anti-inflammatory effects against retinal I/R injury. Materials and methods Obtaining BM Sca-1+ stem cells and culturing All animal experiments were approved by the Institutional Animal Care and Use Committee of Harbin Medical University, and were carried out in accordance with the Statement for the Use of Animals in Ophthalmic and Vision Research by the Association for Research in Vision and Ophthalmology, as well as the Guide for the Care and Use of Laboratory Animals from the National Institutes of Health. Femurs and tibias from wild-type C57BL/6 mice [2–3 months, totally 90 mice] were flushed with phosphate buffered saline (PBS) to obtain nucleated BM stem cells. These cells were then sorted into 2 categories, Sca-1+ or Sca-1− stem cells [24], using a magnetic affinity cell sorting kit, in line with the manufacturer’s instructions (Stem Cell Technology, Canada). After sorting, stem cells were cultured in Iscove's Modified Dulbecco's Medium (Biosharp, China), with 10% (v/v) exosome-depleted fetal bovine serum (FBS; SBI, USA) and 1% antibiotic–antimycotic solution (Beyotime, China), and incubated for 48 h in an incubator, at 37 °C and with 5% CO2, for exosome enrichment. Isolation and characterization of BM stem cell-derived exosomes After 48 h culture, cell supernatants were obtained via centrifugation at 5000×g for 15 min, filtered through 0.45 μm filters, and concentrated by passing them through 100K ultrafiltration tubes (Millipore, USA). Exosome precipitation and purification was performed using Exo-spin™ exosome size-exclusion columns (Cell GS, UK). Their sizes and shapes were examined using transmission electron microscopy (TEM) (Hitachi, HT-7700, Japan). Nanoparticle tracking analysis (NTA) was used to measure exosome diameter and quantities (NanoFCM, N30E, China). Specific exosome markers cluster of differentiation (CD) 9 (1:1000, Cat # 92726, Abcam), CD81 (1:1000, Cat # ET1611-87, Huabio) and CD63 (1:1000, Cat # 217345, Abcam) [25], were detected with Western blot. Establishing the retinal I/R injury mouse model and intravitreal injection of exosomes To establish the retinal I/R injury animal model, old mice (total 48; 18–20 months) were anaesthetized with 5% chloral hydrate, and both eyes were subjected to retinal I/R injury, as previously described [26]. Briefly, a 500 mL IV bottle, containing sterile salt solution, comprised the normal saline reservoir. This bottle was connected to a 32-gauge needle, and the needle was inserted into the anterior chamber of mouse eyes. The reservoir was then hung from an IV pole extension, and elevated at a height of 1.5 m, resulting in the mouse eye being subjected to 110 mmHg of hydrostatic pressure. The eye was then continuously infused with saline for 1 h, after which the needle was removed to allow for the reperfusion of retinal vasculature, resulting in induction of retinal I/R injury. Mice who received retinal I/R in both eyes were randomly assigned into the following groups: normal control without I/R (Normal), I/R, I/R + Sca-1+, and I/R + Sca-1−; both eyes were each injected intravitreally with 2 μL of exosomes right after I/R injury. Visual function detection To assess visual function, we selected mice who had a clear refractive medium in both eyes, following modeling and intravitreal injection. Light/dark box exploration and optomotor response tasks were used, as previously described [9]. For light/dark box exploration, briefly, after 2 h of dark adaptation in the dark box, mice (total 9) were placed into the light box (Fig. 2A), and visual function was evaluated, based on data collected for 10 min, in terms of time spent in the light box, as well as the number of transitions between the dark and light boxes via passing through the doorway between them. For the optomotor response task, the mouse was placed on a platform in the light box, and revolving vertical stripes, at 3 different frequencies, were projected on the surrounding LED screen (Fig. 2D–E); the number of mouse head movements elicited by the vertical stripe rotation, for 5 min, was recorded for each frequency. Visual acuity was based on the highest spatial frequency eliciting this optomotor response. Each mouse was tested 3 times per trial, and the data were averaged for further analysis. Three animals from each experimental group were tested. Retinal thickness measurements Histological analysis was used to measure total retinal thickness, as well as for the 5 retinal layers. Mice were anesthetized at 7 days post-retinal I/R, and received trans-cardiac perfusions of 4% paraformaldehyde (PFA). Eyeballs were removed and fixed in 4% PFA overnight at 4 °C, after limbal paracentesis with a sterile 32 G syringe. Fixed eyeballs were dehydrated using increasing percentages of ethanol, from 50, to 70, to 95, to 100%, then paraffin-embedded and sectioned into 4 μm sections. The sections were stained using a hematoxylin and eosin (H&E)-staining kit, following the manufacturer’s instructions (Beyotime, China). All retinal thickness measurements were performed 2 mm away from the optic disc edge. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay and immunofluorescence staining For the TUNEL assay, mouse eyeballs were harvested at 3 days post-retinal I/R, while they were harvested at 7 days post-injury for immunofluorescence staining. This difference in timepoints is due to the different phases of the I/R injury response [27], in which neuronal apoptosis occurs in the immediate aftermath of the injury, followed by inflammatory reactions, involving immune cell infiltration, as well as subsequent differentiation and polarization into microglia for the clearance of apoptotic neurons. Eyeballs were fixed overnight in 4% PFA at 4 °C, after limbal paracentesis with a sterile 32 G syringe. Subsequently, they were dehydrated using increasing percentages of sucrose, starting at 10%, then 20%, for 2 h each, and finally at 30% for 30 min, all at 4 °C. Eyeballs were then embedded in OCT compound (Sakura Finetek, Japan), and 4 µm transverse sections through the optic disc of the eye were obtained. The resulting tissue sections were fixed with 4% PFA for 20 min, permeabilized with 0.1% Triton X-100 in PBS for 10 min, and blocked with 10% goat serum in PBS for 2 h. For TUNEL staining, the TUNEL assay kit was used (Roche, Switzerland), following the manufacturer’s instructions. As for immunofluorescence staining, the sections were incubated overnight, at 4 °C, with the following antibodies: NeuN (1:50; Cat # ab177487, Abcam), CD16/CD32 (1:50; Cat # ab223200, Abcam), and ionized calcium-binding adapter molecule 1 (Iba-1, 1:50; Cat # ab283319, Abcam). Sections were then incubated with fluorescein (FITC) AffiniPure goat anti-rabbit (1:200; 111-095-003, Jackson) and Red-X-AffiniPure goat anti-mouse immunoglobulin G (IgG) antibodies (1:200; 115-295-003, Jackson). Nuclei were stained using 4′,6-diamidino-2-phenylindole (DAPI; Beyotime, China) at room temperature for 3 min. Slides were mounted using an anti-fade fluorescence mounting medium (Dako; S3023, Denmark), and fluorescence images were obtained via fluorescence microscopy (Leica, Germany). Quantifications of fluorescence intensity, as well as NeuN+ and CD16/CD32+ microglia, were performed using ImageJ, for 3 mice from each treatment group. Retinal flatmounts were carried out by first harvesting and incubating mouse eyeballs for 2 h at 4 °C, followed by dissection of their retinas. Retinas were then placed in 2% Triton X-100, diluted in PBS, for 40 min at − 80 °C, and transferred into blocking buffer (5% normal goat serum in 2% Triton X-100), to be incubated for 2 h at room temperature. Afterwards, they were incubated with Iba-1 and CD16/CD32 primary antibodies, diluted in 2% blocking buffer, overnight at 4 °C, followed by incubation with fluorescein (FITC) AffiniPure goat anti-rabbit (1:200; 111-095-003, Jackson) and Red-X-AffiniPure goat anti-mouse immunoglobulin G (IgG) antibodies (1:200; 115-295-003, Jackson) at room temperature for 2 h. Nuclei were stained using 4′,6-diamidino-2-phenylindole (DAPI; Beyotime, China) at room temperature for 3 min. Retinal flatmounts were then formed by placing the retinas on the slides, vitreous body-side down; the retinas were flattened by applying 4 symmetrical radial incisions, centered on the optic disc, and cover-slipped with Dako fluorescence mounting medium. CD16/CD32+ microglia were quantified, using ImageJ, in a blinded fashion, within a 0.01 mm2 (100 × 100 μm) rectangular region, from the peripheral edge of the retina; 3 regions from each retinal sample were used. BV2 cell culture and uptake of exosomes BV2, an immortalized murine microglial cell line, was maintained in Dulbecco’s minimal essential medium, supplemented with 10% (v/v) FBS and penicillin/streptomycin (100 units and 100 μg/mL) at 37 °C in a 5% CO2 incubator. To establish the lipopolysaccharide (LPS)-induced inflammatory cell model, BV2 at a density of 5 × 104 cells/well were seeded into 6 well plates, cultured overnight, then co-cultured with 100 μg/mL of exomes for 12 h, and exposed to 1 μg/mL LPS (Sigma, USA) for 24 h. To measure exosome uptake, BV2 cells were seeded at a density of 5 × 103 cells/well into a 24-well chamber slide and cultured overnight. Cells were then co-cultured for 2 h with 10 μM of DiI-labeled exosomes (DiI, Beyotime, China), fixed with 4% PFA for 20 min, permeabilized with 0.1% Triton X-100 in PBS for 10 min, and blocked with 10% goat serum in PBS for 2 h. Afterwards, cells were incubated overnight with primary antibodies against Iba-1 at 4 °C, then incubated with FITC AffiniPure goat anti-rabbit IgG secondary antibody for 60 min in darkness at room temperature. Nuclei were stained with DAPI, and cells imaged by fluorescence microscopy. Reverse transcription quantitative real-time PCR (RT-qPCR) Total RNA from retinas, as well as from BV2 cells, representing microglia, were extracted using TRIzol® Reagent (CWBIO, China), then reverse transcribed into cDNA using the Transcriptor First Strand cDNA Synthesis Kit, in accordance with the manufacturer’s instructions (Roche, Switzerland). qPCR was conducted using NCSYB GREEN qPCR Master Mix (NCBIOTECH, China). As for exosomes, total RNA was extracted using the RNAsimple Total RNA Kit (Tiangen, DP419, China), reverse transcribed into cDNA using the miRcute Plus miRNA First-Strand cDNA Kit (Tiangen, KR211, China), and qPCR performed using the miRcute Plus miRNA qPCR Kit (SYBR Green; Tiangen, FP411, China). Primer sequences used were listed in Additional file 1: Table S1. Western blot Radioimmunoprecipitation assay lysis buffer (Beyotime, China) was used to lyse retinas obtained after removing the lens and anterior portions of mouse eyes, as well as exosomes, plus Sca-1+, Sca-1−, and BV2 cells, for protein extraction. Proteins were quantified using the BCA Protein Assay Kit (Solarbio, China), and 15 µg of total protein was separated on SDS-PAGE gels (Epizyme, China), then transferred to polyvinylidene difluoride membranes (Millipore, Germany). Primary antibodies used included Sca-1 (1:1000, Cat # 0804-10, HUABIO), IL-6 (1:1000, Cat # 500286, ZENBIO), TNF-α (1:1000, Cat # 346654, ZENBIO), MEKK3 (1:1000, Cat # 381471, ZENBIO), phospho-SAPK (p-SAPK)/JNK (1:1000, Cat # 4668, CST), JNK (1:1000, Cat # 10023, Proteintech), p–c-Jun (1:1000, Cat # R22955, ZENBIO), c-Jun (1:1000, Cat # R23335, ZENBIO), and GAPDH (1:10000, Cat # 10494-1-AP, Proteintech), and protein levels were quantified by ImageJ. Constructing exosomal RNA libraries and high-throughput sequencing Total RNA from exosomes were extracted using TRIzol (Invitrogen, CA, USA), treated with DNase I (Takara, Kusatsu, Japan) to remove any contaminating DNA, and monitored for degradation and contamination with 1% agarose gel electrophoresis. RNA concentration and purity was measured, and 1 µg of total RNA for each sample was used to generate sequencing libraries, using the NEBNext® Multiplex Small RNA Library Prep Set for Illumina® (NEB, USA). The libraries were then sequenced using Illumina Novaseq 6000, and 50 bp single-end reads were generated. Quality control was conducted on raw reads to obtain clean reads for differential expression analysis of miRNA between different samples, using DEseq2 software (P value< 0.05, |log2foldchange|> 1). Bioinformatics analyses Target genes for differentially-expressed miRNAs were predicted using miRDB (http://www.mirdb.org/), miRWalk (http://mirwalk.umm.uni-heidelberg.de/), microT (http://diana.cslab.ece.ntua.gr/microT/), miRanda (http://www.microrna.org/microrna/home.do), and TargetScan databases (http://www.targetscan.org/). Gene Ontology (GO) functional annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted using the Database for Annotation, Visualization and Integrated Discovery (http://david.niaid.nih.gov). Retrograde labeling of retinal ganglionic cells and quantification Retrograde labeling of retinal ganglion cells was conducted 3 days prior to retinal I/R. Mice (18–20 month, totally 5 mice) were deeply anesthetized and immobilized with a small stereotactic instrument, followed by exposure of their skulls and identification of their bregma. A 22 G needle was then used to drill a hole above the superior colliculus of each hemisphere, and 1 μl of 4% hydroxystilbamidine in PBS (aka Fluoro-Gold; Cat # ab138870, Abcam, USA) was injected into both superior colliculi, using a micro-injector, at 1 mm from the bony surface of the brain. Mice were then sacrificed at days 3 and 7 post-I/R, and retinal flat mounts were prepared. Fluoro-Gold-positive ganglion cells were identified under a fluorescent microscope, and quantitated by ImageJ. Luciferase reporter assay To determine the possible target binding sequences of miR-150-5p on MEKK3 (MAP3K3), targetScan (http://www.targetscan.org/) was used. Both wild-type and mutant pMIR-MAP3K3-3’UTR luciferase reporter plasmids (Guangzhou RiboBio Co., Ltd, China) were obtained for the luciferase reporter assay, in which these plasmids were co-transfected with either miR-150-5p mimic, or negative control (NC) comprising of scrambled random miRNA, into 293T cells, using Lipofectamine 3000 (Thermo Fisher, USA). Dual-luciferase activity was then measured using the Dual-Glo Luciferase Assay System (Promega, USA). Statistical analysis All statistical analyses was conducted using GraphPad Prism version 8.0, and all values were presented as mean ± standard error of the mean (SEM). Student’s t-test was used for comparisons between 2 groups, while for 3 or more groups, one way analysis of variance (ANOVA) was used. Additionally, comparisons to “baseline” measurements was defined in terms of changes pre- and post-I/R injury, while comparisons to “Normal” was in terms of differences between un-injured versus I/R-injured mice. p < 0.05 was considered statistically significant. Results Characterization of BM Sca-1+ and Sca-1− exosomes Western blot analyses confirmed successful isolation of BM Sca-1+ and Sca-1− cells, as well as exosomes (Additional file 2: Fig. S1). Exosomes were then characterized using TEM, NTA and Western blot, where we found that exosomes from both cell types had typical round or cup-shaped morphologies, and were approximately 80 nm in diameter (Fig. 1A). A bell-shape distribution curve for exosome sizes were found for both cell types under NTA, indicating that most exosomes fell within the characteristic size range of 50–150 nm, with the highest peak at 70 nm (Fig. 1B). Western blot analysis showed that those exosomes expressed surface markers CD9, CD81, and CD63, all of which were absent in Sca-1+ or Sca-1− cells (Fig. 1C). Therefore, these findings demonstrated successful isolation of exosomes from both Sca-1+ and Sca-1− cells, adhering to the characteristic morphology, size, and surface marker profiles.Fig. 1 Characterization of bone marrow (BM) cell-derived exosomes. A Transmission electron microscopic (TEM) images of ring-shaped BMC-Sca-1+ and BMC-Sca-1− exosomes. B Nanoparticle Tracking Analysis (NTA) histograms demonstrating the size distribution for BMC-Sca-1+ and BMC-Sca-1− exosomes. C Western blot illustrating the characteristic surface markers CD63, CD9 and CD81 being present on BMC-Sca-1+ and BMC-Sca-1− exosomes, unlike with cells BM Sca-1+ exosomes improved visual behavior in I/R induced retinal damage We first confirmed that exosomes from both Sca-1+ or Sca-1− cells could be taken up by mouse retinal cells and found that they aggregated in the ganglion cell layer (Additional file 3: Fig. S2). The light/dark box exploration test was used to examine mouse visual behavior post-retinal I/R, with or without intravitreal injection of exosomes (Fig. 2A), where it was found that at baseline prior to I/R injury, all 3 animal groups (I/R, I/R + Sca-1+ exo, I/R + Sca-1− exo) preferred to remain in the dark room, and had similar durations in the light room and transition numbers between dark and light rooms (Fig. 2B, C). At 1, 3, and 7 days post-I/R, though, all 3 groups spent longer durations in the light room, and had lower transition numbers (Fig. 2B, C). However, the I/R + Sca-1+ exo group had the shortest duration in the light room, and the highest transition numbers, compared to the other 2 groups during those time periods (Fig. 2B, C).Fig. 2 BM Sca-1+ exosomes improved visual behavior after retinal ischemia–reperfusion (I/R) injuries. A Schematic illustration of the apparatus for the light/dark exploration mouse model. B, C Mice who received Sca-1+ exosomes after I/R (I/R + Sca-1+ exo) were more able to respond to light exposure, compared to those who received Sca-1− exosomes (I/R + Sca-1− exo), or untreated post-I/R. Schematic illustration (D) and photograph (E) depicting the experimental apparatus for optomotor tests. F–H Visual behaviors, in terms of head movements, under 0.1–0.3 cycles per degree (cpd) were better preserved among I/R + Sca-1+ exo mice, compared to I/R and I/R + Sca-1− exo. Data shown as mean ± SEM. n = 3 mice/group for all experiments. **P < 0.01, *P < 0.05 Another test for examining visual behavior was the optomotor response test, entailing the quantification of the number of head movements, under photopic conditions, during the rotation of the grating (Fig. 2D–E). The number of head movements/min, serving as a measure of visual function, decreased in all 3 groups following I/R compared to baseline, no matter the frequency of grating rotation (Fig. 2F–H). However, I/R + Sca-1+ exo had significantly higher head movements at days 3 and 7 days post-I/R, compared to the other 2 groups, at all 3 frequencies, though these differences were not present between I/R + Sca-1+ exo and I/R + Sca-1− exo groups, at 0.2–0.3 cycles/degree (cpd), for day 1 post-I/R (Fig. 2F–H). All these findings thus demonstrate that the application of Sca-1+-derived exosomes resulted in greater maintenance of proper visual behaviors post-I/R, in terms of light/dark box exploration and optomotor responses. Sca-1+ exosomes protect retinal morphology by reducing I/R-induced ganglion cell apoptosis H&E staining was used to examine retinal morphology for uninjured normal (Normal), I/R, I/R + Sca-1+ exo, and I/R + Sca-1− exo groups (Fig. 3A), where it was found that for both the entire retina, as well as the 5 retinal layers of ganglion cell (GCL), inner plexiform (IPL), inner nuclear (INL), outer plexiform (OPL) and outer nuclear layers (ONL), they were all thinner in I/R, compared to Normal (Fig. 3B–G). However, the administration of Sca-1+ exosomes in the I/R + Sca-1+ exo group yielded retinal thicknesses, both in total and for the 5 layers, closer to that of Normal, compared to I/R and I/R + Sca-1− exo groups (Fig. 3B–G). To further examine the basis behind retinal layer thinning post-I/R, TUNEL was performed at 3 days post-I/R, where it was found that compared to baseline, significant increases in TUNEL+/NeuN+ cells were present among I/R, I/R + Sca-1+ exo, and I/R + Sca-1− exo groups, compared to Normal (Fig. 3H, I). However, I/R + Sca-1+ exo had significantly fewer TUNEL+/NeuN+ cells, compared to the other 2 groups (Fig. 3H–I). In line with the findings of TUNEL+/NeuN+ cells being localized in GCL, Fluro-Gold labeling of viable retinal ganglion cells indicated that their numbers significantly decreased at days 3 and 7 post-I/R, compared to baseline, for I/R, I/R + Sca-1+ exo, and I/R + Sca-1− exo groups. Notably, though, I/R + Sca-1+ exo had significantly higher numbers of Fluro-Gold+ cells, compared to the other 2 groups (Additional file 4: Fig. S3A, B). All of these findings thus indicate that intravitreal injection of Sca-1+ exosomes was able to preserve proper retinal layer morphology via lowering I/R-induced GCL cell apoptosis in aged mouse retinas.Fig. 3 Sca-1+ exosomes protect retinal morphology by reducing I/R-induced retinal ganglion cell apoptosis. A Hematoxylin and eosin (H&E) staining of the retina to evaluate total B, ganglion cell (GCL; C), inner plexiform (IPL; D), inner nuclear (INL; E), outer plexiform (OPL; F), and outer nuclear (ONL; G) layer thicknesses, among control (Normal), I/R, I/R + Sca-1+ exo, and I/R + Sca-1− exo groups. Representative immunofluorescence images (H) and quantification I of apoptotic (TUNEL+) retinal neurons (NeuN+), both at baseline and at 3 days post-I/R. Data shown as mean ± SEM. n = 3 retinas/group for all experiments. **P < 0.01, *P < 0.05 Sca-1+ exosomes reduced post-I/R M1 microglial polarization I/R-induced retinal ganglion cell apoptosis may stem from M1 microglial polarization, as M1 microglia have previously been demonstrated to play important roles in post-I/R neurotoxicity [7]. Immunofluoresence staining of Iba-1, representing microglia, and CD16/32, representing M1 polarization, was thus performed, where it was found that compared to Normal, I/R, I/R + Sca-1+ exo, and I/R + Sca-1− exo groups had higher percentages of M1 microglia (Fig. 4A, B). However, compared to the other 2 I/R groups, I/R + Sca-1+ exo had significantly less M1 polarization, being closer to that of Normal (Fig. 4A, B). The same trend was present among the 4 groups when examining retinal flatmounts, in which I/R, I/R + Sca-1+ exo, and I/R + Sca-1− exo had higher M1 microglial percentages versus Normal; M1 microglial polarization levels among I/R + Sca-1+ exo was also lower than for I/R and I/R + Sca-1− exo (Additional file 5: Fig. S4A, B).Fig. 4 Sca-1+ exosomes reduced the occurrence of post-I/R microglial M1 polarization. Representative immunofluorescence images (A) and quantification (B) of M1 versus total microglia, excluding M1, among Normal, I/R, I/R + Sca-1+ exo, and I/R + Sca-1− exo groups. Relative mRNA expression levels of inflammatory factors interleukin-6 (IL-6) (C) and tumor necrosis factor-α (TNF-α) (D) among the 4 treatment groups, as determined by reverse transcription-quantitative polymerase chain reaction (RT-qPCR). Representative Western blot image (E) and analysis of IL-6 (F) and TNF-α (G) protein expression levels among the 4 groups, normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Data shown as mean ± SEM. n = 3 retinas/group for all experiments. **P < 0.01, *P < 0.05 As M1 polarization has also been associated with increased inflammation, we then examined the expression of pro-inflammatory cytokines IL-6 and TNF-α. We found higher levels of IL-6 and TNF-α, in terms of mRNA (Fig. 4C, D) and protein expression (Fig. 4E–G), in the I/R group. However, application of Sca-1+ exosomes lowered expression of these cytokines to levels closer to that of Normal, while Sca-1− exosomes did not exhibit the same effect (Fig. 4C–G). Therefore, Sca-1+ exosome treatment reduced I/R-elicited M1 microglial polarization, which in turn led to lowered retinal mRNA and protein expression of pro-inflammatory IL-6 and TNF-α. miR-150-5p is enriched in Sca-1+ exosomes and regulates microglial polarization via repressing the MAP3K3 (MEKK3)/JNK signal pathway Exosomes have been found to be enriched for miRNAs, which in turn are able to modulate the translation of target mRNAs [28]. To determine whether Sca-1+ exosomes were enriched in miRNAs, and if so, whether they are involved in modulating M1 microglial polarization, high-throughput sequencing was used to compare Sca-1+ and Sca-1− exosome contents. As shown in the volcano plot (Fig. 5A), 486 miRNAs were detected, and significant differences in abundances between Sca-1+ and Sca-1− exosome were found for 27 miRNAs. Out of those 27 miRNAs, 11 were significantly upregulated (fold change ≥ 1.0, p < 0.05), and 16 downregulated, in Sca-1+, compared to Sca-1− exosomes (Fig. 5B). The most abundant miRNA in Sca-1+ exosomes was miR-150-5p, which was then examined in the following studies (Fig. 5C). To determine the downstream targets of the 27 differentially expressed miRNAs, 5 miRNA target prediction databases were used: miRDB, miRWalk, Targetscan, microT, and miRanda, based on the binding affinities of their 3′-UTR regions (Fig. 5D). The target genes were found under GO to be most enriched for synapse organization, pre-synapse, and nucleoside-triphosphatase regulator activity, with respect to biological process (BP), cellular component (CC), and molecular function (MF), respectively (Fig. 5E). KEGG analysis found that the top 10 most enriched signaling pathways were inflammation-related, such as PI3K-Akt, MAPK, and Ras. Additionally, neurodegeneration pathways were enriched, which was in accordance with I/R injury (Fig. 5F). With respect to miR-150-5p, we found that its target matched with the 3′-UTR of MAP3K3, the human version of MEKK3 in mice, which was part of the MAPK signaling pathway (Fig. 5G). These bioinformatics analyses thus suggest that the Sca-1+ exosomes may exert its effects, via miR-150-5p, to repress the MEKK3 signaling pathway, which may subsequently reduce M1 microglial polarization and inflammation.Fig. 5 miR-150-5p is enriched within Sca-1+ exosomes and regulates microglial polarization via targeting MEKK3, the mouse analogue of mitogen-activated protein kinase kinase kinase 3 (MAP3K3). A Volcano plot showing miRNAs that are up- (red) and down (blue)-regulated in BM Sca-1+ exosomes, compared to Sca-1− ones, as determined by high throughput sequencing. B Heatmap showing the 27 differentially-expressed miRNAs between the 2 groups, with up to twofold change in expression, of which 11 were up-regulated (warm colors, orange-red) and 16 down-regulated (cool colors, light–dark blue) in Sca-1+ versus Sca-1− exosomes. C Relative abundance of the 27 miRNAs, in terms of read count, between the 2 groups. D Venn diagram demonstrating that the 27 differentially-expressed miRNA identified in our analysis were also those predicted by different databases. E Gene ontology (GO) enrichment analysis of the target genes for the 27 differentially-expressed miRNAs, in which they fell into 1 of 3 broad categories: biological process (BP), cellular component (CC), and molecular function (MF). F Kyoto genes and genomes (KEGG) enrichment analysis, in the form of a scatter plot, showing pathway associations for the target genes of these 27 miRNAs. G Predicted 3’-UTR binding site for miR-150-5p on the MEKK3 (MAP3K3) gene by TargetScan Increased miR-150-5p levels within Sca-1+ exosomes decreased M1 microglial polarization among BV2 cells in vitro via downregulating MEKK3/JNK signaling To further validate whether Sca-1+ exosomes downregulated microglia M1 polarization via miR-150-5p, RT-qPCR was used to detect relative expression levels of miR-150-5p in Sca-1+ and Sca-1− exosomes. There, we found that miR-150-5p expression was ~ 15 times higher in Sca-1+ versus Sca-1− exosomes (Fig. 6A). Furthermore, we found that Sca-1+ and Sca-1− exosomes could be taken up by BV2 cells, which was an immortalized murine microglial cell line, and no significant differences were found regarding the endocytosis of these 2 exosome types (Additional file 6: Fig. S5). Based on this finding, we established the LPS-induced inflammatory cell model, were we observed that LPS significantly reduced miR-150-5p expression in BV2 cells. This reduced expression, however, was able to be reversed back towards levels found in untreated (Normal) cells upon Sca-1+ exosome administration, but not for Sca-1− exosomes (Fig. 6B). Furthermore, LPS significantly increased expression, both in terms of mRNA and protein levels, for IL-6 and TNF-α, compared to Normal. All of these levels, though, were significantly reduced in LPS + Sca-1+ exo, but not for LPS + Sca-1− exo, whose levels were similar to that of LPS (Fig. 6C–G). These changes were most likely owed to miR-150-5p downregulating MEKK3, as LPS increased mRNA expression of MEKK3, which was lowered by Sca-1+, but not Sca-1− exosome administration (Fig. 6H). Consistent with this finding, Western blot analyses showed that levels of MEKK3, and its downstream targets of p-JNK and p–c-Jun, significantly increased under LPS stimulation, compared to Normal. By contrast, all these proteins decreased in LPS + Sca-1+ exo, but not LPS + Sca-1− exo group (Fig. 6I–L). To further confirm that MAP3K3 was the target gene of miR-150-5p, luciferase reporter assay was carried out, where it was found that luciferase activity decreased following co-transfection with miR-150-5p mimic and wild-type 3′-MAP3K3 UTR luciferase plasmid. By contrast, no changes in luciferase activity were present when the mimic was co-transfected with a luciferase plasmid containing mutant 3′-MAP3K3 UTR (Additional file 7: Fig. S6).Fig. 6 Increased miR-150-5p levels within Sca-1+ exosomes downregulate MEKK3/JNK signaling to decrease M1 microglial polarization among lipopolysaccharide (LPS)-induced BV2 microglial cells in vitro. A Relative expression levels for miR-150-5p within Sca-1+ and Sca-1− exosomes, as determined by qPCR. B Relative expression levels of miR-150-5p within 4 BV2 cell treatment groups: Normal control, LPS, as well as Sca-1+ exosomes (LPS + Sca-1+ exo), or Sca-1− exosomes (LPS + Sca-1+ exo) after LPS. Relative expression levels of inflammatory factors IL-6 (C) and TNF-α (D), as determined by RT-qPCR, among the 4 treatment groups. Representative Western blot images (E) and quantification of IL-6 (F) and TNF-α (G) protein expression levels among the 4 treatment groups, normalized to GAPDH. Relative expression levels of MEKK3 (H), the target gene of miR-150-5p, as determined by RT-qPCR, among the 4 treatment groups. Representative Western blot images (I) and quantification of MEKK3 (J), phosphorylated c-Jun N-terminal kinase (p-JNK)/JNK (K) and p–c-Jun/c-Jun (L) protein expression levels among the 4 treatment groups, normalized to GAPDH. Representative immunofluorescence images (M) and quantification (N) of mean CD16/32+ (M1 polarization biomarker) fluorescence intensity among BV2 cells in the 4 treatment groups. Data shown as mean ± SEM. n = 3/group for all experiments. **P < 0.01, *P < 0.05 We then examined M1 polarization in BV2 cells using immunofluorescence staining for Iba-1 and CD16/32 markers, where the fluorescence intensity for CD16/32, representing M1 polarization among Iba-1+ BV2 cells, significantly increased in LPS versus Normal. However, applying Sca-1+ exosomes decreased CD16/32+ intensity to a greater extent in LPS + Sca-1+ exo, compared to LPS + Sca-1− exo, indicating reduced M1 polarization (Fig. 6M, N). All of these findings from the LPS-stimulation model were in accordance with the results from retinal I/R, indicating that Sca-1+ exosomes could counteract against heightened inflammation, caused by I/R or LPS, via repressing the MEKK3/JNK/Jun pathway to decrease M1 microglial polarization. Increased miR-150-5p levels within Sca-1+ exosomes also decreased M1 microglial polarization in vivo via downregulating MEKK3/JNK signaling Having shown in an in vitro model that Sca-1+ exosomes were able to downregulate the MEKK3/JNK/Jun pathway, and subsequent inflammatory factors, via miR-150-5p, we next aim to validate in vivo in old mouse retinas from an I/R injury model. We first examined miR-150-5p expression levels within the retinas of Normal, I/R, I/R, I/R + Sca-1+ exo, and I/R + Sca-1− exo groups. As expected, I/R mice had significantly lower miR-150-5p levels, compared to Normal. However, I/R + Sca-1+ exo, restored miR-150-5p levels towards that of Normal, while miR-150-5p remained the same as I/R in the I/R + Sca-1− exo group (Fig. 7A). MEKK3 expression, for both mRNA (Fig. 7B) and protein (Fig. 7C, D), exhibited the inverse pattern, in which I/R had significantly higher levels than that of Normal, and these levels significantly decreased upon Sca-1+, but not Sca-1− exosome administration. The same pattern for MEKK3 was also found for the downstream targets p-JNK and p–c-JUN, where their protein levels were significantly higher in I/R, but decreased in IR + Sca-1+ exo. These decreases, however, did not occur in I/R + Sca-1− exo group (Fig. 7E, F). The correspondence between in vitro and in vivo findings thus demonstrates that Sca-1+ exosomes may exert its post-injury anti-inflammatory effects, via miR-150-5p repressing the MEKK3/JNK/Jun pathway.Fig. 7 Increased miR-150-5p levels within Sca-1+ exosomes downregulate MEKK3/JNK signaling within post-I/R retinas in vivo. Relative expression of miR-150-5p (A) and MEKK3 (B) within the retina among the 4 treatment groups: Normal control, I/R, as well as Sca-1+ (I/R + Sca-1+ exo) or Sca-1− (I/R + Sca-1− exo) exosomes administered after I/R injury, as determined by RT-qPCR. Representative Western blot images (C) and quantification of MEKK3 (D), p-JNK/JNK (E), and p–c-Jun/c-Jun (F) protein expression levels among the 4 treatment groups, normalized to GAPDH. Data shown as mean ± SEM. n = 3/group for all experiments. **P < 0.01, *P < 0.05 Discussion BM Sca-1+ stem cells from young donors have been noted to be able to aid in recovery and improve tissue functioning in aged recipients in multiple previous studies, such as one where old mice reconstituted with young Sca-1+ BM cells exhibited enhanced autophagy in aged hearts [29], as well as attenuated stroke-induced neurological dysfunction [30] and radiotherapy-induced cognitive impairments [31]. However, the usage of Sca-1+ stem cells is not without risk, particularly with respect to immune rejection and possible tumorigenesis stemming from radiation exposure prior to reconstitution [32]. Exosomes derived from those cells, though, could serve as a possible approach to mitigate these shortcomings, as they likely serve as the basis behind the beneficial paracrine support of Sca-1+. Indeed, over the past few years, exosome-based therapies have become a promising approach as a cell-free therapy for tissue repair [33]. In this study, we found that exosomes derived from young BM Sca-1+ cells were able to alleviate the effects of retinal I/R injury via reducing ganglion cell apoptosis. In particular, these exosomes were enriched for miR-150-5p, which repressed M1 microglial polarization and inflammatory responses, via suppressing MEKK3/JNK/Jun signaling. All of these changes, in turn, resulted in preserved visual functioning. Previously, we had demonstrated that old chimeric mice containing young BM Sca-1+ stem cell had greater reparative capabilities post-retinal I/R [9], owing to the homing and differentiation of BM Sca-1+ cells into microglia in the retina. Furthermore, these cells had excellent neurotrophic capabilities, displaying high expression levels for various neurotrophic factors [34]. Based on these findings, as well as from multiple other studies showing that stem cell-derived exosomes could serve as a cell-free therapeutic alternative to traditional stem cell therapies [35], we hypothesized that young BM Sca-1+-derived exosomes may exert the same pro-reparative effects as BM Sca-1+ stem cell in the aged retina, and that these effects may also be associated with microglia polarization alterations. Size-exclusion chromatography was used to harvest secreted exosomes from cell culture. We found, consistent with a previous report, that this method was able to isolate exosomes at a high purity, along with being reproducible and scalable for large quantities [36]. Indeed, over the past few years, exosome-based therapies have become a promising approach as a cell-free therapy for tissue repair [33]. They have particularly shown promise when delivered into the retina via intravitreal injection. For instance, one study found that intravitreal injection of exosomes containing the anti-angiogenic peptide KV11 was more effective, compared to injecting KV11 alone, in suppressing pathological angiogenesis within the retina [17]. Furthermore, intravitreal injection of exosomes derived from mesenchymal stem cells were able to promote recovery of retinal laser injury, due to these exosomes down-regulating monocyte chemotactic protein (MCP)-1 expression [37]. In line with these findings, we found in this study that exosomes derived from young BM Sca-1+ cells were able to alleviate the effects of retinal I/R injury via reducing ganglion cell apoptosis. To verify the therapeutic effect of Sca-1+ exosomes, the most intuitive approach is by using visual functional tests. One such test, the light/dark test, is based on the unconditioned preference of rodents for dark over bright environments. Thus, increases in their duration of stay in the light chamber is suggestive of diminished visual acuity [38]. We observed that I/R mice who received Sca-1+ exosomes showed robust aversion to light and greater movements, compared to untreated I/R mice, and I/R mice treated with Sca-1− exosomes. These findings were also consistent with those from the optomotor response test, another simple and rapid method for assessing visual defects in mice [39]. Its operation is based on the fact that most animals turn their heads to stabilize the images on the retina, as a compensation method in a globally moving environment. A greater optomotor response, in the form of more involuntary head movements, was recorded in I/R mice who received Sca-1+ exosomes, indicating that they had better visual acuity. On top of reduced visual functioning, retinal I/R was also found to be associated with decreased retinal thickness and GCL cell density, along with increased apoptosis [40]. In our study, these outcomes were reflected with increased TUNEL+, and decreased Fluoro-Gold+ cells, which represented, respectively, apoptotic [41] and live ganglion cells. All of these I/R-associated changes, however, were reversed upon Sca-1+ exosome administration, indicating that it exerted its reparative effects by reducing ganglion cell apoptosis, in turn preserving retinal thickness and structure. We also found that after intravitreal injection, exosomes aggregated in the ganglion. This aggregation pattern is likely due to the ganglion cell layer being the first cell layer exposed to intravitreal injection, which may trigger active endocytosis by this cell layer. It has been noted by Pollalis et al., though, that intravitreally-injected exosomes are able to enter other retinal layers, including inner plexiform, inner nuclear, outer plexiform, and outer nuclear layers [42]. However, exosome delivery in those layers was observed in the context of targeting choroidal neovascularization, which occurs below the retina; indeed, Pollalis et al. designed these exosomes to specifically target that region. This stands in contrast with our study, which focused on counteracting against I/R induced neuronal cell death within the retina itself. I/R injury is also associated with neuroinflammation and pyroptosis, of which its mechanistic basis may be due to increased M1 microglia, which has been found post-injury in various organs, such as the brain [43, 44], spine [45], heart [46], and retina [47]; these M1 microglia, in turn, release pro-inflammatory cytokines. It has been previously documented, though, that one possible approach to attenuate neuroinflammation was by reducing M1 microglial activity, leading to lowered inflammatory factor production. This modulation of M1 microglia, in turn could be carried out via exosome exposure [48, 49]. Indeed, in our study, we found increased proportions of CD16/32+ cells, which represents pro-inflammatory M1 microglia [50], post-I/R injury. This increase in M1-type microglia was concomitant with elevated levels of pro-inflammatory cytokines IL-6 and TNF-α, both of which were likely secreted by activated M1 microglia in the retina. Furthermore, it was observed that enrichment of the inner retina with Sca-1+ exosomes may be associated with decreases in retinal M1 microglia in retina and inflammatory responses, suggesting that these exosomes may serve as a possible anti-inflammatory treatment when applied intravitreally. To elucidate the underlying molecular mechanisms behind the reparative effects of Sca-1+ exosomes, bioinformatics and molecular analyses were performed. It has been documented in previous studies that miRNA could serve as a mediator responsible for the therapeutic effects of exosomes [51]. Our application of high-throughput sequencing technology helped us identify the putative exosomal miRNAs responsible for the pro-reparative phenotypes that we observed. We found that several hundred miRNAs were present within BM stem cell exosomes, and out of those differentially-expressed miRNAs, miR-150-5p emerged as a candidate of interest. miR-150-5p was previously associated with neurodegeneration, and was able to modulate matrix metalloproteinase-14 and vascular endothelial growth factor expression, serving as a possible therapeutic strategy for rheumatoid arthritis [52]. Furthermore, miR-150-5p in BM stem cell-derived exosomes was able to attenuate myocardial infarction in mice, via its targeting of Bcl-2-associated X protein [53], and slowed myocardial fibrosis progression by targeting early growth response 1 [54]. Additionally, it protected against septic acute kidney injury via downregulating MEKK3 [55]. In line with these findings, we found that miR-150-5p was most highly expressed in Sca-1+ versus Sca-1− exosomes, and that its target genes were related to synapse organization, pre-synapse and neurogenesis under GO. As for signaling pathways, miR-150-5p was most strongly associated, under KEGG, with inflammation-related signaling pathways, and the strongest candidate pathway was MAPK signaling, due to MEKK3 being identified as a common target gene for miR-150-5p in 5 separate databases. We then examined whether exosomes could deliver miRNAs into specific cells and tissues. As previously demonstrated, circulating monocytes could take up exosomes [56], and miRNAs within exosomes were able to be transferred to other types of cells [57]. This was supported by our findings, in which BV2 cells displayed bright red fluorescence, originating from Dil-labelled exosomal membranes [58], thereby indicating that these BV2 cells took up exosomes via endocytosis [59]. Microglial uptake of exosomes was further confirmed by increased abundance of miR-150-5p in BV2 cells after LPS exposure and Sca-1+ exosome administration, as miR-150-5p is usually barely expressed in LPS-stimulated BV2 cells. All these observations therefore demonstrated that Sca-1+ exosomes contained miRNA, particularly miR-150-5p, which could be transferred directly into target cells. The transfer of miRNA into microglia was further confirmed by the presence of significant down regulation of its downstream target gene, MEKK3, which was part of the MEKK3/JNK/c-Jun signaling pathway. MEKK3 downregulation, in turn, was associated with inhibition of JNK and c-Jun phosphorylation, ultimately leading to reduced expression of downstream inflammatory factors IL-6 and TNF-α. The repression of the MEKK3/JNK/c-Jun signaling pathway by miR-150-5p was further confirmed by the lack of CD16/32+, indicating M1 polarization, among LPS-activated BV2 cells co-cultured with in Sca-1+ exosomes. LPS exposure has been documented to polarize microglia towards an M1 phenotype [48], indicating that Sca-1+ exosomes inhibited this polarization in BV2 cells, via delivering miR-150-5p, which subsequently downregulated the MEKK3/JNK/c-Jun signal pathway. This in vitro finding was verified in vivo in a I/R mouse model, in which I/R mice who received Sca-1+ exosomes had increased miR-150-5p levels, as well as lowered MEKK3, p-JNK, and p–c-JUN. It should be noted, though, that miR-150-5p may not be the only factor behind the neuroprotective effects of Sca-1+ exosomes. Indeed, we found that 27 miRNAs had significant differences in expression levels between Sca-1+ and Sca-1− exosomes, and that 11 of those were significantly up-regulated in Sca-1+ exosomes. Therefore, even though miR-150-5p was the most up-regulated in Sca-1+ exosomes, the other 10 miRNAs could also contribute to the neuroprotective effect. For instance, miR-21 has been documented by Deng et al. to exert protective effects against retinal degeneration, particularly via counteracting against photoreceptor apoptosis [60]. Therefore, future studies should be conducted to determine the potential neuroprotective roles that these other exosomal miRNAs may play. Conclusion In this study, we found that exosomes produced by young BM Sca-1+ stem cells were able to counteract against the effects of retinal I/R injury, due to them being enriched for miR-150-5p, compared to those obtained from Sca-1− cells. miR-150-5p acted upon retinal microglia to repress the MEKK3/JNK/c-Jun pathway, resulting in reduced M1 microglial polarization, and subsequent lowered expression of pro-inflammatory cytokines IL-6 and TNF-α, all of which ultimately yielded decreased ganglion cell apoptosis, as well as preservation of proper retinal morphology and visual functioning. Therefore, Sca-1+ exosomes could serve as a possible cell-free treatment approach for retinal I/R injury. Supplementary Information Additional file 1: Table S1. Primer sequences used in the study. Additional file 2: Fig. S1: Western blot illustrating the specificity of stem cell antigen-1 (Sca-1) expression on magnetic bead-sorted Sca-1+ cells and exosomes, which was absent from Sca-1− cells and exosomes. Additional file 3: Fig. S2: Representative immunofluorescence images of bone marrow stem cell-derived Sca-1+ and Sca-1− exosomes (Sca-1+ and Sca-1− exo), labelled with Dil dye (red) within the ganglion cell layer (GCL) of the mouse retina, labelled with NeuN (green), at 0, 12, 24, and 48-h post-exosome injection. INL: inner nuclear layer, ONL: outer nuclear layer. Additional file 4: Fig. S3: Representative images (A) and quantification (B) of viable retinal ganglion cells, labelled by Fluoro-Gold, in I/R, as well as bone marrow stem cell-derived Sca-1+ and Sca-1− exosome groups (IR + Sca-1+ and I/R + Sca-1− exo), at baseline, 3 and 7 days after I/R injury. Data shown as mean ± standard error of the mean (SEM). n = 3/group. **P < 0.01, *P < 0.05. Additional file 5: Fig. S4: Sca-1+ exosomes reduced the occurrence of post-I/R microglial M1 polarization. Representative immunofluorescence images in retinal flatmounts (A) and quantification (B) of M1 versus total microglia, excluding M1, among Normal, I/R, I/R + Sca-1+ exo, and I/R + Sca-1− exo groups. Data shown as mean ± SEM. n = 3/group. **P < 0.01. Additional file 6: Fig. S5: Endocytosis of Sca-1+ and Sca-1− exosomes (represented by Dil dye, red) by BV2 cells (Iba-1+, green), compared to Control without exosome treatment. Additional file 7: Fig. S6: Mitogen-activated protein kinase kinase kinase 3 (MAP3K3) was a direct target of miR-150-5p. A Schematic diagram showing miR-150-5p base-pairing with wild-type (WT), but not with mutant (MUT) versions of the 3’ UTR binding site of MAP3K3. B Luciferase activity decreased following co-transfection with miR-150-5p mimic and wild-type 3’-MAP3K3 UTR luciferase plasmid, while no changes were present following co-transfection of the mimic with mutant luciferase plasmid. Luciferase activity in miR-150-5p negative control (NC, comprising scrambled control miRNA) was set at 1.0. Data shown as mean ± SEM. n = 3/group, **p < 0.01. Abbreviations ANOVA Analysis of variance BM Bone marrow BP Biological process CC Cellular component CD Cluster of differentiation cpd Cycles/degree DAPI 4′,6-Diamidino-2-phenylindole FITC Fluorescein GCL Ganglion cell layer GO Gene Ontology H&E Hematoxylin and eosin Iba-1 Ionized calcium-binding adapter molecule 1 I/R Ischemia–reperfusion injury IgG Immunoglobulin G IL Interleukin INL Inner nuclear layer IOP Intraocular pressure IPL Inner plexiform layer KEGG Kyoto Encyclopedia of Genes and Genomes LPS Lipopolysaccharide MEKK3 Mitogen-activated protein kinase kinase kinase 3 MF Molecular function ncRNA Non-coding ribonucleic acid NTA Nanoparticle tracking analysis ONL Outer nuclear layer OPL Outer plexiform layer PBS Phosphate buffered saline PFA Paraformaldehyde Sca-1 Stem cell antigen 1 SEM Standard error of the mean TEM Transmission electron microscopy TNF Tumor necrosis factor TUNEL Terminal deoxynucleotidyl transferase dUTP nick end labeling assay Acknowledgements We thank Dr. Ren-Ke Li (Division of Cardiovascular Surgery, Toronto General Hospital Research Institute, University Health Network, Toronto, ON, Canada) for his suggestions on this manuscript, Drs. Huijun Gao and Mingsi Tong (Harbin Institute of Technology) for providing light/dark box and optomotor response measurement equipment, Allwegene Technology Company (Beijing, China), Co. Ltd for assistance with high-throughput sequencing, and Alina Yao for assistance with manuscript preparation and editing. Author contributions YW, WYQ and QW contributed to the design of the study, collection and assembly of data, interpretation and wrote the manuscript; XNL contributed to exosome isolation and assisted with bioinformatics data analysis; XHL and XQY contributed to behavior tests and assisted with animal models; YB and YZ assisted with BM stem cell sorting; PL, XLW and YHZ assisted with data acquisition and analysis; ZBS and HPY helped in conceptualization, design, financial support, and final approval of manuscript. All authors read and approved the final manuscript. Funding This work was supported by grants from the National Natural Science Foundation of China (81970799 to ZBS, 82070956 to HPY), the Applied Technology Research and Development Program of Heilongjiang Provincial Science and Technology Department (GA20C008 to HPY), and the Heilongjiang Postdoctoral Scientific Research Developmental Fund (LBH-Q18082 to ZBS). Availability of data and materials The datasets used and analyzed during our study are available from the first authors on reasonable request. Declarations Ethics approval and consent to participate All animal experiments were approved by the Institutional Animal Care and Use Committee of Harbin Medical University (Permit Number: SYDW2019-141), and were carried out in accordance to the Statement for the Use of Animals in Ophthalmic and Vision Research by the Association for Research in Vision and Ophthalmology, as well as the Guide for the Care and Use of Laboratory Animals from the National Institutes of Health. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. The original online version of this article was revised: the last two authors order have been revised. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Yuan Wang, Wan-yun Qin and Qi Wang contributed equally to this work. Change history 7/12/2023 A Correction to this paper has been published: 10.1186/s12951-023-01964-6 ==== Refs References 1. Eltzschig HK Eckle T Ischemia and reperfusion—from mechanism to translation Nat Med 2011 17 11 1391 1401 10.1038/nm.2507 22064429 2. Osborne NN Casson RJ Wood JPM Chidlow G Graham M Melena J Retinal ischemia: mechanisms of damage and potential therapeutic strategies Prog Retin Eye Res 2004 23 1 91 147 10.1016/j.preteyeres.2003.12.001 14766318 3. Liu T Zhang T Yu H Shen H Xia W Adjudin protects against cerebral ischemia reperfusion injury by inhibition of neuroinflammation and blood-brain barrier disruption J Neuroinflammation 2014 14 11 107 10.1186/1742-2094-11-107 4. Chen C Li T Zhao Y Qian Y Li X Dai X Platelet glycoprotein receptor Ib blockade ameliorates experimental cerebral ischemia–reperfusion injury by strengthening the blood–brain barrier function and anti-thrombo-inflammatory property Brain Behav Immun 2018 69 255 263 10.1016/j.bbi.2017.11.019 29195783 5. Yuan L Neufeld AH Activated microglia in the human glaucomatous optic nerve head J Neurosci Res 2001 64 5 523 532 10.1002/jnr.1104 11391707 6. Nakagawa Y Chiba K Diversity and plasticity of microglial cells in psychiatric and neurological disorders Pharmacol Ther 2015 154 21 35 10.1016/j.pharmthera.2015.06.010 26129625 7. Chen D Peng C Ding XM Wu Y Zeng CJ Xu L Interleukin-4 promotes microglial polarization toward a neuroprotective phenotype after retinal ischemia/reperfusion injury Neural Regen Res 2022 17 12 2755 2760 10.4103/1673-5374.339500 35662225 8. Grabert K Michoel T Karavolos MH Clohisey S Baillie JK Stevens MP Microglial brain region-dependent diversity and selective regional sensitivities to aging Nat Neurosci 2016 19 3 504 516 10.1038/nn.4222 26780511 9. Shao Z Wu J Du G Song H Li SH He S Young bone marrow Sca-1 cells protect aged retina from ischaemia–reperfusion injury through activation of FGF2 J Cell Mol Med 2018 22 12 6176 6189 10.1111/jcmm.13905 30255622 10. Kobayashi T Kato-Itoh M Nakauchi H Targeted organ generation using Mixl1-inducible mouse pluripotent stem cells in blastocyst complementation Stem Cells Dev 2015 24 2 182 189 10.1089/scd.2014.0270 25192056 11. Moll G Ankrum JA Kamhieh-Milz J Bieback K Ringdén O Volk HD Intravascular mesenchymal stromal/stem cell therapy product diversification: time for new clinical guidelines Trends Mol Med 2019 25 2 149 163 10.1016/j.molmed.2018.12.006 30711482 12. Gnecchi M Zhang Z Ni A Dzau VJ Paracrine mechanisms in adult stem cell signaling and therapy Circ Res 2008 103 11 1204 1219 10.1161/CIRCRESAHA.108.176826 19028920 13. Harrell CR Jankovic MG Fellabaum C Volarevic A Djonov V Arsenijevic A Pham PV Molecular mechanisms responsible for anti-inflammatory and immunosuppressive effects of mesenchymal stem cell-derived factors Tissue engineering and regenerative medicine 2018 Cham Springer International Publishing 187 206 14. Whiteside TL Tumor-derived exosomes and their role in cancer progression Adv Clin Chem 2016 74 103 141 10.1016/bs.acc.2015.12.005 27117662 15. Elahi FM Farwell DG Nolta JA Anderson JD Preclinical translation of exosomes derived from mesenchymal stem/stromal cells Stem Cells 2020 38 1 15 21 10.1002/stem.3061 31381842 16. Mead B Tomarev S Extracellular vesicle therapy for retinal diseases Prog Retin Eye Res 2020 79 100849 10.1016/j.preteyeres.2020.100849 32169632 17. Dong X Lei Y Yu Z Wang T Liu Y Han G Exosome-mediated delivery of an anti-angiogenic peptide inhibits pathological retinal angiogenesis Theranostics 2021 11 11 5107 5126 10.7150/thno.54755 33859737 18. Mead B Tomarev S Bone marrow-derived mesenchymal stem cells-derived exosomes promote survival of retinal ganglion cells through miRNA-dependent mechanisms Stem Cells Transl Med 2017 6 4 1273 1285 10.1002/sctm.16-0428 28198592 19. van der Merwe Y Faust AE Sakalli ET Westrick CC Hussey G Chan KC Matrix-bound nanovesicles prevent ischemia-induced retinal ganglion cell axon degeneration and death and preserve visual function Sci Rep 2019 9 1 3482 10.1038/s41598-019-39861-4 30837658 20. Cheng L Sharples RA Scicluna BJ Hill AF Exosomes provide a protective and enriched source of miRNA for biomarker profiling compared to intracellular and cell-free blood J Extracell Vesicles. 2014 3 24734 10.3402/jev.v3.23743 21. Momen-Heravi F Bala S Bukong T Szabo G Exosome-mediated delivery of functionally active miRNA-155 inhibitor to macrophages Nanomedicine 2014 10 7 1517 1527 10.1016/j.nano.2014.03.014 24685946 22. Yu N Yong S Kim HK Choi Y Jung Y Kim D Identification of tumor suppressor miRNAs by integrative miRNA and mRNA sequencing of matched tumor–normal samples in lung adenocarcinoma Mol Oncol 2019 13 6 1356 1368 10.1002/1878-0261.12478 30913346 23. Basak I Patil KS Alves G Larsen JP Møller SG microRNAs as neuroregulators, biomarkers and therapeutic agents in neurodegenerative diseases Cell Mol Life Sci 2016 73 4 811 827 10.1007/s00018-015-2093-x 26608596 24. Li SH Sun L Yang L Li J Shao Z Du GQ Young bone-marrow Sca-1+ stem cells rejuvenate the aged heart and improve function after injury through PDGFRβ-Akt pathway Sci Rep 2017 31 7 41756 10.1038/srep41756 25. Kowal J, Arras G, Colombo M, Jouve M, Morath JP, Primdal-Bengtson B, et al. Proteomic comparison defines novel markers to characterize heterogeneous populations of extracellular vesicle subtypes. Proc Natl Acad Sci USA. 2016;113(8). 26. Hartsock MJ Cho H Wu L Chen WJ Gong J Duh EJ A mouse model of retinal ischemia-reperfusion injury through elevation of intraocular pressure JoVE 2016 113 54065 27. Qian Y Conway KL Lu X Seitz HM Matsushima GK Clarke SH Autoreactive MZ and B-1 B-cell activation by Faslpr is coincident with an increased frequency of apoptotic lymphocytes and a defect in macrophage clearance Blood 2006 108 3 974 982 10.1182/blood-2005-12-006858 16861350 28. Pillai RS Bhattacharyya SN Filipowicz W Repression of protein synthesis by miRNAs: how many mechanisms? Trends Cell Biol 2007 17 3 118 126 10.1016/j.tcb.2006.12.007 17197185 29. Yeganeh A Alibhai FJ Tobin SW Lim F Wu J Li S Age-related defects in autophagy alter the secretion of paracrine factors from bone marrow mononuclear cells Aging (Albany NY) 2021 13 11 14687 14708 10.18632/aging.203127 34088884 30. Wlodarek L Alibhai FJ Wu J Li SH Li RK Stroke-induced neurological dysfunction in aged mice is attenuated by preconditioning with young Sca-1+ stem cells Stem Cells 2022 40 564 574 10.1093/stmcls/sxac019 35291015 31. Wlodarek L Cao F Alibhai FJ Fekete A Noyan N Tobin SW Rectification of radiotherapy-induced cognitive impairments in aged mice by reconstituted Sca-1+ stem cells from young donors J Neuroinflammation 2020 17 1 51 10.1186/s12974-019-1681-3 32028989 32. Liu Y Zhao S Luo L Wang J Zhu Z Xiang Q Mesenchymal stem cell-derived exosomes ameliorate erection by reducing oxidative stress damage of corpus cavernosum in a rat model of artery injury J Cell Mol Med 2019 23 11 7462 7473 10.1111/jcmm.14615 31512385 33. Zhang Y Wang Y Shao L Pan X Liang C Liu B Knockout of beta-2 microglobulin reduces stem cell-induced immune rejection and enhances ischaemic hindlimb repair via exosome/miR-24/Bim pathway J Cell Mol Med 2020 24 1 695 710 10.1111/jcmm.14778 31729180 34. Liu X Hou M Zhang S Zhao Y Wang Q Jiang M Neuroprotective effects of bone marrow Sca-1+ cells against age-related retinal degeneration in OPTN E50K mice Cell Death Dis 2021 12 6 613 10.1038/s41419-021-03851-0 34127652 35. Rani S Ryan AE Griffin MD Ritter T Mesenchymal stem cell-derived extracellular vesicles: toward cell-free therapeutic applications Mol Ther 2015 23 5 812 823 10.1038/mt.2015.44 25868399 36. Sidhom K Obi PO Saleem A A review of exosomal isolation methods: is size exclusion chromatography the best option? Int J Mol Sci 2020 21 18 E6466 10.3390/ijms21186466 37. Yu B Shao H Su C Jiang Y Chen X Bai L Exosomes derived from MSCs ameliorate retinal laser injury partially by inhibition of MCP-1 Sci Rep 2016 30 6 34562 10.1038/srep34562 38. Thiels E Hoffman EK Gorin MB A reliable behavioral assay for the assessment of sustained photophobia in mice Curr Eye Res 2008 33 5–6 483 491 10.1080/02713680802130347 18568886 39. Abdeljalil J Hamid M Abdel-mouttalib O Stéphane R Raymond R Johan A The optomotor response: a robust first-line visual screening method for mice Vision Res 2005 45 11 1439 1446 10.1016/j.visres.2004.12.015 15743613 40. Tan J Liu G Lan C Pang IH Luo X Wu S Lentiviral vector-mediated expression of C3 transferase attenuates retinal ischemia and reperfusion injury in rats Life Sci 2021 1 272 119269 10.1016/j.lfs.2021.119269 41. Zhao Y Tan Y Xi S Li Y Li C Cui J A novel mechanism by which SDF-1β protects cardiac cells from palmitate-induced endoplasmic reticulum stress and apoptosis via CXCR7 and AMPK/p38 MAPK-mediated interleukin-6 generation Diabetes 2013 62 7 2545 2558 10.2337/db12-1233 23423573 42. Pollalis D Kim D Nair GKG Kang C Nanda AV Lee SY Intraocular RGD-engineered exosomes and active targeting of choroidal neovascularization (CNV) Cells 2022 11 16 2573 10.3390/cells11162573 36010651 43. Liu X Zhang M Liu H Zhu R He H Zhou Y Bone marrow mesenchymal stem cell-derived exosomes attenuate cerebral ischemia–reperfusion injury-induced neuroinflammation and pyroptosis by modulating microglia M1/M2 phenotypes Exp Neurol 2021 341 113700 10.1016/j.expneurol.2021.113700 33741350 44. Chen AQ Fang Z Chen XL Yang S Zhou YF Mao L Microglia-derived TNF-α mediates endothelial necroptosis aggravating blood brain–barrier disruption after ischemic stroke Cell Death Dis 2019 10 7 487 10.1038/s41419-019-1716-9 31221990 45. Zhang S Yan Y Wang Y Sun Z Han C Qian X Inhibition of MALT1 alleviates spinal ischemia/reperfusion injury-induced neuroinflammation by modulating glial endoplasmic reticulum stress in rats J Inflamm Res 2021 14 4329 4345 10.2147/JIR.S319023 34511971 46. Jing W Tuxiu X Xiaobing L Guijun J Lulu K Jie J LncRNA GAS5/miR-137 is a hypoxia-responsive axis involved in cardiac arrest and cardiopulmonary cerebral resuscitation Front Immunol 2021 12 790750 10.3389/fimmu.2021.790750 35087519 47. Yu Z Wen Y Jiang N Li Z Guan J Zhang Y TNF-α stimulation enhances the neuroprotective effects of gingival MSCs derived exosomes in retinal ischemia-reperfusion injury via the MEG3/miR-21a-5p axis Biomaterials 2022 25 284 121484 10.1016/j.biomaterials.2022.121484 48. Wang Y Gao C Gao T Zhao L Zhu S Guo L Plasma exosomes from depression ameliorate inflammation-induced depressive-like behaviors via sigma-1 receptor delivery Brain Behav Immun 2021 94 225 234 10.1016/j.bbi.2021.02.004 33607235 49. Zavatti M Gatti M Beretti F Palumbo C Maraldi T Exosomes derived from human amniotic fluid mesenchymal stem cells preserve microglia and neuron cells from Aβ Int J Mol Sci 2022 23 9 4967 10.3390/ijms23094967 35563358 50. Zhou T Huang Z Zhu X Sun X Liu Y Cheng B Alpha-1 antitrypsin attenuates M1 microglia-mediated neuroinflammation in retinal degeneration Front Immunol 2018 30 9 1202 10.3389/fimmu.2018.01202 51. Toh WS Lai RC Zhang B Lim SK MSC exosome works through a protein-based mechanism of action Biochem Soc Trans 2018 46 4 843 853 10.1042/BST20180079 29986939 52. Chen Z Wang H Xia Y Yan F Lu Y Therapeutic potential of mesenchymal cell-derived miRNA-150-5p-expressing exosomes in rheumatoid arthritis mediated by the modulation of MMP14 and VEGF J Immunol 2018 201 8 2472 2482 10.4049/jimmunol.1800304 30224512 53. Wu Z Cheng S Wang S Li W Liu J BMSCs-derived exosomal microRNA-150-5p attenuates myocardial infarction in mice Int Immunopharmacol 2021 93 107389 10.1016/j.intimp.2021.107389 33582480 54. Shen J Xing W Gong F Wang W Yan Y Zhang Y MiR-150-5p retards the progression of myocardial fibrosis by targeting EGR1 Cell Cycle 2019 18 12 1335 1348 10.1080/15384101.2019.1617614 31122130 55. Shi L Zhang Y Xia Y Li C Song Z Zhu J MiR-150-5p protects against septic acute kidney injury via repressing the MEKK3/JNK pathway Cell Signal 2021 86 110101 10.1016/j.cellsig.2021.110101 34333083 56. Ismail N Wang Y Dakhlallah D Moldovan L Agarwal K Batte K Macrophage microvesicles induce macrophage differentiation and miR-223 transfer Blood 2013 121 6 984 995 10.1182/blood-2011-08-374793 23144169 57. Qin X Xiang Y Li N Wei B Chen Y Fang D Simultaneous detection of cancerous exosomal miRNA-21 and PD-L1 with a sensitive dual-cycling nanoprobe Biosens Bioelectron 2022 14 216 114636 10.1016/j.bios.2022.114636 58. Kurczy ME Kozole J Parry SA Piehowski PD Winograd N Ewing AG Relative quantification of cellular sections with molecular depth profiling ToF-SIMS imaging Appl Surf Sci 2008 255 4 1158 1161 10.1016/j.apsusc.2008.05.018 19247454 59. Mathew B Ravindran S Liu X Torres L Chennakesavalu M Huang CC Mesenchymal stem cell-derived extracellular vesicles and retinal ischemia–reperfusion Biomaterials 2019 197 146 160 10.1016/j.biomaterials.2019.01.016 30654160 60. Deng CL Hu CB Ling ST Zhao N Bao LH Zhou F Photoreceptor protection by mesenchymal stem cell transplantation identifies exosomal MiR-21 as a therapeutic for retinal degeneration Cell Death Differ 2021 28 3 1041 1061 10.1038/s41418-020-00636-4 33082517

PMC010xxxxxx/PMC10271935.txt

==== Front Drug Alcohol Depend Rep Drug Alcohol Depend Rep Drug and Alcohol Dependence Reports 2772-7246 Elsevier S2772-7246(23)00043-4 10.1016/j.dadr.2023.100173 100173 Full Length Report Disruptions to naloxone training among lay and occupational responders in Maryland during the emergence of COVID-19: Early impacts, recovery, and lessons learned☆ Byregowda Himani hbyrego1@jhu.edu a⁎ Tomko Catherine b Schneider Kristin E. b Russell Erin c Johnson Renee M. a Susukida Ryoko a Rouhani Saba d Parnham Taylor b Park Ju Nyeong e a Department of Mental Health, Johns Hopkins Bloomberg School of Public Health b Department of Health, Behavior and Society, Johns Hopkins Bloomberg School of Public Health c Center for Harm Reduction Services, Maryland Department of Health d Department of Epidemiology, New York University School of Global Public Health e Division of General Internal Medicine, Warren Alpert Medical School, Brown University ⁎ Corresponding author at: Johns Hopkins Bloomberg School of Public Health, Johns Hopkins University, 624 North Broadway, Baltimore, MD 21215, USA. hbyrego1@jhu.edu 16 6 2023 9 2023 16 6 2023 8 10017322 3 2023 1 6 2023 5 6 2023 © 2023 The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/ This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Highlights • Immediate drop in occupational naloxone trainees after COVID-19 stay-at-home order. • No change in lay naloxone trainees immediately after COVID-19 stay-at-home order. • Naloxone trainings recovered over the 12-month period after stay-at-home order. • Strengthening connections between responder types can help maintain naloxone access. • Programs should consider alternate naloxone delivery models for future emergencies. Background Opioid overdose death rates increased during the COVID-19 pandemic. Disruptions in community-based naloxone trainings could have reduced the likelihood of overdose reversal and increased the chances of a fatal overdose. We investigated changes in the number of people trained in naloxone administration and distribution in Maryland before, during, and after COVID-related stay-at-home orders. Methods Data on naloxone training are from the Maryland Department of Health. We used interrupted time series models to estimate changes in average monthly number of people trained: [1] pre-interruption (4/2019–3/2020), [2] 1-month post-interruption (4/2020–5/2020), and [3] 12-month post-interruption (4/2020–3/2021). Trainees were classified as lay (e.g., people who use drugs) or occupational (e.g., law enforcement officers and harm reduction workers) responders. Results There were 101,332 trainees; 54.1% lay, 21.5% occupational, and 23.4% unknown responder status. We observed a decrease in the average monthly number of trainees in the pre-interruption period (-235, p<0.001), a larger decrease of 93.2% during the 1-month post-interruption (-846, p = 0.013), and an increase during the 12-month post-interruption (+217, p<0.001). There was a significant decrease among occupational responders 1-month post-interruption, and a significant increase among lay responders in the 12-month post-interruption period. Conclusions Findings suggest a marked decrease in naloxone trainees immediately after stay-at-home order, followed by a moderate rebound in the 12-months after stay-at-home order. The decrease in occupational responders trained may have limited access to naloxone, but would likely have been offset by increases in number of lay responders trained. Strengthening lay and occupational responder connections could maintain naloxone distribution during public health crises. Keywords Naloxone COVID-19 Harm reduction Overdose prevention ==== Body pmc1 Introduction Opioid overdose continues to be a major public health problem globally, and in the US (World Health Organization, 2021), and there are indications that the COVID-19 pandemic or its associated consequences (e.g., closures of businesses and workplaces) exacerbated the problem (Centers for Disease Control and Prevention, n.d.). The US drug overdose death rate increased by 31% from 2019 to 2020 (i.e., 21.6 to 28.3 per 100,000 persons), the year the pandemic emerged (Hedegaard et al., 2021). The mean emergency department (ED) visit rate for opioid overdose in the US was significantly higher in March-October of 2020 compared to the same time period in 2019 (i.e., 336.7 per 100,000 vs. 220.4 per 100,000) (Holland et al., 2021). Explanations for the increase in overdose during the emergence of the COVID-19 pandemic include solitary drug use during isolation, decreased access to substance use disorder (SUD) prevention and treatment services, changes in the composition of illicit opioids and patterns of use, and increased social and economic stress (Faust et al., 2021; Friedman and Akre, 2021). An additional explanation is that there were disruptions in harm reduction programs that distribute naloxone and offer instruction in its use. A decline in naloxone training programs could have reduced the likelihood of overdose reversals and increased the chances of a fatal overdose. Several studies are investigating how COVID-19 impacted overdose, so as to generate information about how to mitigate increases in overdose during public health emergencies. In the present study, we examine changes in the number of people trained by health officials to distribute and administer naloxone before, during, and after the implementation of the COVID-19-related statewide stay-at-home order in the state of Maryland. 1.1 Community-Based naloxone training and distribution Naloxone is an opioid antagonist that rapidly reverses the effects of an opioid overdose. Distribution of naloxone and training on how to use it has been cited by CDC and the World Health Organization as a key strategy for reducing opioid overdose fatalities (Centers for Disease Control and Prevention, 2022a; World Health Organization, 2021; Doe-Simkins et al., 2014; Walley et al., 2013). To ensure use of naloxone to prevent overdose deaths, community programs prioritize training and distribution to people who are likely to be in a position to reverse an overdose. This includes people who use drugs (PWUD), their family and friends, and people who come into contact with PWUD through their occupation. Typically, people working in community-based distribution of naloxone receive training from state or local agencies. As part of their training, they receive: [1] education about what naloxone is and how to administer it, and [2] naloxone kits to distribute. A brief naloxone training session that educates community members and health and safety workers on overdose identification and response is sufficient to improve comfort and competence in recognizing and administering naloxone in the event of an overdose (Behar et al., 2015). Several countries permit naloxone to be sold over-the-counter (World Health Organization, 2021), and recently, the naloxone nasal spray was approved for over-the-counter use without a prescription in the US (Food and Drug Administration, 2023). However, naloxone was regulated as a prescription medication in the US during and after the emergence of the COVID-19 pandemic, and community programs were the main source of naloxone. To facilitate community distribution of naloxone, states and cities have “standing orders”, which enable community programs to distribute naloxone in compliance with state regulations with oversight from state agencies. 1.2 Pandemic-related changes in community-based naloxone training and distribution The stay-at-home orders and social distancing protocols that were implemented as part of the public health response to COVID-19 pandemic may have impacted community-based naloxone programs (Courtemanche et al., 2020). In the US, the COVID-19 pandemic emerged in March 2020, and there were widespread closures of schools, clinics, and workplaces by the beginning of April 2020 (Raifman et al., 2020; Skinner et al., 2022). Community programs and harm reduction organizations were forced to limit operating hours and faced challenges in naloxone distribution efforts (Glick et al., 2020). Decreases in naloxone training programs during the early days of the pandemic would likely have resulted in a decrease in access to naloxone in the community. However, there are indications that organizations pivoted operating procedures and revised outreach efforts to ensure service provision despite stay-at-home orders and social distancing protocols (Collins et al., 2020). Program adaptations included telephone-based opioid overdose education, online naloxone trainings, drive-through services, and expansion of mail-based naloxone distribution (Barnett et al., 2021; Courser and Raffle, 2021; French et al., 2021; Hughes et al., 2022; Krawczyk et al., 2021). It remains unclear whether there was sufficient access to naloxone during the emergence of the COVID-19 pandemic. 1.3 Current study We investigate whether there were disruptions in community-based naloxone training in the state of Maryland after the emergence of the COVID-19 pandemic. Maryland is among the US states that have been highly affected by the opioid overdose crisis, due in part to the widespread proliferation of illicitly manufactured fentanyl (IMF) in the state's illicit opioid drug market. Maryland's opioid overdose death rate rose steeply during COVID-19, from 34.0 deaths per 100,000 in 2019 to 40.4 deaths per 100,000 in 2020 (Centers for Disease Control and Prevention, 2022c; 2023). Maryland has a robust statewide program for community-based naloxone distribution led by the Center for Harm Reduction Services (CHRS), within the Maryland Department of Health (MDH). CHRS authorizes individuals and local entities to dispense naloxone in community settings, including local health departments, nonprofit organizations, law enforcement agencies, SUD treatment programs, and syringe service programs. Dispensing outside of the patient-provider relationship is permitted through a state standing order (Maryland Code Health, n.d.). CHRS provides training on naloxone administration and distributes naloxone to local entities free of charge to ensure broader access among people who are likely to witness or experience an overdose. CHRS took steps to maintain operations during the COVID-19 pandemic, although the extent to which training and distribution was sustained is not known. The purpose of this study was to examine whether there was a decrease in naloxone training after the emergence of the COVID-19 pandemic. We compare the number of people who received training in naloxone administration and distribution from CHRS before, during, and after the implementation of Maryland's stay-at-home order, which was issued in March 2020. We specifically investigate changes based on different types of trainees, i.e., lay responders (e.g., PWUD and their close contacts) and occupational responders (harm reduction workers, community volunteers, and law enforcement officials). 2 Material and methods 2.1 Data on naloxone training and distribution We used CHRS program data on naloxone training and distribution from April 2019 to March 2021. CHRS monitors and coordinates community naloxone distribution efforts and provides technical assistance and guidance to local entities engaged in distribution, i.e., Overdose Response Programs (ORPs). ORPs are required by the state of Maryland to submit monthly reports of naloxone training events to CHRS through a web-based form. The data represent naloxone training and distribution by ORPs and does not reflect the total amount of naloxone dispensed in the entire state by other sources, such as pharmacies. Because CHRS does not collect any identifiable personal information, individuals in the dataset are not necessarily unique individuals, as someone may receive multiple trainings and doses of naloxone from a single ORP over a given time period. Analysis of program data was exempt from review by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board. 2.2 Study variables We calculated the monthly number of people trained in overdose response education and naloxone administration by ORPs from April 2019 through March 2021. The date of naloxone training or distribution events was used to derive the number of people trained per month. Although doses of naloxone distribution are a more common measure in the literature, examining the number of people trained allows the estimation of the capacity needed at the state-level to conduct naloxone trainings. The study period was divided into pre- and post-interruption periods based on the date of issuance of Maryland's stay-at-home order, with “interruption” referring to implementation of the order. The statewide stay-at-home order was issued on March 30, 2020 and lasted until May 15, 2020 (State of Maryland, 2020). The interruption period was April 2020, which included the only full month of the stay-at-home order. The pre-interruption period (April 2019-March 2020) represented the 12 months prior to the implementation of the stay-at-home order. We derived a 12-month post-interruption period (April 2020-March 2021). Within the post-interruption period, we also derived a 1-month post-interruption period (April 2020-May 2020) to examine immediate changes. Trainees were classified as: (1) lay responders including PWUD, family and friends of PWUD, and others interested in receiving training; (2) occupational responders including law enforcement officials, harm reduction workers, community volunteers, and people who interact with PWUD in an occupational capacity, or (3) unknown. The occupational responder group does not include EMS personnel, who receive naloxone and training in its use through their organizations. The total sample included lay and occupational responders, as well as trainees with unknown responder status. 2.3 Statistical analysis Descriptive statistics were used to summarize the total number of people trained per month, number of people trained by responder group, and doses of naloxone distributed. Next, Interrupted Time Series (ITS) analysis was conducted to examine whether the implementation of COVID-19 related stay-at-home order was associated with a change in the number of naloxone trainees. We chose this approach because the pandemic-related stay-at-home order was well-defined and implemented statewide, and because data on the number of naloxone trainees before and after the stay-at-home order was available. We estimated the average monthly number of trainees during the pre-interruption period (April 2019 to March 2020), the 12-month post-interruption period (April 2020 to March 2021), and the 1-month post-interruption period (April 2020 to May 2020). We used the ITS model to assess the statistical significance (p<0.05) of change in the average monthly number of trainees: [1] during the 12-month post-interruption period relative to the pre-interruption period, and [2] during the 1-month post-interruption period relative to the pre-interruption period. Separate ITS models were conducted for occupational responders, lay responders, and for the full sample of trainees. To account for monthly fluctuations, a smoothing technique was used to average estimates of the monthly number of trainees. Newey-West standard errors were used to adjust for heteroskedasticity and autocorrelation (Linden, 2015), and the Cumby–Huizinga test was used to assess autocorrelation and include an appropriate lag term (Cumby and Huizinga, 1992). Coefficients represent the average change per unit (i.e., month) for the time-period; negative numbers indicate an average decrease and positive numbers indicate an average increase. All statistical analyses were performed in Stata/MP version 17.0 using the itsa package for ITS analysis. 3 Results 3.1 Number of trainees per month during the pre- and post-interruption periods From April 2019 to March 2021, CHRS-affiliated ORPs held 6,325 naloxone training events. At those events, 101,332 people were trained and trainees were supplied with 295,067 doses of naloxone for use or community distribution. Among the trainees, 54.1% were lay responders, 21.5% were occupational responders, and 23.4% had unknown responder status. The number of trainees per month is summarized in the Supplemental Table. There were 57,811 trainees during the pre-interruption period (Table 1). Approximately one-half were lay responders (50.8%) and 28% were occupational responders. The mean number of trainees per month during the pre-interruption period was 4,818, and the mean was higher among lay versus occupational responders. There were 43,521 trainees during the 12-month post-interruption period. Most (58.6%) were lay responders (n = 25,501) and 12.9% were occupational responders (n = 5,623).Table 1 Number of people trained in overdose response and naloxone distribution before and after COVID-19 related stay-at-home orders in Maryland. Table 1 N Mean, per Month Median (Range) % Change: Pre- vs. Post-Interruption Pre-Interruption (Apr. 2019-Mar. 2020) Lay 29,343 2,445.25 2,211 (1,672–4,100) — Occupational 16,194 1,349.50 1,472 (623–1,729) — Unknown 12,274 1,022.83 741 (351–3,199) — Total 57,811 4,817.58 4,561 (3,203–7,259) — 1-Month Post-Interruption (Apr. 2020-May 2020) Lay 2,293 191.08 (—) (776–1,517) −92.2% Occupational 529 44.08 (—) (234–295) −96.7% Unknown 1,092 91.00 (—) (474–618) −91.1% Total 3,914 326.17 (—) (14,84–2,430) −93.2% 12-Months Post-Interruption (Apr. 2020-Mar. 2021) Lay 25,501 2,125.08 2,035 (776–3,265) −13.1% Occupational 5,623 468.58 444 (181–929) −65.3% Unknown 12,397 1,033.08 769 (368–2,863) +1.0% Total 43,521 3,626.75 3,410 (1,484–6,553) −24.7% When comparing the 1-month post-interruption period to the pre-interruption period, there was a 93.2% decrease in the average number of monthly trainees. A percent decrease of greater than 90% was observed for both lay and occupational responders. When comparing the 12-month post-interruption period to the pre-interruption period, the percent decrease in average number of monthly trainees was 24.7%. The percent decrease was higher for occupational responders (65.3%) than for lay responders (13.1%). 3.2 Estimated trends in the number of trainees Table 2 shows estimates of change in the average monthly number of trainees during the pre- and post-interruption periods that were derived from the ITS models; estimates are plotted in Fig. 1, Fig. 2. There was an estimated decrease in the monthly number of trainees during the pre-interruption period (−235, 95% CI: −329, −143) and the 1-month post-interruption period (−846, 95% CI: −1,490, −202). However, there was an estimated increase in the monthly number of trainees during the 12-month post-interruption period (+217, 95% CI: +176, +258).Table 2 ITS model-based estimates of change in the average monthly number of trainees before and after COVID-19 related stay-at-home orders in Maryland. Table 2 Coefficient 95% Confidence Interval p-value Pre-Interruption (Apr. 2019-Mar. 2020) Lay −102 −132, −72 <0.001 Occupational −58 −94, −23 0.003 Total −235 −329, −143 <0.001 1-Month Post-Interruption (Apr. 2020-May 2020) Lay −198 −551, +155 0.260 Occupational −557 −878, −236 0.002 Total −846 −1,490, −202 0.013 12-Months Post-Interruption (Apr. 2020-Mar. 2021) Lay +102 +57, +147 <0.001 Occupational +14 −3, +30 0.100 Total +217 +176, +258 <0.001 Fig. 1 Plot of ITS model-based estimates of change in the average monthly number of trainees before and after COVID-19 related stay-at-home orders in Maryland. Fig. 1 Fig. 2 Plot of ITS model-based estimates of change in the average monthly number of trainees before and after COVID-19 related stay-at-home orders in Maryland, by type of responder. Fig. 2 Estimated decreases in the monthly number of trainees during the pre-interruption period were modest for both lay (−102, 95% CI: −132, −72) and occupational (−58, 95% CI: −94, −23) responders. By contrast, estimates of decreases during the 1-month post-interruption period were substantial among occupational responders (−557, 95% CI: −878, −236), but not among lay responders (−198, 95% CI: −551, +155). During the 12-month post-interruption period, estimates indicate an increase in the monthly number of trainees among lay responders (+102, 95% CI: +57, +147), but not among occupational responders (+14, 95% CI: −3, +30). 4 Discussion Community-based programs that provide training in naloxone administration and give trainees naloxone to use and/or distribute to others are critical for overdose prevention. There are indications that there were disruptions to naloxone training programs during the emergence of the COVID-19 pandemic, and we explored trends in the number of trainees in the present study. Specifically, we investigated changes in average monthly number of people trained in overdose prevention before, during, and after the implementation of COVID-related stay-at-home orders in Maryland. Nearly one-half of the trainees were lay responders, 21.5% were occupational responders, and the rest had unknown responder status. We found a significant decrease in the number of people trained immediately following the pandemic-related stay-at-home orders, which appeared to be driven by a reduction in the number of occupational responders trained. There was a subsequent recovery in the number of trainees by March 2021, likely driven by increases in trainees who were lay responders. Our findings align with previous investigations that describe the impact of the COVID-19 pandemic on harm reduction service operations; specifically, several studies note reduced capacity to deliver naloxone training programs and missed patient contacts during suspension of services (McDonald et al., 2022; Noyes et al., 2021; Swann et al., 2022). Training of both lay and occupational responders are an important component of maintaining naloxone availability in the community. We examined changes in training patterns separately among each group. During the study period, Maryland ORPs trained 54,844 lay responders and 21,817 occupational responders. There are several possible reasons why naloxone trainings declined among occupational responders but not lay responders. Occupational responders may have been forced to prioritize COVID-19 response activities, leaving less time for overdose prevention training. For example, law enforcement departments have reported suspension of in-service trainings, community outreach initiatives, and reassignment of personnel to high traffic public areas to maintain public order (Jennings and Perez, 2020). Reduced operation hours and pandemic-related service disruptions at harm reduction organizations including syringe service programs may have also affected trainings among some occupational responders during the pandemic. It is also plausible that hiring freezes entailed less training need for new employees of these organizations. Trends in naloxone trainings among lay responders recovered quickly while trainings among occupational responders did not recover to pre-pandemic levels after 12 months. Lay responders trained through Maryland ORPs are reached primarily through harm reduction organizations such as syringe services programs and peer outreach organizations. Our findings underscore that harm reduction organizations may be more resilient and effective in reaching populations that are in need of naloxone especially in times of crisis. The rapid recovery of lay responder trainings within six months of the pandemic suggests that ORPs were able to adapt their operations. For example, some programs have added telephone-based, mail-based, and virtual naloxone trainings to their programming (Hughes et al., 2022; Krawczyk et al., 2021). There was also a significant downtrend in naloxone trainings in the 12 months before implementation of the COVID-related stay-at-home order. A statewide expansion of the Naloxone Standing Order in 2017 eliminated naloxone training and prescription requirements for obtaining naloxone and expanded access to naloxone (Taylor et al., 2022). Although the standing order only allowed physicians and pharmacists to dispense naloxone without requiring a prescription or training, it may have inadvertently reduced demand for community-based naloxone trainings. It is also plausible that the demand for trainings stabilized by 2020 since community members and professionals may have received training in previous years. However, more research is needed to understand the effect of standing order policies on community naloxone training and distribution. Our study has some limitations. Naloxone training data were provided by ORPs and delays in reporting could have led to measurement error. Data missingness due to lack of timely reporting could also have affected the quality of the data. To mitigate data quality issues, CHRS contacted program administrators at ORPs across the state to verify data. The dataset did not contain information on whether an individual receiving naloxone training was a first-time trainee and did not allow differentiation between first-time and repeat trainees. Number of people trained does not equate to the number of people reached because secondary naloxone distribution was common during the pandemic and may have led to an underestimate of the people reached. Generalizability of findings beyond Maryland may be limited. Finally, results should be viewed considering varied timing of stay-at-home orders versus broader closures in businesses, schools, and prohibitions on large social gatherings. While the stay-at-home order in Maryland ended in May 2020, some counties chose to extend stay-at-home orders while broader closures or limitations on operations largely continued across the state. The overall results indicate a decrease in naloxone trainings immediately after the implementation of the stay-at-home order, but the slower rebound among occupational responders may reflect continued limitations on some operations affecting occupational responders that may not have affected lay responders in the same way. Nonetheless, this study examines a unique database and provides important public health implications. The decrease in naloxone trainings and slow recovery among occupational responders compared to lay responders indicate the need for strengthening the relationship between lay and occupational responders during public health crises. Lay responders have demonstrated their willingness to maintain naloxone supply in the community and play a crucial role in filling gaps in coverage. Therefore, it could be useful to explore models to increase engagement among lay individuals, such as community-based organizations or non-profits. In case of future pandemics or other emergencies that occur while the opioid overdose epidemic continues, when traditional models of care may be inadmissible or inaccessible, local harm reduction organizations and health departments should consider targeted outreach services, online trainings, and mobile delivery of naloxone that proved successful at continuing access to naloxone (Bolinski et al., 2022). Programs should create contingency plans for staff and supply shortages by maintaining additional stock of naloxone and cross training staff members (Noyes et al., 2021). At the policy level, given that naloxone and other harm reduction services were designated as “essential services” during the pandemic, such organizations should be provided state and federal support and funding to continue operations. It is imperative to identify and fund strategies that were successful in supporting harm reduction organizations and broadening naloxone access to equip for future emergencies. 5 Conclusions In this study conducted in Maryland, we found a decrease in community-based naloxone training immediately following the implementation of COVID-19 related stay-at-home order, driven by a substantial decrease in training among occupational responders. However, naloxone trainings recovered to pre-pandemic levels particularly among lay responders, indicating that overdose response programs are resilient and effectively adapted their operations to maintain naloxone access in the community. Declaration of Competing Interest None Appendix Supplementary materials Image, application 1 Acknowledgements The authors appreciate the work of the Overdose Response Program Staff, at the Center for Harm Reduction Services (CHRS) within Maryland Department of Health, including Zoe Renfro, Kyle Kenny, Elizabeth Murphy, Margaret Rybak, and Dana Heilman for her contributions while serving at CHRS. We are also grateful to the staff working in local naloxone distribution programs across Maryland, who gathered the data used in this report. ☆ Role of Funding Source: This research received funding from the Overdose Data to Action (OD2A) project is supported by Cooperative Agreement number 6NU17CE924961 from the Centers for Disease Control and Prevention (CDC) to Maryland Department of Health, Public Health Services. Dr. Tomko is supported by a training grant from the National Institute of Drug Abuse (T32DA007292). Contents are solely the responsibility of the authors and do not necessarily represent the official views of the CDC or NIH. Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.dadr.2023.100173. ==== Refs References Barnett B.S. Wakeman S.E. Davis C.S. Favaro J. Rich J.D. Expanding Mail-Based Distribution of Drug-Related Harm Reduction Supplies Amid COVID-19 and Beyond Am. J. Public Health 111 6 2021 1013 1017 33950718 Behar E. Santos G.-M. Wheeler E. Rowe C. Coffin P.O. Brief overdose education is sufficient for naloxone distribution to opioid users Drug Alcohol Depend. 148 2015 209 212 25595053 Bolinski R.S. Walters S. Salisbury-Afshar E. Ouellet L.J. Jenkins W.D. Almirol E. Van Ham B. Fletcher S. Johnson C. Schneider J.A. Ompad D. Pho M.T. The Impact of the COVID-19 Pandemic on Drug Use Behaviors, Fentanyl Exposure, and Harm Reduction Service Support among People Who Use Drugs in Rural Settings Int. J. Environ. Res. Public Health 19 4 2022 10.3390/ijerph19042230 Centers for Disease Control and Prevention. (2022). Evidence-Based Strategies for Preventing Opioid Overdose: what's Working in the United States. https://www.cdc.gov/drugoverdose/featured-topics/evidence-based-strategies.html. Centers for Disease Control and Prevention. (2022). The drug overdose epidemic: behind the numbers. https://www.cdc.gov/opioids/data/index.html. Centers for Disease Control and Prevention. (2022). 2019-2020 Drug Overdose Death Rate Percent Change Map. (2022, October 7). https://www.cdc.gov/drugoverdose/deaths/2019-2020-increase.html. Centers for Disease Control and Prevention. (2023). SUDORS Dashboard: fatal Overdose Data. (2023, February 8). https://www.cdc.gov/drugoverdose/fatal/dashboard/index.html. Collins A.B. Ndoye C.D. Arene-Morley D. Marshall B.D.L. Addressing co-occurring public health emergencies: the importance of naloxone distribution in the era of COVID-19 Int. J. Drug Policy 83 2020 102872 Courser M.W. Raffle H. With crisis comes opportunity: unanticipated benefits resulting from pivots to take-home naloxone (THN) programs during the COVID-19 pandemic J. Subst. Abuse Treat. 122 2021 108220 Courtemanche C. Garuccio J. Le A. Pinkston J. Yelowitz A. Strong Social Distancing Measures In The United States Reduced The COVID-19 Growth Rate Health Aff. 39 7 2020 1237 1246 Cumby R.E. Huizinga J. Testing the Autocorrelation Structure of Disturbances in Ordinary Least Squares and Instrumental Variables Regressions Econometrica: Journal of the Econometric Society 60 1 1992 185 195 Doe-Simkins M. Quinn E. Xuan Z. Sorensen-Alawad A. Hackman H. Ozonoff A. Walley A.Y. Overdose rescues by trained and untrained participants and change in opioid use among substance-using participants in overdose education and naloxone distribution programs: a retrospective cohort study BMC Public Health 14 2014 297 24684801 Faust J.S. Du C. Mayes K.D. Li S.-X. Lin Z. Barnett M.L. Krumholz H.M. Mortality From Drug Overdoses, Homicides, Unintentional Injuries, Motor Vehicle Crashes, and Suicides During the Pandemic, March-August 2020 JAMA 326 1 2021 84 86 34019096 French R. Favaro J. Aronowitz S.V. A free mailed naloxone program in Philadelphia amidst the COVID-19 pandemic Int. J. Drug Policy 94 2021 103199 Friedman J. Akre S. COVID-19 and the Drug Overdose Crisis: uncovering the Deadliest Months in the United States, January‒July 2020 Am. J. Public Health 111 7 2021 1284 1291 33856885 Glick S.N. Prohaska S.M. LaKosky P.A. Juarez A.M. Corcorran M.A. Des Jarlais D.C The Impact of COVID-19 on Syringe Services Programs in the United States AIDS Behav. 24 9 2020 2466 2468 32333209 Hedegaard H. Miniño A.M. Spencer M.R. Warner M. Drug Overdose Deaths in the United States 2021 10.15620/cdc:112340 1999–2020. NCHS Data Brief, no 428 Holland K.M. Jones C. Vivolo-Kantor A.M. Idaikkadar N. Zwald M. Hoots B. Yard E. D'Inverno A. Swedo E. Chen M.S. Petrosky E. Board A. Martinez P. Stone D.M. Law R. Coletta M.A. Adjemian J. Thomas C. Puddy R.W. …Houry D. Trends in US Emergency Department Visits for Mental Health, Overdose, and Violence Outcomes Before and During the COVID-19 Pandemic JAMA Psychiatry 78 4 2021 372 379 33533876 Hughes T.M. Kalicki A. Huxley-Reicher Z. Toribio W. Samuels D.L. Weiss J.J. Herscher M. Wang L. A medical student-led model for telephone-based opioid overdose education and naloxone distribution during the COVID-19 pandemic Subst. Abus. 43 1 2022 988 992 35435812 Jennings W.G. Perez N.M. The Immediate Impact of COVID-19 on Law Enforcement in the United States American Journal of Criminal Justice: AJCJ 45 4 2020 690 701 32837154 Krawczyk N. Fawole A. Yang J. Tofighi B. Early innovations in opioid use disorder treatment and harm reduction during the COVID-19 pandemic: a scoping review Addict Sci Clin Pract 16 1 2021 68 34774106 Linden A. Conducting Interrupted Time-series Analysis for Single- and Multiple-group Comparisons Stata J 15 2 2015 480 500 Maryland Code Health - General § 13-3101. (n.d.). Justia Law. Retrieved October 28, 2022, from https://law.justia.com/codes/maryland/2017/health-general/title-13/subtitle-31/section-13-3101/. McDonald R. Eide D. Abel-Ollo K. Barnsdale L. Carter B. Clausen T. Day E. Fonseca F. Holmén E. Horsburgh K. Kelleher M. Kåberg M. Ladenhauf M. McAuley A. Metrebian N. Neale J. Parkin S. Ratcliffe K. Rintoul C. …Strang J. A rapid assessment of take-home naloxone provision during COVID-19 in Europe Int. J. Drug Policy 107 2022 103787 Noyes E. Yeo E. Yerton M. Plakas I. Keyes S. Obando A. Gaeta J.M. Taveras E.M. Chatterjee A. Harm Reduction for Adolescents and Young Adults During the COVID-19 Pandemic: a Case Study of Community Care in Reach Public Health Rep. 136 3 2021 301 308 33673755 Raifman J., Nocka K., Jones D., et al. (2020). COVID-19 US State Policy Database. Published online June 4, 2020. doi:10.3886/E119446V1. Accessed September 19, 2022. https://statepolicies.com/. Skinner A. Flannery K. Nocka K. Bor J. Dean L.T. Jay J. Lipson S.K. Cole M.B. Benfer E.A. Scheckman R. Raderman W. Jones D.K. Raifman J. A database of US state policies to mitigate COVID-19 and its economic consequences BMC Public Health 22 2022 1124 35659285 Swann W.L. Schreiber T.L. Kim S.Y. McNeely H. Hong J.H. Perceived impact of COVID-19 on prevention, treatment, harm reduction, and recovery services for opioid use disorder: national survey of local health departments Subst. Abus. 43 1 2022 993 998 35435800 Taylor M. Pradhan A. Ogando Y.M. Shaya F. Impact of the naloxone standing order on trends in opioid fatal overdose: an ecological analysis Am. J. Drug Alcohol Abuse 2022 1 9 34932401 The State of Maryland. (2020). Order of the Governor of the State of Maryland Number 20-03-30-0. https://governor.maryland.gov/wp-content/uploads/2020/03/Gatherings-FOURTH-AMENDED-3.30.20.pdf. U.S. Food and Drug Administration. (2023). FDA Approves First Over-the-Counter Naloxone Nasal Spray. Retrieved May 22, 2023, from https://www.fda.gov/news-events/press-announcements/fda-approves-first-over-counter-naloxone-nasal-spray. Walley A.Y. Xuan Z. Hackman H.H. Quinn E. Doe-Simkins M. Sorensen-Alawad A. Ruiz S. Ozonoff A. Opioid overdose rates and implementation of overdose education and nasal naloxone distribution in Massachusetts: interrupted time series analysis BMJ 346 2013 f174 23372174 World Health Organization. (2021). Opioid overdose. https://www.who.int/news-room/fact-sheets/detail/opioid-overdose.

PMC010xxxxxx/PMC10275623.txt

==== Front IEEE Open J Eng Med Biol IEEE Open J Eng Med Biol 0076400 OJEMB IOJEA7 IEEE Open Journal of Engineering in Medicine and Biology 2644-1276 IEEE 37332482 10.1109/OJEMB.2023.3275051 OJEMB-00005-2023 Article Reverse Correlation Uncovers More Complete Tinnitus Spectra Hoyland Alec Hoyland Alec ahoyland@wpi.edu https://orcid.org/0009-0000-0363-1244 Barnett Nelson V. Barnett Nelson V. nbarnett@wpi.edu Roop Benjamin W. Roop Benjamin W. broop@wpi.edu Alexandrou Danae Alexandrou Danae dalexandrou@luc.edu Caplan Myah Caplan Myah macaplan@wpi.edu Mills Jacob Mills Jacob jmills@wpi.edu https://orcid.org/0000-0003-2610-2884 Parrell Benjamin Parrell Benjamin bparrell@wisc.edu Chari Divya A. Chari Divya A. Divya.Chari@umassmemorial.org https://orcid.org/0000-0001-8162-2693 Lammert Adam C. Lammert Adam C. alammert@wpi.edu The review of this letter was arranged by Editor P. Bonato. (Corresponding author: Adam C. Lammert.) Department of Biomedical Engineering (BME) Worcester Polytechnic Institute (WPI) 8718 Worcester MA 01609 USA Clarifai, Inc. Wilmington DE 19808 USA BME Worcester Polytechnic Institute 8718 Worcester MA 01609 USA Neuroscience Program Worcester Polytechnic Institute 8718 Worcester MA 01609 USA Stritch School of Medicine Loyola University Chicago 2456 Chicago IL 60660 USA Department of Communication Sciences and Disorders and the Waisman Center University of Wisconsin 5228 Madison WI 53707 USA University of Massachusetts Chan Medical School 12262 Worcester MA 01609 USA Massachusetts Eye and Ear Infirmary 1866 Boston MA 02114 USA 2023 18 5 2023 4 116118 06 1 2023 29 3 2023 25 4 2023 25 4 2023 14 6 2023 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ 2023 Author https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ ojemb-lammert-3275051.xml ojemb-lammert-3275051.pdf ojemb-lammert-3275051.pdf Goal: This study validates an approach to characterizing the sounds experienced by tinnitus patients via reverse correlation, with potential for characterizing a wider range of sounds than currently possible. Methods: Ten normal-hearing subjects assessed the subjective similarity of random auditory stimuli and target tinnitus-like sounds (“buzzing” and “roaring”). Reconstructions of the targets were obtained by regressing subject responses on the stimuli, and were compared for accuracy to the frequency spectra of the targets using Pearson's \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$r$\end{document}. Results: Reconstruction accuracy was significantly higher than chance across subjects: buzzing: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$0.52 \pm 0.27$\end{document} (mean \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\pm$\end{document} s.d.), \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$t(9) = 5.766$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$p < 0.001$\end{document}; roaring: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$0.62 \pm 0.23$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$t(9) = 5.76$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$p < 0.001$\end{document}; combined: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$0.57 \pm 0.25$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$t(19) = 7.542$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$p < 0.001$\end{document}. Conclusion: Reverse correlation can accurately reconstruct non-tonal tinnitus-like sounds in normal-hearing subjects, indicating its potential for characterizing the sounds experienced by patients with non-tonal tinnitus. Reverse correlation tinnitus behavioral assay University of Massachusetts Center for Clinical and Translation Science This work was supported by the University of Massachusetts Center for Clinical and Translation Science. ==== Body pmcI. Introduction Tinnitus the perception of sound in the absence of any corresponding external stimulus—affects up to 50 million people in the U.S. [1], a third of whom experience functional cognitive impairment and diminished quality of life [2], [3]. Clinical guidelines for tinnitus management involve targeted exposure to external sounds as part of sound therapy or cognitive behavioral therapy [4]. Critically, patient outcomes improve when the employed external sounds are closely informed by the patient's internal tinnitus experience [5], [6], [7], [8]. However, existing strategies for characterizing tinnitus sounds, such as pitch matching (PM), are best suited for patients whose tinnitus resembles pure tones (e.g., ringing) [9], [10], [11], [12]. There is a need for methods to characterize tinnitus sounds in the estimated 20–50% of patients with nontonal (e.g., buzzing, roaring) tinnitus [13], [14]. Nontonal tinnitus sounds are presumed to be complex and heterogeneous [12], although few characteristics have been firmly established. Therefore, we base our approach on reverse correlation (RC), an established behavioral method [15], [16], [17] for estimating internal perceptual representations that is unconstrained by prior knowledge about the representations themselves. RC asks participants to render subjective judgments over ambiguous stimuli, and reconstructs the latent representation by regressing subject responses onto the stimuli. RC is closely related to Wiener theory, which has inspired “white-noise” approaches to system characterization in physiology [18], [19] and engineering [20]. Here, we validate RC as a method for characterizing individuals' internal representations of tinnitus more completely. To that end, normal-hearing participants completed an augmented RC experiment, comparing random stimuli to a target tinnitus-like sound, yielding frequency spectrum estimates of their tinnitus representation. The estimated spectra were subsequently validated against the target tinnitus sounds. Our results demonstrate, for the first time, that tinnitus-like sounds with complex spectra can be accurately estimated using RC. II. Materials and Methods A. Stimuli The frequency space of the stimuli was partitioned into \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$b=8$\end{document} Mel-spaced frequency bins, which divide the frequency space between \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$[100,\;13,000]$\end{document} Hz into contiguous segments of equal amplitude (i.e., “rectangular” bins). Reconstruction detail increases with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$b$\end{document}, but \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$b=8$\end{document} provides a good approximation to the chosen target sounds (cf. Fig. 3). Fig. 1. Diagram of the experimental protocol. Subjects listen to a series of random stimuli, each preceded by a target sound. Subjects compare the stimulus to the target, and respond either “yes” or “no” depending on their perceived similarity. The recorded stimulus-response pairs are used to form an estimate of the target. Fig. 2. Human reconstruction accuracy is significantly above baseline, but is not optimal. Random, human, and ideal reconstruction accuracies are shown as violin plots with box plots overlaid. The median is a white dot, the ordinate of the blue points are the Pearson's \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$r$\end{document} values. Fig. 3. Reconstructions of the tinnitus spectra capture many salient features of the target sounds. The black trace indicates the target, while colored traces plot exemplar human reconstructions. The top row shows standardized power levels within each frequency bin. The bottom row maps the 8-dimensional bin domain to a 11025-dimensional frequency domain with the unbinned power spectral density of the targets shown in gray. For each stimulus, [2,7] bins were randomly “filled” with power 0 dB. “Unfilled” bins were assigned \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$-100$\end{document} dB. All frequencies were assigned random phase. Inverse Fourier transform of the constructed spectrum yields a 500 ms stimulus waveform.1 B. Target Sounds Two spectrally complex and complementary target sounds (“buzzing” and “roaring”) were downloaded from the American Tinnitus Association [21] and truncated to 500 ms in duration (power-spectral densities are displayed in the botton subplots of Fig. 3). C. Experiment Ten (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$n=10$\end{document}) normal-hearing subjects listened to A-X trials containing a target sound (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$A$\end{document}) followed by a stimulus (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$X$\end{document}). \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$X$\end{document} was randomly generated for each trial, while \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$A$\end{document} remained the same within a block of 100 trials (Fig. 1). Subjects completed two (2) blocks per target sound (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$p=200$\end{document} total trials per subject). Subjects were told that some stimuli had \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$A$\end{document} embedded in them, and were instructed to respond “yes” to such stimuli, otherwise “no.” Subjects listened over earphones at a self-determined comfortable level. Presentation level was not recorded in this study. Procedures were approved by the UMass IRB. D. Reconstruction A subject performing \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$p$\end{document} RC trials with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$b$\end{document} frequency bins produces a stimulus matrix \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\Psi \in \mathbb {R}^{p \times b}$\end{document} and a response vector \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$y \in \lbrace 1,-1\rbrace ^{p}$\end{document}, where 1 corresponds to a “yes” response and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$-1$\end{document} to a “no.” RC classically assumes the subject response model: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{} \begin{equation*} y = sign(\Psi x), \tag{1} \end{equation*} \end{document} where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$x \in \mathbb {R}^{\mathrm{b}}$\end{document} is the subject's internal representation of interest (i.e., of their tinnitus). Inverting this model yields: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{} \begin{equation*} \hat{x} = \frac{1}{n} \Psi ^{\mathrm{T}} y \tag{2} \end{equation*} \end{document} which is a restricted form of the Normal equation under the assumption that the stimulus dimensions are uncorrelated [16]. E. Validation The experimental paradigm allows for direct validation of the reconstructions \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\hat{x}_{buzzing}$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\hat{x}_{roaring}$\end{document}. We represent the spectra of the target sounds as vectors \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$s_{buzzing} \in \mathbb {R}^{\mathrm{b}}$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$s_{roaring} \in \mathbb {R}^{\mathrm{b}}$\end{document} using the same frequency bins as the stimulus with power equal to the mean power at frequencies within that bin. Pearson's \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$r$\end{document} between \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$s_{buzzing}$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$s_{roaring}$\end{document} and their corresponding reconstructions quantifies reconstruction accuracy. One-sample t-tests were performed on the mean Fisher-transformed Pearson's \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$r$\end{document} values across subjects to assess significant differences from zero. F. Synthetic Subjects To establish bounds on human performance, additional experiments were run with two simulated subjects who give either ideal or random responses. Each experiment ran for \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$p=200$\end{document} trials and was repeated 1000 times. The ideal subject gives responses following: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{} \begin{equation*} y_{i} = {\begin{cases}1 & \text{if } \Psi _{i} s \geq Q(0.5; \Psi s) \\ -1 & \text{otherwise} \end{cases}} \tag{3} \end{equation*} \end{document} for \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$i \in 1,{\ldots }, p$\end{document}, where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$Q(x, y)$\end{document} is the quantile function for \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$x \in [0, 1]$\end{document} of the similarity calculation \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\Psi s$\end{document}, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\Psi _{i}$\end{document} is the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$i\text{th}$\end{document} column of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\Psi$\end{document}. Thus, the ideal subject has precise knowledge of every stimulus and responds according to (3). The random subject responds \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$y_{i} \in \lbrace 1,-1\rbrace$\end{document} with uniform random probability, thus ignoring the stimulus entirely. III. Results Fig. 2 shows the distribution of Pearson's \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$r$\end{document} for human, ideal, and random subject responses. Human accuracy is statistically significantly higher than random chance and for some subjects, approaches the ideal case. Accuracy from the random subject was \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$0.00 \pm 0.44$\end{document} (mean \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$\pm$\end{document} st.dev.) for buzzing and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$0.00 \pm 0.39$\end{document} for roaring, while mean accuracy from human responses was significantly different from 0 in all conditions: buzzing: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$0.52 \pm 0.27$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$t(9) = 5.766$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$p < 0.001$\end{document}; roaring: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$0.62 \pm 0.23$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$t(9) = 5.76$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$p < 0.001$\end{document}; combined: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$0.57 \pm 0.25$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$t(19) = 7.542$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$p < 0.001$\end{document}. From Fig. 2, it appears that the distribution of buzzing results differs from that of the roaring results, however the difference between buzzing and roaring is not statistically significant (two-way ANOVA across subjects (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$F(13) = 2.94$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$p > 0.05$\end{document}) and target signals (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$F(1) = 2.44$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}$p > 0.05$\end{document})). Fig. 3 plots the most accurate human reconstructions over the target sound spectra. IV. Conclusion Our results show that RC can accurately reconstruct the frequency spectrum of tinnitus-like sounds relevant to non-tonal tinnitus, and therefore represent a proof of concept for using RC to characterize non-tonal tinnitus. Subjects completed the required number of trials within ten minutes, indicating that this procedure could be conducted within a single clinical visit. RC may therefore be useful as the basis for a clinical assay to characterize a wider variety of tinnitus percepts than currently possible. Reconstruction accuracies observed here are below the simulated ideal, which may be attributed to noisy responses universally observed in applications of RC, and which may be mitigated by further optimizing the experimental protocol, stimulus generation, and reconstruction method. For example, recent approaches to improving RC reconstruction methods can boost efficiency, noise robustness and overall accuracy [22]. Future work will focus on more comprehensive validation of this approach, using larger sample sizes, more target sounds, and stricter control of sound presentation level. Acknowledgment The authors declare no competing interests. 1 Software for the experiments and analysis was written in MATLAB and is freely available at https://github.com/alec-hoyland/tinnitus-reconstruction/ ==== Refs References [1] J. M. Bhatt, H. W. Lin, and N. Bhattacharyya, “Prevalence, severity, exposures, and treatment patterns of tinnitus in the United States,” JAMA Otolaryngol.–Head Neck Surg., vol. 142 , no. 10 , pp. 959–965, Oct. 2016.27441392 [2] D. M. Nondahl , “The impact of tinnitus on quality of life in older adults,” J. Amer. Acad. Audiol., vol. 18 , no. 3 , pp. 257–266, Mar. 2007.17479618 [3] S. Tegg-Quinn, R. J. Bennett, R. H. Eikelboom, and D. M. Baguley, “The impact of tinnitus upon cognition in adults: A systematic review,” Int. J. Audiol., vol. 55 , no. 10 , pp. 533–540, Oct. 2016.27240696 [4] D. E. Tunkel , “Clinical practice guideline: Tinnitus,” Otolaryngol.–Head Neck Surg., vol. 151 , no. 2_suppl , pp. S1–S40, Oct. 2014.25273878 [5] H. Okamoto, H. Stracke, W. Stoll, and C. Pantev, “Listening to tailor-made notched music reduces tinnitus loudness and tinnitus-related auditory cortex activity,” Proc. Nat. Acad. Sci., vol. 107 , no. 3 , pp. 1207–1210, Jan. 2010.20080545 [6] R. Schaette, O. König, D. Hornig, M. Gross, and R. Kempter, “Acoustic stimulation treatments against tinnitus could be most effective when tinnitus pitch is within the stimulated frequency range,” Hear. Res., vol. 269 , no. 1 , pp. 95–101, Oct. 2010.20619332 [7] A. Stein , “Inhibition-induced plasticity in tinnitus patients after repetitive exposure to tailor-made notched music,” Clin. Neuriophysiol., vol. 126 , no. 5 , pp. 1007–1015, May 2015. [8] S. Schoisswohl, J. Arnds, M. Schecklmann, B. Langguth, W. Schlee, and P. Neff, “Amplitude modulated noise for tinnitus suppression in tonal and noise-like tinnitus,” Audiol. Neurotol., vol. 24 , no. 6 , pp. 309–321, 2019. [9] P. Neff, B. Langguth, M. Schecklmann, R. Hannemann, and W. Schlee, “Comparing three established methods for tinnitus pitch matching with respect to reliability, matching duration, and subjective satisfaction,” Trends Hear., vol. 23 , 2019, Art. no. 2331216519887247. [10] J. A. Henry, K. E. James, K. Owens, T. Zaugg, E. Porsov, and G. Silaski, “Auditory test result characteristics of subjects with and without tinnitus,” J. Rehabil. Res. Develop., vol. 46 , no. 5 , pp. 619–632, 2009. [11] O. C. Ukaegbe, F. T. Orji, B. C. Ezeanolue, J. O. Akpeh, and I. A. Okorafor, “Tinnitus and its effect on the quality of life of sufferers: A Nigerian cohort study,” Otolaryngol.–Head Neck Surg., vol. 157 , no. 4 , pp. 690–695, 2017.28695761 [12] D. Vajsakovic, M. Maslin, and G. D. Searchfield, “Principles and methods for psychoacoustic evaluation of tinnitus,” in The Behavioral Neuroscience of Tinnitus. Berlin, Germany: Springer, Feb. 2021, pp. 419–451. [13] J. A. Henry, “Measurement” of tinnitus,” Otol. Neurotol., vol. 37 , no. 8 , Sep. 2016, Art. no. e276. [14] A. Noreña, C. Micheyl, S. Chéry-Croze, and L. Collet, “Psychoacoustic characterization of the tinnitus spectrum: Implications for the underlying mechanisms of tinnitus,” Audiol. Neuro- otol., vol. 7 , pp. 358–69, Nov. 2002. [15] A. Ahumada and J. Lovell, “Stimulus features in signal detection,” J. Acoustical Soc. Amer., vol. 49 , no. 6B , pp. 1751–1756, Jun. 1971. [16] F. Gosselin and P. G. Schyns, “Superstitious perceptions reveal properties of internal representations,” Psychol. Sci., vol. 14 , no. 5 , pp. 505–509, Sep. 2003.12930484 [17] W. O. Brimijoin, M. A. Akeroyd, E. Tilbury, and B. Porr, “The internal representation of vowel spectra investigated using behavioral response-triggered averaging,” J. Acoustical Soc. Amer., vol. 133 , no. 2 , pp. EL118–EL122, Feb. 2013. [18] D. Ringach and R. Shapley, “Reverse correlation in neurophysiology,” Cogn. Sci., vol. 28 , no. 2 , pp. 147–166, 2004. [19] P. Z. Marmarelis and V. Z. Marmarelis, “The white-noise method in system identification,” in Analysis of Physiological Systems: The White-Noise Approach (Computers in Biology and Medicine Series), P. Z. Marmarelis and V. Z. Marmarelis, Eds. Boston, MA, USA: Springer, 1978, pp. 131–180. [20] L. Ljung, System Identification: Theory for the User. London, U.K.: Pearson Education, Dec. 1998. [21] “Listen to sample tinnitus sounds (ATA),” Jun. 14, 2022. [Online]. Available: https://www.ata.org/listen-sample-tinnitus-sounds [22] A. Compton, B. W. Roop, B. Parrell, and A. C. Lammert, “Stimulus whitening improves the efficiency of reverse correlation,” Behav. Res. Methods, to be published, doi: 10.3758/s13428-022-01946-w.

PMC010xxxxxx/PMC10277748.txt

==== Front J Nematol J Nematol jofnem jofnem Journal of Nematology 0022-300X 2640-396X Sciendo 37342200 jofnem-2023-0017 10.2478/jofnem-2023-0017 Article Morphological and Molecular Characterization of Diplogasteroides sp., a Cryptic Population of the Haslacheri Group (Diplogastridae), and Parasitorhabditis terebranus (Rhabditidae) from Korea Mwamula Abraham Okki 12 Lee Sang Myeong 3 Jung Young Hak 3 Lee Ho-wook 4 Kim Yi Seul 12 Kim Young Ho 12 Lee Dong Woon whitegrub@knu.ac.kr 124 1 Department of Entomology, Kyungpook National University, Sangju, 37224, Republic of Korea 2 Research Institute of Invertebrate Vector, Kyungpook National University, Sangju 37224, Republic of Korea 3 SM Biovision Co., Jinju, 52849, Republic of Korea 4 Department of Ecological Science, Kyungpook National University, Sangju, 37224, Republic of Korea 2 2023 23 5 2023 55 1 2023001729 1 2023 © 2023 Abraham Okki Mwamula et al., published by Sciendo 2023 Abraham Okki Mwamula et al., published by Sciendo https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. Abstract Diplogasteroides sp., a cryptic population of D. haslacheri, and Parasitorhabditis terebranus were reported from the frass of Monochamus alternatus galleries in dead Pinus thunbergii for the first time in Korea. Females and males are morphologically characterized and their linked DNA barcodes (18S-rRNA, 28S-rRNA, ITS-rRNA and COI) supplied. Females and males of the two species from Korea conform to the original species descriptions from Europe and the USA, with variations in a few details in morphometrics. Specifically, Diplogasteroides sp. is morphologically very similar to D. haslacheri. However, it cannot be designated as D. haslacheri due to the existence of cryptic species complex within the haslacheri group (D. haslacheri, D. asiaticus, D. nix, D. andrassyi, and D. carinthiacus), a condition requiring hybridization studies to test species identity within the group. Based on analysis of COI sequences, differences among these cryptic species are evident. Thus, in addition to hybridization tests, the COI might be a powerful DNA barcoding marker for the precise identification of these cryptic species within the genus. Additionally, this is the first molecular characterization of P. terebranus, and the species is herein recorded for the first time outside its type locality. Keywords Diplogasteroides morphology Parasitorhabditis terebranus taxonomy ==== Body pmcXylophagous insects, mainly the longhorn beetles (Cerambycidae), bark beetles from subfamily Scolytinae (family Curculionidae), and jewel beetles (Buprestidae) represent the most economically important agents of tree decline in pine ecosystems (Wood, 1982; Sauvard, 2007; Meshkova et al., 2018; Andreieva et al., 2020). For instance, the genus Monochamus (Cerambycidae) is best known for its fundamental role in the epidemiology of pine wilt disease (Mamiya, 1983). Specific species of the genus are responsible for the dissemination of the pinewood nematode, Bursaphelenchus xylophilus (Togashi, 2008). In Korea, for instance, M. alternatus and M. saltuarius are the main vectors of the pine wood nematode, and the former is mainly distributed in the southern region and the latter in the central region (Kwon et al., 2006). Additionally, Monochamus beetles are considered secondary forest pests that mainly attack weakened or dying trees (Cesari et al., 2005). Apart from this significant phoretic association, several entomophilic and entomogenous nematodes are closely associated with various xylophagous insects (Nickle, 1973; Massey, 1974; Nguyen and Hunt, 2007; Susoy et al., 2013). Their relationships with insects can be categorized into different groups such as saprobiotism, commensalism, phoresy, and various levels of parasitism (Tenkáčová and Mituch, 1987; Grucmanová and Holuša, 2013). The latter group is of importance in limiting the populations of bark beetles in forest ecosystems, and are therefore potential biological control agents (Nguyen and Hunt, 2007; Grucmanová and Holuša, 2013). The family Diplogastridae Micoletzky, 1922, constitutes morphologically diverse genera whose feeding habits are highly variable (Kanzaki et al., 2016). Many species of Diplogastridae are associated with xylophagous insects; and the types of associations are known to vary among genera (Poinar, 1975). For example, many members of Pristionchus are known to have necromenic association with scarab beetles (Herrmann et al., 2006); Parasitodiplogaster spp. are parasites of fig wasps (Giblin-Davis et al., 2006); and members of Diplogasteroides de Man, 1912 and Micoletzkya Weingärtner, 1955 are commonly associated with bark beetles (Coleoptera: Scolytidae) (Rühm, 1965; Massey, 1974). More precisely, species of the genus Diplogasteroides de Man, 1912 are known to have a phoretic relationship with bark beetles, especially species of Dendroctonus Erichson, 1836 and Ips De Geer, 1775 (Massey, 1974). Several species have also been isolated from Monochamus beetles: for example, D. andrassyi from M. grandis (Kanzaki et al., 2013), D. asiaticus from M. alternatus (Kanzaki et al., 2015), and D. nix from M. urussovii (Kanzaki et al., 2016). The genus is characterized by a monomorphic stoma with three small, rod-like dorsal teeth and nine pairs of genital papillae in males. However, some other characters are significantly divergent among species, and the genus is considered to be a paraphyletic taxon (Sudhaus and Fürst von Lieven, 2003). On the other hand, Parasitorhabditis, a genus that belongs to the family Rhabditidae Orley, 1880, is also constituted by species that form parasitic or phoretic relationships with bark beetles, especially of the family Cerambycidae and Scolytidae (Poinar, 1975). Species of the genus are generally known to be associated with insect digestive tract and the frass in the beetle galleries (Grucmanová and Holuša, 2013). They are generally benign to their hosts, but there have been some reports of harm involving both cerambycid (Chapman, 1964) and scolytid beetles (Tomalak et al., 1989). Parasitorhabditis species are associated with beetles of the genera Dendroctonus Erichson, 1836; Ips De Geer, 1775; Dryocoetes Eichhoff, 1864; Hylurgops LeConte, 1876; Hylastes Erichson, 1836; Hylurgus Latreille, 1807; Pityogenes Bedel, 1888; Crypturgus Erichson, 1836; Cryphalus Erichson, 1836; Phloeosinus Chapuis, 1869; Polygraphus Erichson, 1836; Scolytus Geoffroy, 1762 (Massey, 1974), and some members of the family Cerambycidae (Sudhaus, 1974). Species of Parasitorhabditis and Diplogasteroides are taxonomically confounded, with no clear apomorphy for the genus (in Diplogasteroides) and with phenotypic characters overlapping with other closely related genera (Sudhaus and Fitch, 2001; Sudhaus and Fürst von Lieven, 2003). Integrative taxonomy considering both morphological characters and DNA-based inferences provide a better supported approach in nematode identification (Subbotin et al., 2005; Mwamula et al., 2022). However, entomophilic nematodes have not been given the necessary attention, and DNA sequence data of many of these species are still unavailable. During a survey in the pine forest ecosystem in Korea, a diplogastrid and rhabditid nematode were recovered from the frass in beetle galleries in dead pine, and identified as Diplogasteroides sp. (a cryptic population of D. haslacheri) and P. terebranus. The two species are herein characterized using morphological and molecular techniques. Materials and Methods Nematode populations and extraction Nematode populations were extracted from the frass retrieved from M. alternatus galleries in dead Pinus thunbergii wood taken from Gonri Island in Tongyeong, Gyeongsangnam-do Province, Korea. Nematodes were extracted from the frass samples using the Baermann funnel method (Baermann, 1917). The collected nematode suspension was examined under a Nikon SMZ 1000 stereomicroscope (Nikon) and specimens belonging to Diplogasteroides and Parasitorhabditis were picked out and subsequently characterized to species level based on inferences from morphometrics and DNA sequence data. Morphological characterization Twenty female and 20 male specimens of each of the two species were heat-killed, fixed, and mounted to pure glycerin (Seinhorst, 1959). Morphometrics and photomicrographs were taken using a Zeiss imager Z2 microscope (Carl Zeiss) fitted with Axio-vision Material Science Software for Research and Engineering software (Carl Zeiss). Schematic illustrations were made under a drawing tube before being redrawn using CorelDRAW® software version 24. Species delineation was done following original species descriptions, and the key presented by Massey (1974). Morphometric data were also compared with the closest species from other species descriptions (see Carta et al., 2010; Kanzaki et al., 2015, 2016). Molecular characterization DNA was extracted from single female specimens using a DNA extraction kit (WizPure™) as detailed by Iwahori et al. (2000). The nearly full-length 18S-rRNA gene was amplified as two partially overlapping fragments using two primer sets: 988F (5′-CTCAAAGATTAAGCCATGC-3′) and 1912R (5′-TTTACGGTCAGAACTAGGG-3′), 1813F (5′-CTGCGTGAGAGGTGAAAT-3′) and 2646R (5′-GCTACCTTGTTACGACTTTT-3′) (Holterman et al., 2006); D2A (5′-ACAAGTACCGTGAGGGAAAGTTG-3′) and D3B (5′-TCGGAAGGAACCAGCTACTA-3′) (Nunn, 1992) amplified the D2-D3 expansion segment of 28S-rRNA; COI-F1 (5′-CCTACTATGATTGGTGGTTTTGGTAATTG-3′) and COI-R2 (5′-GTAGCAGCAGTAAAATAAGCACG-3′) primer set (Kanzaki and Futai, 2002) amplified partial COI gene, and TW81 (5′-GTTTCCGTAGGTGAACCTGC-3′) and AB28 (5′-ATATGCTTAAGTTCAGCGGGT-3′) (Curran et al., 1994) amplified the ITS-rRNA gene of P. terebranus. Polymerase chain reaction (PCR) was performed with a PCR cycler (T100™, Bio-Rad), the PCR program being set as follows: initial denaturation at 95°C for 5 min, 35 cycles at 95°C for 30 s, followed by an annealing step at 53°C for 30 s (D2A/D3B, COI-F1/COI-R2 and TW81/AB28 primer sets) and 52°C for 30 s (988F/1912R and 1813F/2646R primer sets); 72°C for 1 min (D2A/D3B, COIF/COIR and TW81/AB28 primer sets) and 72°C for 80 s (988F/1912R and 1813F/2646R primer sets); and finally one cycle at 72°C for 5 min (D2A/D3B, COI-F1/COI-R2 and TW81/AB28 primer sets) and 72°C for 10 min ((988F/1912R and 1813F/2646R primer sets). The PCR products were purified using a PCR purification kit (Qiagen), and subsequently quantified using a quick-drop spectrophotometer (Molecular Devices). The purified products were used for direct sequencing in both directions using the same primers as specified above. DNA sequencing was performed at Macrogen. The obtained and edited DNA sequences were submitted to the GenBank database under the accession numbers OQ704207–OQ704210 (18S-rRNA), OQ291287–OQ291290 (28S-rRNA), OQ305560–OQ305564 (ITS-rRNA) and OQ281740–OQ281743 (COI gene). Phylogenetic analysis The newly obtained sequences (18S-rRNA, D2-D3 expansion segment of 28S-rRNA, ITS-rRNA and COI gene) of Diplogasteroides sp. and P. terebranus were aligned using ClustalX 1.83 (Thompson et al., 1997) along with the corresponding comparison sequence data sets of other close species within the genera Diplogasteroides and Parasitorhabditis, and species of other closely related genera published in GenBank (Kanzaki et al., 2013, 2015; Valizadeh et al., 2017; Bhat et al., 2020; Girgan et al., 2021). Sequences from Rhabditoides inermiformis and Leptolaimus donsi were used as outgroup taxa for constructing the phylogenetic tree for the 18S-rRNA gene. Rhabditoides regina and Ascaridia columbae were used as outgroup taxa for phylogenetic inferences of D2-D3 expansion segment of 28S-rRNA and COI gene, respectively. The generated alignments were analyzed with Bayesian inference (BI) using MrBayes 3.2.6 (Ronquist et al., 2012). The general time reversible substitution model with estimation of invariant sites and assuming a gamma distribution with four categories (GTR + I + G) was chosen as the appropriate nucleotide substitution model for the three analyses. Bayesian inference (BI) analysis for each gene was initiated with a random starting tree, and run with four chains for 1 × 106 generations. Posterior probabilities were estimated using the Markov chain Monte Carlo (MCMC) method, and consensus trees were generated with a 50% majority rule. The generated trees were subsequently visualized and edited using FigTree v1.4.4 software. Intraspecific and interspecific sequence variations were checked using PAUP* 4.0a169 (Swofford, 2003). Results Diplogasteroides sp. (Figs. 1 & 2) Figure 1: Photomicrographs of Diplogasteroides sp. (A-C, F, and G) and Parasitorhabditis terebranus (D, E, and H-J). Diplogasteroides sp. A, B: Variation in female anterior body region; C: Female tail; F and G: Variation in male tail and copulatory apparatus shape. Parasitorhabditis terebranus D, E, and H: Variation in female anterior body region; I: Male tail with copulatory apparatus; J: Female tail; (Scale bars: A-J = 10 μm). Figure 2: Illustrations of Diplogasteroides sp. (A-H): A: Female whole body; B: Male whole body; C: Male tail region including copulatory apparatus and the arrangement of genital papillae; D-F: Variation in shape of gubernaculum; G: Female tail region; H: Female anterior region, including the shape of pharynx. Abbreviations: Genital papillae arrangement from ventral side (v + number); lateral/dorsal sides (v3d; ad; pd); phasmid (ph); and deirid (d). Measurements See Table 1. Table 1. Comparison of morphometries of Diplogasteroides sp. and Parasitorhabditis terebranus from Korea with topotype populations.a Character Diplogasteroides sp. D. haslacheri Parasitorhabditis terebranus Korea (Current study) Rühm (1956), and Kanzaki et al. (2002) Korea (Current study) USA (Massey, 1974) ♀♀ ♂♂ ♀♀b ♂♂ ♀♀ ♂♂ ♀♀ ♂♂ N 20 20 ? ? 20 20 ? ? L 1061,3±106.9 (855.3-1258.3) 909.O±82.8 (777.6-1068.1) 825-1005 750-825 109δ.4±78.6 (922.4-1244.4) 1015.8±96.9 (841.0-1195.6) 770-810 750 A 29.4±2.1 (25.1-32.3) 29.2±0.8 (27.7-30.5) 26.1 26.2-26.8 26.4±2.9 (20.5-30.0) 26.0±2.2 (21.1-29.4) 19.7-20.1 19.5 B 6.2±0.5 (5.4-7.5) 5.8±0.4 (5.2-6.3) 5.6-6.1 5.1-5.2 5.5±0.4 (4.6-6.1) 5.2±0.4 (4.6-6.1) 4.2-4.3 4.1 C 11.7±0.8 (9.8-13.1) I4.9±0.9 (13.2-16.2) 9.4-9.6 12.4-15.3 27.1±2.2 (22.8-30.2) 27.4±2.3 (22.1-30.5) 26.2-27.3 27.3 c’ 4.2±0.3 (3.5-4.7) 2.5±0.2 (2.2-2.9) 5.0 2.4 1.9±0.2 (1.6-2.1) 1.7±0.1 (1.4-1.9) - - V 51.6±1.1 (50.3-53.2) - 50.3-52.3 - 93.3±0.5 (91.9-93.9) - 93 - Lip height - - - - 3.3±0.1 (3.0-3.6) 3.1 ±0.2 (2.8-3.5) - - Lip diameter 11.9±0.5 (10.6-12.6) 11.9±0.6 (11.0-12.6) - - 12.6±0.7 (11.6-14.0) 12.1 ±0.7 (11.0-13.5) - - Stoma 11.1±0.8 (9.8-12.3) 11.0±0.7 (9.0-11.8) - - 20.9±1.4 (18.5-23.0) 20.3±1.1 (18.0-22.0) 21 - Anterior to median bulb valve 98.2±6.3 (87.6-112.7) 90.4±5.3 (82.4-102.2) - - 107.2±7.4 (95.8-120.2) 101,4±5.1 (93.3-112.9) - - Excretory pore 1340±8.3 (123.0-155.4) 126.8±5.7 (114.8-138.0) - - 1540±8.7 (135.0-162.7) 153.6±8.6 (136.0-164.0) - - Pharynx length 171.4±11.9 (154.1-194.2) 156.3±8.6 (145.0-170.0) - - 201.0±9.3 (191.3-223.0) 194.3±8.1 (181.3-208.5) - - Maximum body diameter 36.2±3.5 (29.0-42.1) 31.7±3.0 (27.4-37.7) - - 41.9±4.7 (36.0-51.5) 39.1±2.7 (33.3-43.0) - - Vulval body diameter 38.1±3.6 (33.6-43.6) - - - 35.2±2.9 (30.8-42.6) - - - Vulva to anus length - - - - 32.5±2.6 (28.0-38.0) - - - Vulva to tail tip - - - - 73.1±3.5 (64.0-77.9) - - - Anal / cloacal body diameter 21.5±1.4 (19.2-24.8) 24.2±1.6 (22.0-27.0) - - 21.4±1.5 (18.9-24.5) 21.8±1.6 (19.8-24.9) - - Tail length 90.5±9.5 (73.0-104.6) 60.8±4.6 (50.0-69.5) - - 40.6±3.0 (36.0-45.8) 37.0±2.5 (31.5-41.0) - - Spicules - 35.5±2.1 (30.0-38.0) - 27-30 - 36.6±1.7 (33.0-38.6) - 34 Gubernaculum - 17.9±1.5 (15.7-21.0) - 15-17 - 16.5±0.6 (15.2-18.0) - 17 Tail spike length - 19.1 ±2.3 (16.0-23.4) - - - - - - aAII measurements are in micrometers and in the form mean ± standard deviation (range). bDip¦ogasteroides has¦acheri and Parasitorhabditis terebranus topotype populations. Description Female (n = 20) General habitus agreeing with descriptions and morphometrics of D. haslacheri (Fuchs, 1931; Rühm, 1956; Kanzaki et al., 2002, 2013, 2015). Body cylindrical, moderate to slender, tapering anteriorly and posteriorly. Cuticle moderate in thickness, finely to moderately annulated. Lip region not offset, weakly separated into six sectors; each possessing a setiform labial sensilla. Stoma tube-like, separated into cheilostom, gymnostom, and stegostom sections. Cheilostom short, with thickened and ring-like anterior part and short tube-like posterior part. Gymnostom well cuticularized, tube-like, almost twice as long as the cheilostom. Stegostom separated into pro-meso-, meta-, and telostegostom subsections. Metastegostom bearing three small dorsal teeth. Procorpus muscular, occupying ca 50% of the corresponding body diameter. Metacorpus well differentiated, with a muscular median bulb. Anterior part of pharynx (pro- and metacorpus) little longer than the posterior part of pharynx (isthmus and basal bulb). Nerve ring surrounding the mid- or slightly posterior part of the isthmus. Excretory pore location variable; around the posterior part of the isthmus or anterior end of the basal bulb. Deirid small, located posterior to excretory pore, visible around the base of cardia. Postdeirid not observed. Lateral field unclear, very weakly discernible. Gonads paired. Ovaries reflexed along their entire length and identical as illustrated (Fig. 2A). Each branch basically arranged from ovary to a gonoduct (oviduct connected to a wider proximal part, the spermatheca and uterus) and vagina. Oviduct tube-like. Spermatheca long, appearing oblong in shape, sometimes filled with well-developed sperm. Uterus connected to the vagina. Vagina appear perpendicular to body surface, with a constriction muscle, the sphincter, and dilator muscles at its junction with the uterus. Vulva porelike, with protuberant lips in lateral view. Rectum ca one anal body diameter in length. Three rectal glands visible at prerectum (intestine-rectum junction). Phasmid visible, laterally located ca 23 to 28 μm posterior to anal opening, or ca 1.2 to 1.4 anal body diameter posterior to anal opening. Tail elongated, conical, smoothly tapering to a rounded or finely pointed tip. Male (n = 20) Generally similar to female in general morphology except for sexual characters. Testis single, anterior end straight or reflexed; the reflexed part occupying ca 5% to 10% of total genital tract length. Vas deferens tube-like, joining the rectum and forming a narrow cloacal tube terminating into the cloaca. Spicules paired, separate; manubrium rounded, separated by a weak constriction; calomus-lamina (shaft-blade) complex clearly bent at one-third of its length from the manubrium, posterior part narrowing to a bluntly pointed straight or slightly ventrally curved tip. Gubernaculum ca one-half of the total spicule in length, water drop-shaped, possessing a pointed distal end in lateral view. Tail conical, smoothly tapering and possessing a spike at the tail end. Nine pairs of genital papillae and an additional small papilla on the precloacal lip present. First subventral pair (v1) located ca 1 to 1.3 cloacal body diameter anterior to the cloacal opening; second subventral (v2) pair located close to just anterior to the cloacal opening; third lateral (v3d) located adjacent or just posterior to the cloacal opening; fourth subventral (v4) pair located close to the third just posterior to the cloacal opening; fifth lateral (ad) pair around the midpoint between the cloacal opening and the base part of the tail spike; sixth to eighth (subventral v5-7) pairs located close to each other, anterior to the base of the tail spike, ninth dorsal (pd) pair almost on same level with the eighth pair (subventral v7). Phasmids located between the fifth lateral (ad) pair and the v5-7 set, ca 12 to 16 μm from the base of the tail spike. Tail spike length less than cloacal body diameter. Tail spike also shorter than the distance from the cloaca to base of the spike. Remarks The current morphometric data of the currently studied population agree well with the descriptions of D. haslacheri by Fuchs (1931), Rühm (1956), and the comparative studies of Kanzaki et al. (2002, 2013, 2015), albeit the relatively long body (855.0 to 1258.0 versus 825 to 1005 μm); and relatively longer spicules (30.0 to 38.0 versus 27 to 30 μm). However, as noted by Kanzaki et al. (2013), quantitative morphometric characters are known to vary within the same species depending on culture/host conditions. Qualitative characters are more reliable for morphological comparisons within the genus as suggested by Andrássy (1984) and Sudhaus and Fürst von Lieven (2003). All the qualitative characters in the current population are evidently very similar to D. haslacheri descriptions. Despite the indistinguishable qualitative characters, the current population cannot be designated as D. haslacheri. The presence of a cryptic species complex within the haslacheri group requires hybridization studies to allow more conclusive tests of species identity (Kanzaki et al., 2013, 2015), a process that is currently not feasible due to unavailability of an authentic population of D. haslacheri. Additionally, D. haslacheri was previously described from broadleaf trees infected by some bark beetles (i.e., frass of mainly Scolytus mali, Scolytidae) and cerambycid beetles (Cerambyx scopolii and Leiopus neburosus) in Germany (Fuchs, 1931; Rühm, 1956). In contrast, the population described herein was associated with a different isolation source, the frass of M. alternatus galleries in dead Pinus thunbergii. It is therefore treated here as a cryptic species of D. haslacheri group until hybridization tests with D. haslacheri are feasible. Parasitorhabditis terebranus (Massey, 1974) (Figs. 1 and 3) Figure 3: Illustrations of Parasitorhabditis terebranus (A-G): A: Male whole body; B: Female whole body; C: Female anterior region; D and F: Female tail region including variation in vulval lip shape; E: Male tail in right lateral view, with arrangement of bursal rays; G: Male tail in ventral view, with arrangement of bursal rays. Measurements See Table 1. Description Female (n = 20) General habitus agreeing with description of Massey (1974), with variation in a few details in morphology and morphometrics. Habitus generally straight to ventrally arcuate when heat-killed. Cuticle fine to moderately annulated. Lips angular to rounded, ca 3.5 μm high and 12.5 μm wide, with moderately prominent papillae. Stoma 21.0 (18.5 to 23.0) μm in depth. Anterior tips of prorhabdions bent in most individuals. Remnants of metarhabdions with two visible subventral teeth, and two subdorsal teeth, visible in lateral view. Esophagus muscular throughout, without median bulb. Lumen of corpus heavily sclerotized with transverse ridging. Procorpus equal or slightly less in length to isthmus and basal bulb. Nerve ring at mid- or slightly posterior to midisthmus. Excretory pore obscure, slightly anterior to, or at level of nerve ring. Lateral field hardly discernible. Ovary single, anteriorly directed, reflexed ca one-third to half its length. Oocytes arranged in one to three rows in distal part of ovary, and well-developed oocytes arranged in a single row in proximal part. Oviduct tube-like and relatively long, connected to the wider proximal part, the uterus. Uterus often with one or multiple embryos irregularly arranged; length of uterus varying according to presence or absence of embryo. Vagina short, oblique with muscular walls. Lips of vulva mostly protuberate (not protuberate in three of observed specimen). Vulva a transverse slit; vulval-anal distance ca 28.0 to 38.0 μm, ca equal to or less than vulval body diameter, distinctly less than tail length. Rectum ca equal to anal body diameter. Phasmid indistinct under light microscope. Tail length ca twice anal body diameter, conoid to cupola-shaped, ending in spike with a finely rounded tip. Male (n = 20) Generally similar to female in general morphology except for sexual characters. Lips angular to rounded, ca 3.0 μm high and 12.0 μm wide, with moderately prominent papillae. Stoma as in female in depth. Testis single, reflexed, ventrally or dorsally extending up to two-thirds of body length. Vas deferens tube-like, joining the rectum and forming a narrow cloacal tube terminating into the cloaca. Spicules slender, nearly straight to minimally curved, with short distal fusion. Spicule tip bluntly rounded, not curved. Gubernaculum slender, slipper-shaped, ca half or less than spicule length. Tail peloderan, conoid to an acute terminus. Bursa peloderan. Bursal fan peloderan with 10 pairs of bursal rays, with 2 + (3+1) + 4 typical ray pattern (Figs. 1I, 3E, and 3G). First two bursal rays (papillae), located precloacally, with the first of the two reaching the edge of the bursa. Postcloacally, three (third to fifth) pairs positioned close together, located just after cloacal opening, with the third and fourth appearing fused in lateral view. The sixth bursal papilla appear much shorter and uniquely thicker than all others, located singly, closer to the preceding trio; the seventh, eighth, and ninth rays ending just slightly short of bursal edge, with the ninth appearing a little longer than the other two. The last pair (10th pair) also appear shortened than the preceding pairs. Remarks The morphology and morphometrics of the studied population agree well with the descriptions of Massey (1974) except for the relatively long body (922.0 to 1244.0 versus 770 to 810 μm). Also, nine bursal rays were recorded in the original species description. Indeed, in the lateral view, the three (third through fifth) pairs are positioned very close to each other; the third and fourth bursal rays often appearing more fused together (Figs. 1I and 3E), giving the impression of nine bursal rays in some specimens. However, 10 bursal rays are evident, and more discernible in ventral view (Fig. 3G). It is also important to note that the sixth bursal papilla (much shorter and uniquely thicker than all others) represents the phasmids (see Kiontke and Sudhaus, 2000; Sudhaus and Fitch, 2001). Massey (1974) listed diagnostic characters as the tail terminus shape, procorpus being equal in length to isthmus and basal bulb, and the unique prorhabdions as the main diagnostic characters of the species. These characters are evidently similar in the current population. The species has its type locality in Nacogdoches, Texas; with the black turpentine beetle, Dendroctonus terebrans in loblolly pine as the type habitat. The species is herein characterized morphologically and molecularly for the first time out of its type locality. Molecular characterization and phylogenetic relationships The two partially overlapping fragments of 18S-rRNA gene from each species yielded amplicons of ca 1650 bp in length, and amplification of the partial D2-D3 expansion segment of 28S-rRNA, partial ITS-rRNA and the partial CO1 genes yielded single amplicons of ca 650 to 750 bp. There was no intraspecific variation (0.0%) within the two newly obtained 18S-rRNA partial sequences of Diplogasteroides sp. (OQ704207 and OQ704208), and the top hits in the GenBank BLAST homology search for these two sequences (OQ704207 and OQ704208) comprised the member species of the cryptic group; D. andrassyi (AB808722), D. nix (LC145091), D. asiaticus (LC027672), and D. luxuriosae (LC099973). The new sequences differed from D. nix (LC145091) and D. asiaticus (LC027672) by 2 bp (0.1%); from D. andrassyi (AB808722) by 3 bp (0.1%) and from D. luxuriosae (LC099973) by 8 bp (0.4%). The two 18S-rRNA sequences of P. terebranus (OQ704209 and OQ704210) also showed no intraspecific variation (0.0%), and differed from the closely clustered species: P. obtusa isolate BRS_SSU18A (KJ705089), Parasitorhabditis sp. (AF083028) and another P. obtusa population (EU003189) by 4 bp (0.5%), 8 bp (0.5%) and 9 bp (0.6%), respectively. Fifty-one partial and nearly full-length 18S-rRNA gene sequences, including the four newly obtained sequences from this study and sequences of species from other closely related genera published in GenBank comprised the data set for phylogenetic analysis. Phylogenetic relationships as inferred from Bayesian analysis of the 18S-rRNA gene sequence data set with GTR + I + G substitution model are shown in Fig. 4. Figure 4: Bayesian tree inferred under the GTR + I + G model from I8S-rRNA sequences of Diplogasteroides spp., Parasitorhabditis spp., and other closely related species from other genera. Posterior probability values exceeding 50% are given on appropriate clades. The studied population is indicated in bold. Outgroup taxon: Rhabditoides inermiformis and Leptolaimus donsi. The top 28S-rRNA gene BLASTN hits for the two sequences of Diplogasteroides sp. (OQ291287 and OQ291288) included D. asiaticus, D. nix, D. andrassyi, and D. luxuriosae 28S-rRNA gene sequences, all with identities of 98% to 99%. There was no intraspecific variation (0.0%) within the two newly obtained sequences of Diplogasteroides sp. The newly obtained sequences differed from D. asiaticus (MN736552) D. nix (LC145090), and D. andrassyi (AB808723) 28S-rRNA gene sequences by 5 bp (0.7%) and from D. luxuriosae (LC099975) by 9 bp (1.2%). The top 28S-rRNA gene BLASTN hits for the sequences of P. terebranus (OQ291289 and OQ291290) included the four sequences of P. obtusa (MG865784, MF288651, EF990724, and KM245037) with identities of 96% to 97%, differing from MG865784 and MF288651 by 14 bp (2.6%); and EF990724 and KM245037 by 20 bp (3.8%). There was no intraspecific sequence variation (0.0%) within the obtained sequences of P. terebranus. Seventy-three partial 28S-rRNA gene sequences (the four newly obtained sequences from this study and 69 sequences of closely related species within the genus and species from other closely related genera published in GenBank) constituted the data set for phylogenetic analysis. Phylogenetic relationships as inferred from Bayesian analysis of the 28S-rRNA gene sequence data set with GTR + I + G substitution model are shown in Fig. 5. Figure 5: Bayesian tree inferred under the GTR + I + G model from D2-D3 expansion segment of 28S-rRNA partial sequences of Diplogasteroides spp., Parasitorhabditis spp., and other closely related species from other genera. Posterior probability values exceeding 50% are given on appropriate clades. The studied population is indicated in bold. Outgroup taxon: Rhabditoides regina. Only P. terebranus ITS-rRNA gene was successfully amplified (OQ305560 to OQ305564). There were no sequences of any member species within the genus Parasitorhabditis in GenBank for comparison with the newly obtained sequences. The closest accession was a representative of Rhabdias delangei population, with identity of 90%. Bayesian analysis of the partial ITS-rRNA gene was omitted due to the limited number (or unavailability) of sequences with close homology to the newly obtained sequences in the GenBank database. The closest accession in BLASTN hits for the COI gene sequences of Diplogasteroides sp. (OQ281740 and OQ281741) was a population representative of D. asiaticus (LC027676) with 92% identity; differing by 51 bp (8.3%). The newly obtained sequences differed from other closely clustered species; D. nix, D. luxuriosae and D. andrassyi by 65 bp (10.6%), 69 bp (11.2%) and 72 bp (11.7%), respectively. Diplogasteroides sp. (OQ281740 and OQ281741) also differed from the distant D. nasuensis (LC027677) by 87 bp (14.2%). No intraspecific sequence variation was evident in the obtained COI gene sequences of Diplogasteroides sp. The COI gene sequences of P. terebranus herein represent the first amplification of the gene for the genus, and therefore, there were no definitive COI gene sequences for Parasitorhabditis in GenBank for comparison. However, the two generated sequences (OQ281742 and OQ281743) showed relative homology with COI gene sequences of species of other genera available in GenBank. These included; Teratorhabditis synpapillata (LN827630), Cylicostephanus goldi (AP017681), Cylicostephanus minutus (MT409394), and Allodiplogaster sudhausi (KT355738), from which significant differences of 82 bp (12.6%), 108 to 110 bp (16.8% to 17.0%), 111 bp (17.1%) and 354 to 356 bp (56.7% to 57.0%) were recorded. Thirty-five COI gene sequences of various species within the related genera constituted the data set for phylogenetic analysis. Phylogenetic relationships, as inferred from Bayesian analysis of the data set with GTR + I + G substitution model, are shown in Fig. 6. Figure 6: Bayesian tree inferred under the GTR + I + G model from COI partial sequences of Diplogasteroides spp., Parasitorhabditis spp., and members of closely related genera. Posterior probability values exceeding 50% are given on appropriate clades. The studied population is indicated in bold. Outgroup taxon: Ascaridia columbae. Discussion Bayesian analysis of the 18S-rRNA and D2-D3 expansion segment of Diplogasteroides was highly concordant with the analysis of Kanzaki et al. (2015). The 18S-rRNA and D2-D3 expansion segment sequences generated from amplicons of Diplogasteroides sp. are very close to those of D. asiaticus, D. nix, D. andrassyi, and D. luxuriosae. These species belong to the clade that is equivalent to the genus Rhabdontolaimus Fuchs, 1931, a junior synonym of Diplogasteroides (Sudhaus and Fürst von Lieven, 2003). The group is currently represented by nine nominal species, including D. carinthiacus (Fuchs, 1931) Rühm in Körner, 1954; D. haslacheri (Fuchs, 1931) Rühm in Körner, 1954; D. janae Massey, 1962; D. adephagus (Massey, 1974) Sudhaus & Fürst von Lieven, 2003; D. frontalis (Massey, 1974) Sudhaus & Fürst von Lieven, 2003; D. luxuriosae Kanzaki & Ide, 2016; D. andrassyi Kanzaki, Tanaka, Hirooka & Maehara, 2013; D. asiaticus Kanzaki, Woodruff, Akiba & Maehara, 2015; and D. nix Kanzaki, Sakamoto, & Maehara, 2016. All these species are known to share several characters, including a water droplet-shaped gubernaculum, ventrally bent spicules one-half to one-third along the spicule length from the manubrium and paired gonads (Kanzaki et al., 2013, 2015). The species are differentiated mainly by the positions of the second genital papillae, in combination with other qualitative characters. However, D. haslacheri, D. asiaticus, D. nix, D. andrassyi, and D. carinthiacus share an identical arrangement of genital papillae and other diagnostic characters such as the elongated, conical female tails (Kanzaki et al., 2013, 2015). According to Kanzaki et al. (2015), these species may be distinguished by the relative length of the male tail spike, the position of nerve ring and position of excretory pore, in addition to hybridization tests. But, as observed in this study, the position of excretory pores appears to be a variable character, especially when a high number of specimens is examined. Based on DNA inferences, particularly the nearly full-length 18S-rRNA and D2-D3 region, very limited interspecific sequence variations of 1 to 7 bp have been recorded among these species. However, similar to cryptic species among plant parasitic nematodes (see Subbotin, 2015; Mwamula et al., 2020), COI gene sequences are more promising, with significant interspecific variation of 8.3% (51 bp) to 11.7% (72 bp). Thus, in addition to future hybridization studies, the COI gene might be a powerful, discriminating DNA barcoding marker for precise identification of these cryptic species within the genus, despite the existence of haplotypes in some diplogastrid groups, as detailed by Kanzaki et al. (2020). On the other hand, morphometrics of the studied population of P. terebranus agree well with the descriptions of Massey (1974) except for the relatively long body, higher a ratio (20.5 to 30.0 versus 19.7 to 20.1), and the number of bursal rays (nine recorded in the original species description versus 10 in the current study). As noted by Sudhaus and Fitch (2001), and Carta et al. (2010), species belonging to the genus Parasitorhabditis have one of the most variable male tail ray patterns among the genera of Rhabditida. According to Andrássy (1983), the number of papillae is constant in the genus Parasitorhabditis: 10 pairs (two pairs preanal). This is echoed by Sudhaus and Fitch (2001), who stated this as being the main diagnostic character of the genus, although plesiomorphic features of the genus include, among others, a peloderan bursa, supported by 10 pairs of bursal papillae, two of which are located precloacally, and with the grouping of papillae varying by species. However, the type species, P. obtusa (Fuchs, 1915) Chitwood & Chitwood, 1950 has always been described with no exact number of bursal papillae, with ranges of 8 to 12 recorded in the available published literature (see Massey, 1974; Valizadeh et al., 2017). In the P. terebranus population described here, the third and fourth bursal rays appear too close and may give an impression of nine bursal rays in some specimens. However, 10 bursal rays are clearly evident, as illustrated in the species description. Our results therefore suggest that it is important to examine a large number of specimens before drawing definitive conclusions on species identity based on this set of characters. Additionally, there is need to re-examine the type specimen or characterize populations from type localities of the various species within the genus, which have been described with variations in the number of bursal rays (papillae). These variations, especially among species described by Massey (1974) are not thoroughly discussed in the review of Andrássy (1983) and Sudhaus and Fitch (2001). In conclusion, integrative taxonomic identification, considering both morphometric and molecular characterization of the studied populations of various species, is necessary to allow comprehensive comparisons with the respective type populations. This will supplement and resolve the current generic compendia within this taxonomically confounded group of nematodes. Acknowledgments This study was carried out with the support of the R & D Program for Forest Science Technology (Project No. 2021333D10-2223-CD02) provided by the Korea Forest Service (Korea Forestry Promotion Institute). ==== Refs Literature Cited Andrássy I. 1983 . A taxonomic review of the suborder Rhabditina (Nematoda: Secernentia) . Paris : Orstom . Andrássy I. 1984 . Klasse Nematoda (Ordnungen Monhysterida, Desmoscolecida, Araeolaimida, Chromadorida, Rhabditida) . Stuttgart, Germany : Gustav Fischer Verlag . Andreieva O. Korma O. Zhytova O. Martynchuk I. , and Vyshnevskyi A. 2020 . Beetles and nematodes associated with wither Scots pines . Lesnicky Casopis 66 : 49 - 59 . Baermann G. 1917 . Eine einfache Methode zur Auffindung von Ankylostomum (nematoden) Larven in Erdproben . Geneeskundig Tijdschrift voor Nederlandsch-Indie 57 : 131 - 137 . Bhat A. H. Srivastava S. Rana A. Chaubey A. K. Machado R. A. , and Abolafia J. 2020 . Morphological, morphometrical, and molecular characterization of Metarhabditis amsactae (Ali, Pervez, Andrabi, Sharma and Verma, 2011) Sudhaus, 2011 (Rhabditida, Rhabditidae) from India and proposal of Metarhabditis longicaudata as a junior synonym of M. amsactae . Journal of Nematology 52 : 1 - 23 . Carta L. K. Bauchan G. Hsu C. Y. , and Yuceer C. 2010 . Description of Parasitorhabditis frontali n. sp.(Nemata: Rhabditida) from Dendroctonus frontalis Zimmermann (Coleoptera: Scolytidae) . Journal of Nematology 42 : 46 - 54 . 22736836 Cesari M. Marescalchi O. Francardi V. , and Mantovani B. 2005 . Taxonomy and phylogeny of European Monochamus species: Ffirst molecular and karyological data . Journal of Zoological Systematics and Evolutionary Research 43 : 1 - 7 . Chapman J. A. 1964 . Nematode infestation and sex difference in response to log odours, in the cerambycid beetle, Leptura obliterata (Haldeman) . Bi-Monthly Progress Report Forestry , Canada 20 ( 3 20230017 ). Fuchs A. G. 1931 . Die genera: 1. Rhabditolaimus Fuchs, 2. Neodiplogaster Cobb, 3. Tylenchodon Fuchs . Zentralbulatt far das gesamte Forstwesen 57 : 177 - 194 . Giblin-Davis R. M. Ye W. Kanzaki N. Williams D. Morris K. , and Thomas W. K. 2006 . Stomatal ultrastructure, molecular phylogeny, and description of Parasitodiplogaster laevigata n. sp.(Nematoda: Diplogastridae), a parasite of fig wasps . Journal of Nematology 38 : 137 - 149 . 19259439 Girgan C. Du Preez G. Fourie H. , and Rashidifard M. 2021 . Morphological and molecular characterization of Goodey, 1929 (Nematoda: Diplogastridae) from South Africa: First report . Journal of Nematology 53 : 1 - 12 . 34179819 Grucmanová Š. , and Holuša J. 2013 . Nematodes associated with bark beetles, with focus on the genus Ips (Coleoptera: Scolytinae) in Central Europe . Acta Zoologica Bulgarica 65 : 547 - 556 . Herrmann M. Mayer W. E. , and Sommer R. J. 2006 . Nematodes of the genus Pristionchus are closely associated with scarab beetles and the Colorado potato beetle in Western Europe . Zoology 109 : 96 - 108 . 16616467 Holterman M. van der Wurff A. van den Elsen S. van Megen H. Bongers T. Holovachov O. Bakker J. , and Helder J. 2006 . Phylum-wide analysis of SSUrDNA reveals deep phylogenetic relationships among nematodes and accelerated evolution toward crown clades . Molecular Biology and Evolution 23 : 1792 - 1800 . 16790472 Iwahori H. Kanzaki N. , and Futai K. 2000 . A simple, polymerase chain reaction-restriction fragment length polymorphism-aided diagnosis method for pine wilt disease . Forest Pathology 30 : 157 - 164 . Kanzaki N. Futai K. 2002 A PCR primer set for determination of phylogenetic relationships of Bursaphelenchus species within the xylophilus group Nematology 4 35 41 Kanzaki N. Minagawa N. , and Futai K. 2002 . Description of Rhabdontolaimus psacotheae n. sp.(Diplogasterida: Diplogasteroididae), isolated from the yellow-spotted longicorn beetle, Psacothea hilaris (Coleoptera: Cerambycidae) and fig trees, Ficus carica . [In Japanese]. Journal of Nematology 32 : 7 - 12 . Kanzaki N. Ozawa M. Ota Y. , and Degawa Y. 2020 . Four species associated with two mass-occurring populations . Journal of Nematology 52 : 1 - 10 . Kanzaki N. Sakamoto H. , and Maehara N. 2016 . Diplogasteroides nix n. sp.(Nematoda: Diplogastridae), a cryptic species related to D. andrassyi, isolated from Monochamus urussovii (Coleoptera: Cerambycidae) from Hokkaido, Japan, with remarks on body surface structures . Nematology 18 : 753 - 773 . Kanzaki N. Tanaka R. Hirooka Y. , and Maehara N. 2013 . Description of Diplogasteroides andrassyi sp. n.(Rhabditida, Diplogastridae), associated with Monochamus grandis and Pinaceae trees in Japan . Journal of Nematode Morphology and Systematics 16 : 35 - 47 . Kanzaki N. Woodruff G. C. Akiba M. , and Maehara N. 2015 . Diplogasteroides asiaticus n. sp. is associated with Monochamus alternatus in Japan . Journal of Nematology 47 : 105 - 115 . 26170472 Kiontke K. , and Sudhaus W. 2000 . Phasmids in male Rhabditida and other secernentean nematodes . Journal of Nematode Morphology and Systematics 3 : 1 - 37 . Kwon T. Lim J. Sim S. Kwon Y. Son S. Lee K. Kim Y. Park J. Shin C. Ryu S. Lee C. Shin S. Chung Y. , and Park Y. 2006 . Distribution patterns of Monochamus alternatus and M. saltuarius (Coleoptera: Cerambycidae) in Korea . Journal of Korean Forest Society 95 : 543 - 550 . Mamiya Y. 1983 . Pathology of the pine wilt disease caused by Bursaphelenchus xylophilus . Annual Review of Phytopathology 21 : 201 - 220 . Massey C. L. 1962 . New species of Diplogasteridae (Nematoda) associated with bark beetles in the United States . Proceedings of the Helminthological Society of Washington 29 : 67 - 75 . Massey C. L. 1974 . Biology and taxonomy of nematode parasites and associates of bark beetles in the United States No. 446 . Washington, DC : Forest Service, U.S. Department of Agriculture . Meshkova V. Borysenko O. , and Pryhornytskyi V. 2018 . Forest site conditions and other features of Scots pine stands favorable for bark beetles . Proceedings of the Forestry Academy of Sciences of Ukraine 16 : 106 - 114 . Mwamula A. O. Lee G. Kim Y. H. Kim Y. H. Lee K. S. , and Lee D. W. 2020 . Molecular phylogeny of several species of Hoplolaimina (Nematoda: Tylenchida) associated with turfgrass in Korea, with comments on their morphology . Nematology 23 : 559 - 576 . Mwamula A. O. Lim T. H. Kim Y. Lee H. W. Kim Y. H. , and Lee D. W. 2022 . Morphological plasticity in the rice root nematode, Hirschmanniella oryzae (van Breda de Haan, 1902) Luc & Goodey, 1964 from Korea, with inferences from its ribosomal and mitochondrial DNA . European Journal of Plant Pathology 164 : 337 - 352 . Nguyen K. B. , and Hunt D. J. ( 2007 ). Entomopathogenic nematodes: Systematics, phylogeny and bacterial symbionts . Nematology Monographs and Perspectives 5 . Leiden, The Netherlands : Brill . Nickle W. R. 1973 . Identification of insect parasitic nematodes: A rewiew . – Experimental Parasitology 33 : 303 - 317 . 4574674 Nunn G. B. 1992 . Nematode molecular evolution . Ph.D. dissertation, University of Nottingham , Nottingham, UK . Poinar G. O. Jr . 1975 . Entomogenous nematodes: A manual and host list of insect-nematode associations . Leiden, The Netherlands : Brill . Ronquist F. Teslenko M. van der Mark P. Ayres D. L. Darling A. Höhna S. Larget B. Liu L. Suchard M. A. , and Huelsenbeck P. J. 2012 . MrBayes 3.2: Efficient Bayesian phylogenetic inference and model choice across a large model space . Systematic Biology 61 : 539 - 542 . 22357727 Rühm W. 1956 . Die Nematoden der Ipiden . Parasitologische Schriftenreihe 6 : 1 - 437 . Rühm W. 1965 . Zur „Wirtskreiserweiterung “einer mit Borkenkäfern (Scolytoidea, Col.) vergesellschafteten Nematodenart . Zeitschrift für Parasitenkunde 26 : 230 - 253 . 5875147 Sauvard D. 2007 . General biology of bark beetles . Pp. 63 - 88 Lieutier F. , ed. Bark and wood boring insects in living trees in Europe: A synthesis . Dordrecht : Springer . Seinhorst J. W. 1959 . A rapid method for the transfer of nematodes from fixative to anhydrous glycerin . Nematologica 4 : 67 - 69 . Subbotin S. A. 2015 . Heterodera sturhani sp. n. from China: A new species of the Heterodera avenae species complex (Tylenchida: Heteroderidae) . Russian Journal of Nematology 23 : 145 - 152 . Subbotin S. A. Vovlas N. Crozzoli R. Sturhan D. Lamberti F. Moens M. , and Baldwin J. G. 2005 . Phylogeny of Criconematina Siddiqi, 1980 (Nematoda: Tylenchida) based on morphology and D2-D3 expansion segments of the 28S-rRNA gene sequences with application of a secondary structure model . Nematology 7 : 927 - 944 . Sudhaus W. 1974 . Zur Systematik, Verbreitung, Oekologie und Biologie neuer und wenig bekannter Rhabditiden (Nematoda) 2. Teil. Zoologische Jahrbucher . Abteilung fur Systematik, Okologie und Geographie der Tiere 101 : 417 - 465 . [In German]. Sudhaus W. , and Fitch D. 2001 . Comparative studies on the phylogeny and systematics of the Rhabditidae (Nematoda) . Journal of Nematology 33 : 1 - 70 . 19265873 Sudhaus W. , and Fürst von Lieven A. 2003 . A phylogenetic classification and catalogue of the Diplogastridae (Secernentea, Nematoda) . Journal of Nematode Morphology and Systematics 6 : 43 - 90 . Susoy V. Kanzaki N. , and Herrmann M. 2013 . Description of the bark beetle associated nematodes Micoletzkya masseyi n. sp. and M. japonica n. sp. (Nematoda: Diplogastridae) . Nematology 15 : 213 - 231 . Swofford D. L. 2003 . PAUP*: Phylogenetic analysis using parsimony and other methods . Version 4.0a169. Sunderland, MA : Sinauer Associates . Tenkáčová I. , and Mituch J. 1987 . Nematodes new for the fauna of the Czechoslovak Socialist Republic with the affinity to scolytids (Coleoptera: Scolytidae) . Helminthologia 24 : 281 - 291 . Thompson J. D. Gibson T. J. Plewniak F. Jeanmougin F. , and Higgins D. G. 1997 . The CLUSTAL_X windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools . Nucleic Acids Research 25 : 4876 - 4882 . 9396791 Togashi K. 2008 . Vector-nematode relationships and epidemiology in pine wilt disease . Pp. 162 - 183 Vieira P. and Mota M. , eds. Pine wilt disease: A worldwide threat to forest ecosystems . Dordrecht : Springer . Tomalak M. Welch H. E. , and Galloway T. D. 1989 . Parasitism of Parasitorhabditis obtusa and P. autographi (Nematoda: Rhabditidae) in the digestive tract of their bark beetle (Coleoptera: Scolytidae) hosts . Journal of Invertebrate Pathology 53 : 57 - 63 . Valizadeh A. Goldasteh S. Rafiei-Karahroodi Z. , and Pedram M. 2017 . First record of the genus Parasitorhabditis Fuchs, 1937 (Rhabditida, Nematoda) from Iran with notes on morphological and molecular characters of the Iranian population of P. obtusa (Fuchs, 1915) Chitwood & Chitwood, 1950 . Zootaxa 4353 : 591 - 600 . 29245507 Wood D. L. 1982 . The role of pheromones, kairomones, and allomones in the host selection and colonization behavior of bark beetles . Annual Review of Entomology 27 : 411 - 446 .

PMC010xxxxxx/PMC10288305.txt

==== Front J Nematol J Nematol jofnem jofnem Journal of Nematology 0022-300X 2640-396X Sciendo 37359195 jofnem-2023-0021 10.2478/jofnem-2023-0021 Research Paper Virulence of Two Isolates of Meloidogyne enterolobii (Guava Root-Knot Nematode) from North Carolina on Cotton Lines Resistant to Southern Root-Knot Nematode (M. incognita) and Reniform Nematode (Rotylenchulus reniformis) Gaudin Amanda G. Wubben Martin J. martin.wubben@usda.gov McCarty Jack C. Jr. Jenkins Johnie N. USDA-ARS, Crop Science Research Lab, Genetics and Sustainable Agriculture Research Unit, 150 Twelve Lane, Mississippi State, MS 39762 USA. This paper was edited by Guiping Yan. 2 2023 23 6 2023 55 1 2023002126 8 2022 © 2023 Amanda G. Gaudin et al., published by Sciendo 2023 Amanda G. Gaudin et al., published by Sciendo https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. Abstract Meloidogyne enterolobii [the guava root-knot nematode (RKN)] is an emerging plant-parasitic nematode that poses a threat to Upland cotton (Gossypium hirsutum) production in the southeastern United States. Like other RKN spp., M. enterolobii has a wide host range and proven ability to overcome resistance sources that have helped protect crops from other Meloidogyne spp., including the southern RKN (Meloidogyne incognita). In this study we evaluated the virulence of two North Carolina M. enterolobii isolates on Upland cotton germplasm lines having resistance quantitative trait loci (QTL) to RKN (M240 RNR, MRk-Rn-1) and/or reniform nematode (Rotylenchulus reniformis) (M713 Ren1, MRk-Rn-1) in comparison to their susceptible recurrent parents (DPL61, SG747). Multiple assays using eggs or J2 as inoculum demonstrated that both isolates reproduced equally well on all germplasm lines, producing reproductive factor (RF) values ≥ 6 on the otherwise nematode-resistant lines. Measurements of seedling growth in control and inoculated containers suggested that existing nematode-resistance QTL may offer a level of tolerance to M. enterolobii infection that should be further explored in greenhouse and field environments. Meloidogyne enterolobii infection of SG747 and MRk-Rn-1 showed nearly identical stages of symptom and nematode development over a time-course of 24 days. These data demonstrate that existing RKN and RN resistance QTL available in elite cotton varieties to producers are most likely insufficient in preventing yield loss due to M. enterolobii and that future research should focus on (i) understanding the M. enterolobii–cotton interaction at the molecular level, and (ii) screening novel germplasm collections to identify resistance loci. Keywords Cotton resistance root-knot nematode Meloidogyne ==== Body pmcThe guava root-knot nematode (Meloidogyne enterolobii) is a sedentary endoparasitic species that parasitizes many important crops, including guava, tomato, sweetpotato, and cotton. Since its initial discovery in China, M. enterolobii has been identified in Africa, southeast Asia, and the Americas (Yang and Eisenback, 1983; Elling, 2013). Initially, the nematode was misidentified as two different species, M. mayaguensis and M. enterolobii with M. mayaguensis being documented as an important crop pest on tomato and peppers (Brito et al., 2004). However, it has since become accepted that M. mayaguensis and M. enterolobii are the same species of nematode based on the sequence identity of mitochondrial DNA (Xu et al., 2004; Castagnone-Sereno, 2012). The M. enterolobii life cycle is like that of other Meloidogyne spp. (Elling, 2013). Mature females deposit eggs into a gelatinous matrix secreted by the rectal glands. Development within the egg results in the first-stage juvenile that molts inside the egg to become the infective second-stage juvenile (J2). The J2 infective life stage emerges from the eggs and infects host plant roots preferentially through the root tips and migrates intercellularly through the root cortex. The J2 becomes sedentary and a collection of feeding cells, called “giant cells,” are formed near the endodermis and function to supply nutrition. The giant cell is a single cell that undergoes multiple rounds of mitosis without cytokinesis, creating a large multinucleate cell specialized for nematode feeding. Molting through the J3 and J4 life stages results in mature female development and egg deposition. M. enterolobii reproduces by parthenogenesis and has a similar chromosomal makeup as the southern root-knot nematode (RKN) M. incognita (Cetintas et al., 2007; Collett et al., 2021). However, the time it takes to complete a full life cycle and reproduce is variable when compared to related species, sometimes taking longer and other times being significantly faster, depending on isolate genotype, soil temperature, host plant species, and other environmental factors (Collett et al. 2021). In the contiguous United States, M. enterolobii was first reported in Florida, being initially classified as M. mayaguensis, where it was morphologically and molecularly differentiated from other common Meloidogyne spp., e.g., M. incognita and M. arenaria (Brito et al., 2004). In the late 2010s, M. enterolobii was discovered in North and South Carolina, and Louisiana (Ye et al., 2013; Rutter et al., 2019; Philbrick et al. 2020). Recently, M. enterolobii was intercepted in Louisiana on sweetpotato storage roots imported from North Carolina, confirming previous suspicions of a link between M. enterolobii spread and sweetpotato planting stock (Thiessen, 2018; Rezende et al., 2022). Because of its wide geographical distribution and host range, M. enterolobii is considered an important nematode species and is on the European and Mediterranean Plant Protection Organization (EPPO) quarantine list because of high pathogenicity (Castagnone-Sereno, 2012). M. enterolobii is of special concern because it can infect and proliferate on crop varieties that have resistance to other RKN species. This ability has been confirmed in sweetpotato, tomato, potato, and soybean (Brito et al., 2007; Cetintas et al., 2008). In a study comparing Meloidogyne pathogenicity on tomato, M. enterolobii was found to cause more severe root galling than other Meloidogyne spp., including southern RKN (Cetintas et al., 2007). The Mi-1 gene in tomato confers resistance to M. incognita, M. arenaria, and M. javanica; however, this gene was ineffective against M. enterolobii (Brito et al., 2007). Furthermore, the N gene in bell pepper and the Tabasco gene in pepper, both of which confer resistance to M. incognita, M. arenaria, and M. javanica, did not protect against M. enterolobii (Brito et al., 2007). The potential spread of M. enterolobii is an emerging problem for cotton growers in the southeastern U.S. and management options need to be identified to prevent significant crop loss. Decades of research focused on developing cotton germplasm resistant to M. incognita and reniform nematode (RN) (summarized in McCarty et al., 2021) recently yielded the deployment of elite Upland cultivars resistant to both species. For M. incognita, resistance is mediated by two loci, qMi-C11 and qMi-C14, whereas resistance to RN is largely controlled by a single locus on chromosome 21 (Renbarb2) (Gutiérrez et al., 2010; Wubben et al., 2017). With the availability of these nematode-resistant cultivars to cotton producers, it is vital that we understand how these resistance loci impact the infection and reproduction of M. enterolobii on cotton. Toward this end, our laboratory secured the necessary phytosanitary permits from the State of Mississippi Plant Advisory Board and the Animal and Plant Health Inspection Service (APHIS) to import two North Carolina isolates of M. enterolobii to test for potential differences in virulence on southern RKN and RN-resistant cotton germplasm lines. Materials and Methods Nematode cultures Southern RKN (Meloidogyne incognita) race 3 and M. enterolobii cultures were maintained on the Upland cotton germplasm line “M8” in separate growth chambers. Meloidogyne enterolobii isolate “NC.1” was received from Dr. William Rutter (USDA-ARS) and isolate “18-5126” from Dr. Eric Davis (North Carolina State University) under APHIS permit number P526P-19-02617. Details describing the isolation and characterization of these isolates can be found in Rutter et al. (2021) and Schwarz et al. (2021), respectively. A quarantine permit was obtained from the State of Mississippi Department of Agriculture and Commerce that allowed experiments to be conducted in growth chambers at our facility. Soil and tissues of M. enterolobii-infected plants were autoclaved prior to disposal. Water used for washing soil from infected roots was passed through a #500-sieve before going down a drain. Cotton germplasm and inoculation experiments The cotton germplasm lines M240-RNR (PI 592511) and M713 Ren1 (PI 665928) are highly resistant to RKN and RN, respectively (Shepherd et al., 1996; McCarty et al., 2013). The germplasm line M Rk-Rn-1 (PI 678938) combines the RKN and RN resistance present in M240-RNR and M713 Ren1 into a single genotype (McCarty et al., 2017). Susceptible checks were “Deltapine61” (DPL61; PI 607174) and “Sure-Grow747” (SG747; PI 656375). DPL61 served as the recurrent parent for M240-RNR, while SG747 was the recurrent parent for M713 Ren1 and M Rk-Rn-1. Meloidogyne infection experiments were conducted using 16 oz plastic Solo® cups containing approximately 500 cm3 of an autoclaved mixture of 2-part sand:1-part Wickham sandy loam soil. Seeds were scarified, wrapped in wet paper towels, and incubated overnight at 30°C to promote germination. Seeds with radicles of around 0.5 cm were planted two per Solo® cup. For experiments using nematode eggs as inoculum, M. incognita or M. enterolobii eggs were collected from infected roots of culture cotton plants according to Hussey and Barker (1973). One day after planting, each Solo® cup was inoculated with 10,000 eggs delivered as 4 × 1 mL aliquots of a 2,500 egg/mL solution pipetted into four holes (two holes flanking each seedling). For experiments using M. enterolobii second-stage juveniles (J2) as inoculum, eggs of each isolate were collected as described above and layered over a #500 nylon mesh submerged in distilled water. Hatched J2 were collected daily over a period of 5 to 7 days. Each Solo® cup was inoculated with 2,000 J2 as 4 × 1 mL aliquots of a 500 J2/mL solution pipetted into four holes (two holes flanking each seedling). All experiments were performed in a Percival PGC-9/2 growth chamber at a constant temperature of 30°C under a 16 hr day/8 hr night cycle. Except for time course-dependent studies, all experiments were halted six weeks post-inoculation and root tissues collected. Gall severity was measured on a 1–10 scale (1=no galls present and 10=large galls present on entire root system). Root tissues were weighed and the eggs extracted, stained, and counted as previously described (Gaudin and Wubben, 2021). The reproductive factor (RF) was calculated as total eggs collected ÷ initial inoculum. Acid fuchsin staining and visualization of infected roots was performed as previously described (Wubben et al., 2020). Measurements of plant growth traits (stem height, shoot fresh weight, leaf number and node number) were performed six-weeks after inoculation. Stem height was measured from the stem base to the apical meristem. Statistical analyses All experiments were conducted according to a completely randomized design with five to six replications per treatment. Data collected from each experiment were analyzed separately using the general linear model function and analysis of variance with SAS version 9.4 (SAS Institute). Egg per gram root values were log-transformed. Interaction effects were determined for nematode population × line for the initial three experiments, and timepoint × nematode × line was determined for the fourth experiment. Means were separated at P ≤ 0.05 using least squared means with the Tukey-Kramer adjustment for multiple comparisons. Results Reproduction of Meloidogyne enterolobii isolates on RKN and RN-resistant germplasm In our first assay, eggs were used as inoculum to compare the reproduction of M. enterolobii to that of RKN on susceptible (DPL61 and SG747) and RKN/RN resistant germplasm (M240 RNR, M713 Ren1, and M Rk-Rn-1). The interaction of cotton line with nematode isolate was significant at the α ≤ 0.05 for RF, log10(eggs g−1 root), and gall score (Fig. 1). As expected, the RF of M. incognita was significantly lower on M240 RNR and MRk-Rn-1 compared to the other germplasm lines (Fig. 1A). In contrast, both M. enterolobii isolates showed an RF > 6 on M240 RNR and MRk-Rn-1, with isolate NC.1 having a significantly greater RF versus isolate 18-5126 (Figure 1A). When inoculated with isolate NC.1, there was no difference in RF among DPL61, SG747, M240 RNR, and MRk-Rn-1 (Fig. 1A). Isolate 18-5126 showed similar RF values across all germplasm lines except for DPL61 which showed greater reproduction compared to the RKN/ RN-resistant lines but SG747 (Fig. 1A). Curiously, the RN-resistant line M713 Ren1 showed the lowest RF for isolate NC.1 versus all other lines (Fig. 1A). The log10(eggs g−1 root) and gall score values showed trends similar to those of RF, with a few notable exceptions (Figs. 1A, B). M713 Ren1, in contrast to RF, showed no difference in log10(eggs g−1 root) or gall score between M. enterolobii isolates. Likewise, no differences between 18-5126 and NC.1 were observed for log10(eggs g−1 root) or gall score on M240 RNR and M Rk-Rn-1, in contrast to their respective RF values. In fact, the log10(eggs g−1 root) of M. enterolobii-inoculated M240 RNR, MRk-Rn-1, and M713 Ren1 were not different from RKN-inoculated DPL61 and SG747 (Fig. 1B). Overall, gall scores revealed a similar trend in which no differences were found in DPL61 and SG747 under the pressure of different nematode species (Fig. 1C). Lines M240 RNR and MRk-Rn-1 tended to show less severe galling in response to M. enterolobii compared to DPL61, SG747, and M713 Ren1; however, the galling observed was clearly within what would be expected in a susceptible genotype (Fig. 1C). Figure 1: Virulence of Meloidogyne enterolobii isolates 18-5126 and NC.1 on cotton germplasm lines resistant to M. incognita (RKN) (M240 RNR), resistant to RN (M713 Ren1), or resistant to both RKN and RN (M Rk-Rn-1) in comparison to a greenhouse-maintained RKN race 3 population. Also included were nematode-susceptible obsolete cultivars Deltapine61 (DPL61) and SureGrow747 (SG747). Eggs were used as inoculum and virulence was measured as (A) RF, (B) log10 eggs g−1 root, and (C) gall score index (1–10 scale). Tukey-Kramer multiple comparison was used to determine significant differences between means at α = 0.05 as denoted by different letters (n = 6). Figure 2: Virulence of Meloidogyne enterolobii isolates 18-5126 and NC.1 on cotton germplasm lines resistant to M. incognita (RKN) (M240 RNR), resistant to RN (M713 Ren1), or resistant to both RKN and RN (M Rk-Rn-1) using second-stage juveniles as inoculum. Also included were nematode-susceptible obsolete cultivars Deltapine61 (DPL61) and SureGrow747 (SG747). Virulence was measured as (A) RF ± SE, (B) log10 eggs g−1 root ± SE, and (C) gall score index ± SE (1–10 scale) of six replicates. A second assay was performed using hatched J2 as inoculum to account for possible differences in egg hatch between isolates 18-5126 and NC.1. The results of this experiment showed that no significant differences in RF and log10(eggs g−1 root) were detected between isolates for any of the germplasm lines (Figs. 2A, B). The interaction effect of cotton line × M. enterolobii isolate was not significant; however, the isolates were significantly different for gall score, with 18-5126 generally producing a higher gall score than NC.1 (Fig. 2C). Based on the two assays, we concluded that isolates 18-5126 and NC.1 did not differ significantly in their virulence across the germplasm lines; therefore, only isolate NC.1 was used in subsequent experiments. Effects of M. enterolobii isolate NC.1 on cotton seedling growth At the conclusion of both assays described above, we noticed that lines M240 RNR and MRk-Rn-1 tended to appear “healthier” compared to DPL61 and SG747 in terms of leaf number, stem height, and general appearance. To help determine the impact of M. enterolobii infection on cotton seedling growth, a third experiment was conducted that compared plant growth traits of inoculated plants to their uninoculated control counterparts. Because our experiments were limited to growth chambers, we decided to use the following measures: stem height (mm), shoot fresh weight (g), leaf number, and stem height/node number (Table 1). Significant decreases in stem height due to M. enterolobii infection were observed for DPL61, SG747, and MRk-Rn-1 (Table 1). M240 RNR and M713 Ren1 also showed a decrease in stem height; however, this decrease was not significant. A similar pattern was observed for stem height/node number, where only M240 RNR and M713 Ren1 did not show a significant reduction in response to infection. Stem height/node number is commonly used as an indicator of stress on a cotton plant (Lu et al., 2014). Fresh shoot weight was only significantly impacted in SG747. For leaf number, M. enterolobii infection showed a significant effect for SG747 and MRk-Rn-1. Table 1. Effect of Meloidogyne enterolobii infection on cotton growth under growth chamber conditions. Presented are mean ± SE after 6 weeks of growth under uninoculated (control) or inoculated conditions (n = 5). Control Inoculated Stem Height (mm) DPL61 124.1 ± 3.2 79.3 ± 3.7** M240 RNR 147.8 ± 4.6 133.8 ± 22.8 SG747 143.5 ± 2.8 89.6 ± 2.1** M713 Ren1 144.0 ± 17.0 131.0 ± 12.8 MRk-Rn-1 157.6 ± 10.5 111.6 ± 5.4** Shoot Fresh Weight (g) DPL61 2.49 ± 0.13 1.94 ± 0.23 M240 RNR 2.75 ± 0.23 2.01 ± 0.27 SG747 3.23 ± 0.45 1.48 ± 0.19** M713 Ren1 2.58 ± 0.22 3.39 ± 0.42 MRk-Rn-1 2.99 ± 0.74 1.83 ± 0.11 Leaf Number DPL61 3.7 ± 0.1 3.8 ± 0.4 M240 RNR 3.8 ± 0.1 3.5 ± 0.5 SG747 3.5 ± 0.2 2.4 ± 0.3** M713 Ren1 3.9 ± 0.3 3.8 ± 0.3 MRk-Rn-1 4.1 ± 0.3 3.1 ± 0.2** Stem Height/Node Number DPL61 28.9 ± 0.5 21.1 ± 1.8** M240 RNR 33.2 ± 1.9 33.2 ± 1.5 SG747 40.0 ± 2.3 26.5 ± 0.9** M713 Ren1 31.7 ± 4.7 27.6 ± 2.2 MRk-Rn-1 40.5 ± 2.3 29.5 ± 1.1** ** Significantly different (P ≤ 0.05) from corresponding control mean. Figure 3: Time-point dependent reproduction of the Meloidogyne enterolobii isolate NC.1 on cotton germplasm lines M240 RNR and MRk-Rn-1 in comparison to the susceptible cultivar SureGrow747 (SG747). Total eggs were extracted at 35-, 42-, 49-, and 56-days post inoculation (DPI). Isolate NC.1 reproduction is presented as (A) RF (total eggs extracted/initial inoculum) and (B) log10 eggs g−1 root. Significant differences between means at α = 0.05 noted by different letters (n = 6). Eggs were extracted from the inoculated plants in this experiment and were counted. The RF was similar between DPL61, SG747, and M240 RNR (25.1, 23.0, and 25.1, respectively). In contrast, M713 Ren1 and MRk-Rn-1 both showed RF values of 12.2, significantly lower than the others. However, this difference disappeared when log10(eggs g−1 root) was determined and showed no differences between any of the lines. Time-course reproduction of isolate NC.1 In a final assay, we measured M. enterolobii reproduction on SG747, M240 RNR, and MRk-Rn-1 at four time points: 35, 42, 49, and 56 days post-inoculation (DPI). Previous experiments had measured reproduction at six weeks (42-DPI); however, measuring at multiple time points may shed light on the relative speed of infection and development of M. enterolobii in different germplasm lines. In general, reproduction decreased over time for each germplasm line (Fig. 3). The first time point (35-DPI) showed the highest level of reproduction in each line and revealed a significantly higher RF in SG747 versus M240 RNR and MRk-Rn-1 (Fig. 3A). This difference was also observed at 42-DPI but largely disappeared by 49- and 56-DPI (Fig. 3A). Analysis of log10(eggs g−1 root) did not reveal a significant line × time point interaction, and only slight differences were detected between lines and time points (Fig. 3B). In contrast to the RF data, log10(eggs g−1 root) revealed no significant differences between lines within each time point, with the single exception of M240 RNR at 56-DPI. Time course of isolate NC.1 infection in SG747 and MRk-Rn-1 seedlings Meloidogyne enterolobii isolate NC.1-infected SG747 and MRk-Rn-1 roots were collected and stained with acid fuchsin at 4, 8, 12, 16, 20, and 24 DPI (Fig. 4). The timing of infection and rate of juvenile development was nearly identical between SG747 and MRk-Rn-1. At 4-DPI, infective J2 had penetrated the root tips of both lines and migrated intercellularly through the cortex. By 8-DPI, root swelling was evident and sedentary parasitic J2 were observed. Nematode development progressed rapidly after 8-DPI such that severe root galling and J4 females were present in both lines at 12- and 16-DPI. Mature females and egg deposition was evident by 20-DPI and manifested throughout the root system of both lines by 24-DPI. Figure 4: Time-course of Meloidogyne enterolobii isolate NC.1 infection of SureGrow747 and MRk-Rn-1. Infected roots were collected at 4-, 8-, 12-, 16-, 20-, and 24-days after inoculation (DAI) and stained with acid fuchsin. Discussion Meloidogyne enterolobii is an emerging pathogen in the cotton production areas of the southeastern U.S. Reports from Brazil indicated that cotton varieties expressing resistance to southern RKN were susceptible to populations of M. enterolobii (Galbieri et al., 2020). Multiple isolates of M. enterolobii have been collected from the Carolinas and evaluated in greenhouse experiments for differences in virulence on a range of sweetpotato genotypes (Schwarz et al., 2020; Rutter et al., 2021). In this study, we evaluated two such isolates regarding their reproduction on cotton germplasm lines having strong resistance to southern RKN, mediated by qMi-C11 and qMi-C14, and/or RN, mediated by the Renbarb2 QTL. In general, the two North Carolina isolates used in our study, “18-5126” and “NC.1,” reproduced to similar levels on all cotton germplasm lines tested, with slight variations between experiments. Both isolates were able to overcome southern RKN and RN resistance QTL and induce severe root galling and high RF values. These data suggest the 18-5126 and NC.1 isolates are genetically similar and may originate from the same original introduction into North Carolina. A similar conclusion was determined by testing four North Carolina isolates on sweetpotatoes, which showed no differences in virulence (Schwarz et al., 2020). In contrast, a South Carolina M. enterolobii isolate, “SC.1,” was discovered that showed zero reproduction on Upland cotton (Rutter et al., 2021); however, the nature of this cotton resistance to SC.1 remains unknown. Meloidogyne enterolobii has a history of evading or overcoming RKN spp. resistance in many crops, making the results of this study not wholly unexpected. In a study on sweetpotatoes, M. enterolobii reproduced well on multiple cultivars that were previously identified as moderately resistant to RKN (Brito et al., 2020). The tomato Mi-1 resistance gene is effective for the management of southern RKN and two other root-knot species, M. arenaria and M. javanica. Despite this multi-specie protection, M. enterolobii reproduced on Mi-1 tomato cultivars like susceptible tomato varieties (Kiewnick et al., 2009). Similarly, the N gene in peppers provides resistance to southern RKN, M. arenaria, and M. javanica, but this gene was also overcome by M. enterolobii (Kiewnick et al., 2009). We determined existing cotton nematode resistance genes were ineffective against M. enterolobii, much like the Mi-1 gene in tomato and the N gene in pepper. The ability of M. enterolobii to overcome the resistance of related nematode species, particularly M. incognita, is problematic for future nematode management efforts (Kiewnick et al., 2009). In regard to cotton, the mechanism of defense evasion needs to be examined to better determine potential control options. For example, on the molecular level, M. enterolobii is either going undetected by qMi-C11 and qMi-C14, or these QTL are being triggered, but the resulting defense is ineffective against M. enterolobii, possibly via effector triggering susceptibility. Understanding how M. enterolobii is overcoming the resistance is important for continuing the search for novel resistance. With the knowledge that current Upland cotton RKN and RN-resistant genes are ineffective, the search for M. enterolobii resistance will require looking elsewhere and searching for novel genes and allele combinations. Extensive screening has been a successful, albeit labor-intensive, tool used to identify nematode resistance genes in cotton. For example, the currently used RN resistance was originally identified in G. barbadense accession GB-713 during such germplasm screenings, leading to the development of RN-resistant Upland cotton lines (Robinson et al., 2004). Common cotton breeding methods are effective for novel cotton line development. In early RKN resistance screenings, two moderately resistant lines (Clevewilt-6 and Mexico Wild Jack Jones) were crossed to develop the highly resistant line Auburn 634, which had qMi-C11 and qMi-C14 resistance loci (Shepherd, 1974; Gutiérrez et al., 2010). Other tools like random mating are often used to break negative genetic linkages by outcrossing between multiple diverse parents (Gutiérrez et al., 2006). Linkages are broken by forming new combinations of alleles and potentially new phenotypes. For example, these novel combinations can then be screened for potential resistance or tolerance to M. enterolobii. Historically, superior gene combinations have been generated in randomly mated populations (Gutiérrez et al., 2006). In summary, the present study indicated that existing nematode resistance QTL in cotton are ineffective against M. enterolobii. Due to experimental restrictions limiting work to only growth chambers, these data cannot address questions regarding the effect of season-long exposure of cotton germplasm lines to M. enterolobii infection. In addition to germplasm screening for resistance, future work should incorporate commercially available cotton varieties having the same nematode resistance QTL as described in this study, into field trials to examine any potential tolerance afforded by existing resistance QTL. Acknowledgments The authors would like to thank Drs. John Brooks and Dana Miles (USDA-ARS) for their helpful review of the manuscript. The authors would also like to thank Dr. Don Jones at Cotton Incorporated for providing financial support for this project in the form of a Cotton Incorporated Fellowship awarded to Ms. Amanda Gaudin (Project Number 18-191). Disclaimer Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer. ==== Refs Literature Cited Brito J. A. Desaeger J. Dickson D. W. 2020 Reproduction of Meloidogyne enterolobii on selected root-knot nematode resistant sweetpotato (Ipomoea batatas) cultivars Journal of Nematology 52 1 6 Brito J. A. Powers T. O. Mullin P. G. Inserra R. N. Dickson D. W. 2004 Morphological and molecular characterization of Meloidogyne mayaguensis isolates from Florida Journal of Nematology 36 232 240 19262811 Brito J. A. Stanley J. D. Kaur R. Cetintas R. Di Vito M. Thies J. A. Dickson D. W. 2007 Effects of the Mi-1, N and tabasco genes on infection and reproduction of Meloidogyne mayaguensis on tomato and pepper genotypes Journal of Nematology 39 327 332 19259507 Castagnone-Sereno P. 2012 Meloidogyne enterolobii (= M. mayaguensis): profile of an emerging, highly pathogenic, root-knot nematode species Nematology 14 133 138 Cetintas R. Brito J. A. Dickson D. W. 2008 Virulence of four Florida isolates of Meloidogyne mayaguensis to selected soybean genotypes Nematropica 38 127 135 Cetintas R. Kaur R. Brito J. A. Mendes M. L. Nyczepir A. P. Dickson D. W. 2007 Pathogenicity and reproductive potential of Meloidogyne mayaguensis and M. floridensis compared to three common Meloidogyne spp Nematropica 37 21 32 Collett R. L. Marais M. M. Aneel M. D. Ashidifard M. R. Fourie H. F. 2021 Meloidogyne enterolobii, a threat to crop production with particular reference to sub-Saharan Africa: an extensive, critical and updated review Nematology 23 3 247 285 Elling A. A. 2013 Major emerging problems with minor Meloidogyne species Phytopathology 103 1092 1102 23777404 Galbieri R. Davis R. F. Scoz L. B. Belot J. L. Skantar A. M. 2020 First report of Meloidogyne enterolobii on cotton in Brazil Plant Disease 104 2295 Gaudin A. G. Wubben M. J. 2021 Genotypic and phenotypic evaluation of wild cotton accessions previously identified as resistant to root-knot (Meloidogyne incognita) or reniform nematode (Rotylenchulus reniformis) Euphytica 217 207 Gu M. Bui H. X. Ye W. Desaeger J. A. 2021 First report of Meloidogyne enterolobii on sweet potato in Florida, USA Nematropica 51 36 40 Gutiérrez O. A. Bowman D. T. Cole C. B. Jenkins J. N. McCarty J. C. Wu J. Watson C E. 2006 Development of random-mated populations using bulked pollen methodology: cotton as a model Journal of Cotton Science 10 175 179 Gutiérrez O. A. Jenkins J. N. McCarty J. C. Wubben M. J. Hayes R. W. Callahan F. E. 2010 SSR markers closely associated with genes for resistance to root-knot nematode on chromosomes 11 and 14 of Upland cotton Theoretical and Applied Genetics 121 1323 1337 20607210 Hussey R. S. Barker K. R. 1973 A comparison of methods for collecting inocula for Meloidogyne spp., including a new technique Plant Dis Rep 57 1025 1028 Kiewnick S. Dessimoz M. Franck L. 2009 Effects of the Mi-1 and the N root-knot nematode-resistance gene on infection and reproduction of Meloidogyne enterolobii on tomato and pepper cultivars Journal of Nematology 41 134 139 22661786 Lu P. Davis R. F. Kemerait R. C. van Iersel M. W. Scherm H. 2014 Physiological effects of Meloidogyne incognita infection on cotton genotypes with differing levels of resistance in the greenhouse Journal of Nematology 46 352 359 25580028 McCarty J. C. Jr. Jenkins J. N. Wubben M. J. Gutiérrez O. A. Hayes R. W. Callahan F. E. Deng D. 2013 Registration of three germplasm lines of cotton derived from Gossypium barbadense L. accession GB713 with resistance to reniform nematode Journal of Plant Registrations 7 220 223 McCarty J. C. Jr. Jenkins J. N. Wubben M. J. Hayes R. W. Callahan F. E. Deng D. 2017 Registration of six germplasm lines of cotton with resistance to the root-knot and reniform nematodes Journal of Plant Registrations 11 168 171 McCarty J. C. Jr. Jenkins J. N. Saha S. Wubben M. J. 2021 History of USDA-ARS cotton host plant resistance and breeding research at Mississippi State, MS Journal of Cotton Science 25 101 157 Philbrick A. N. Adhikari T. B. Louws F. J. Gorny A. M. 2020 Meloidogyne enterolobii, a major threat to tomato production: current status and future prospects for its management Frontiers in Plant Science 11 1 13 32117356 Rezende J. S. Clark C. A. Sistrunk M. W. Watson T. 2022 Interception of Meloidogyne enterolobii on sweetpotato in Louisiana Nematropica 52 1 20230021 Robinson A. F. Bridges A. C. Percival A. E. Edward A. 2004 New sources of resistance to the reniform (Rotylenchulus reniformis) and root-knot (Meloidogyne incognita) nematode in upland (Gossypium hirsutum L.) and Sea Island (G. barbadense L.) cotton Journal of Cotton Science 8 191 197 Rutter W. B. Skantar A. M. Handoo Z. A. Mueller J. D. Aultman S. P. Agudelo P. 2019 Meloidogyne enterolobii found infecting root-knot nematode resistant sweetpotato in South Carolina, United States Plant Disease 103 775 2019 Rutter W. B. Wadl P. A. Mueller J. D. Agudelo P. 2021 Identification of sweet potato germplasm resistant to pathotypically distinct isolates of Meloidogyne enterolobii from the Carolinas Plant Disease 105 3147 3153 33599516 Schwarz T. Li C. Ye W. Davis E. 2020 Distribution of Meloidogyne enterolobii in eastern North Carolina and comparison of four isolates Plant Health Progress 21 91 96 Schwarz T. R. Li C. Yencho G. C. Pecota K. V. Heim C. R. Davis E. L. 2021 Screening sweetpotato genotypes for resistance to a North Carolina isolate of Meloidogyne enterolobii Plant Disease 105 1101 1107 32880532 Shepherd R. L. 1974 Transgressive segregation for root-knot nematode resistance in cotton Crop Science 14 872 875 Shepherd R. L. McCarty J. C. Jenkins J. N. Parrott W. L. 1996 Registration of nine cotton germplasm lines resistant to root-knot nematode Crop Science 36 3 820 Thiessen L. 2018 Impact of guava root knot nematode internal quarantine on tobacco and field crops in NC NC State Extension, North Carolina State University go.ncsu.edu/readext?557138. Wubben M. J. Gaudin A. G. McCarty J. C. Jenkins J. N. 2020 Analysis of cotton chromosome 11 and 14 root-knot nematode resistance quantitative trait loci effects on root-knot nematode postinfection development, egg mass formation, and fecundity Phytopathology 110 927 932 31961253 Wubben M. J. McCarty J. C. Jr. Jenkins J. N. Callahan F. E. Deng D. D. 2017 Individual and combined contributions of the Renbarb1, Renbarb2, and Renbarb3 quantitative trait loci to reniform nematode (Rotylenchulus reniformis Linford & Oliveira) resistance in upland cotton (Gossypium hirsutum L.) Euphytica 213 1 9 Xu J. Liu P. Meng Q. Long H. 2004 Characterisation of Meloidogyne species from China using isozyme phenotypes and amplified mitochondrial DNA restriction fragment length polymorphism European Journal of Plant Pathology 110 309 315 Yang B. Eisenback J. D. 1983 Meloidogyne enterolobii n. sp. (Meloidogynidae), a root-knot nematode parasitizing pacara earpod tree in China Journal of Nematology 15 381 391 19295822 Ye W. M. Koenning S. R. Zhuo K. Liao J. L. 2013 First report of Meloidogyne enterolobii on cotton and soybean in North Carolina, United States Plant Disease 97 1262

PMC010xxxxxx/PMC10306098.txt

==== Front Front Public Health Front Public Health Front. Public Health Frontiers in Public Health 2296-2565 Frontiers Media S.A. 37388157 10.3389/fpubh.2023.1188741 Public Health General Commentary Commentary: Establishing the college Return to Learn team for concussion: a practical approach Bevilacqua Zachary W. 1 * McPherson Jacob 2 1Department of Exercise Science, Rochester Institute of Technology, Rochester, NY, United States 2Department of Rehabilitation Sciences, University at Buffalo, Buffalo, NY, United States Edited by: Robert C. Lynall, University of Georgia, United States Reviewed by: Michelle Weber Rawlins, San Diego State University, United States; Melissa Kossman, University of Southern Mississippi, United States *Correspondence: Zachary W. Bevilacqua zwbihst@rit.edu 02 6 2023 2023 11 118874117 3 2023 19 5 2023 Copyright © 2023 Bevilacqua and McPherson. 2023 Bevilacqua and McPherson https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. A Commentary on Establishing the college Return to Learn team for concussion: a practical approachconcussion Return to Learn college higher education policy disabilities FERPA multidisciplinary team section-at-acceptancePublic Health Policy ==== Body pmcIntroduction Researchers have only recently begun to investigate Return to Learn (RTL) considerations among college students with concussion. Initial investigations were focused on academic reintegration timeframes (1), faculty and staff knowledge of concussion (2, 3), and impacts on academics (4). Since then, investigations have addressed more precise aims, namely faculty perspectives of RTL and classroom accommodations (5, 6), peer-mentoring programs for students recovering from concussion (7), the Athletic Trainers' role in RTL (8), and factors mediating reading performance (9, 10). Most recently, preliminary consensus recommendations were introduced (11), signifying that college RTL is distinct from K-12, and requires unique attention. Consensus recommendations for college-RTL discuss routine check-ins with a multidisciplinary team (i.e., academic counselor, physician, instructor, and student-advocate) (11). Multidisciplinary teams have become a cornerstone of concussion management, yet a multidisciplinary approach in the college setting inherently encounters barriers, which have yet to be focally addressed in the literature. Specifically, the Family Educational Rights and Privacy Act (FERPA) disallows open communication between medical and academic entities on campus (12); the core characteristic of a RTL partnership. In turn, the onus of maintaining interdisciplinary communication is situated firmly with the student. To be successful, a RTL team must overcome such barriers if they hope to stay informed and properly chaperone students through a RTL protocol. To that end, we will discuss a reasonable approach to establishing communication between college-RTL team members (i.e., medical provider, disability services, faculty, etc.) in order to efficiently facilitate a student's RTL progress. Protocol Literature discusses RTL as both the reintroduction of academic work post-concussion, and a state of recovery. Our protocol will refer to RTL in the former, as we do not define academic recovery, but instead outline RTL proceedings until such a time when they are no longer required. The following approach is comprised of three stages: receiving the diagnosis, activating the team, and follow-up. Receiving the diagnosis Previous literature suggests that symptoms of concussion may persist for longer durations (i.e., >14 days) among student-athletes who receive fewer medical check-ins (13). Factors such as a pre-injury anxiety/mood symptoms (14), symptom severity (15), and sex (16) alter symptom resolution and time away from the classroom as well. If left unsupervised, the effects of such variables may have damaging repercussions on a student's semester and academic progress, especially when considering the pace of learning expected at the collegiate level. College-RTL already demands a mean 18.3 days when under medical care (1), with females requiring longer than males (16). Overall, 26% of college students will require more than 14 days to RTL (n = 1,974) (16), with up to 13% needing ≥35 days (1, 16). Therefore, to attenuate academic time lost to avoidable symptom prolongation, we encourage all students to seek care if a concussion is suspected. For college students, campus health centers are an excellent option, with Athletic Trainers representing the preferred route for student-athletes. If financial accessibility is a concern, the majority of college institutions offer health insurance plans to students that can be financed with financial aid packages. Additionally, the Affordable Care Act allows college-aged (i.e., <26 years old) individuals to remain insured through their parents (17). If barriers persist, Dean of Students Offices or other administrative entities on campus can provide assistance to students seeking concussion care. Whatever the means, obtaining medical supervision is the priority, and represents the first point-of-contact for students with concussion (Figure 1A). Figure 1 RTL team schematic. (A) Receiving the diagnosis. (B, C) Activating the team. (D) Follow-up. Activating the team Medical evaluation provides obvious benefits, but it simultaneously triggers the next step of the RTL process (Figure 1B). Upon diagnosis, providers can offer students the option to sign a FERPA consent form (Figure 2) permitting them with the authority to share a student's diagnosis and treatment plan with the campus' disability office, or similar entity. Currently, the autonomous nature of the college environment often leaves students responsible for disseminating this information on their own, and is the product of FERPA restrictions (12); however, obtaining written approval is quickly accomplished in clinic, and will immediately permit interdisciplinary communication while alleviating students of this burden. FERPA consent could even merge into pre-participation paperwork for student-athletes, affording ongoing permission to Athletic Trainers and other members of the healthcare team. Likewise, initiation of cognitive reintegration is recommended no later than 48 h post-injury (18), so speed with communication is important. Lastly, it is speculated that campus disability offices lack the ability to assist students in a timely manner (11); therefore, integrating this form into clinical practice will allow team members to rapidly select classroom accommodations. In fact, dialogue between provider and disability staff could determine which accommodations will best support clinical findings (i.e., symptoms of photophobia managed by wearing sunglasses in class). Students pursuing accommodations throughout their recovery would likely find convenience in a protocol for authorizing open communication between the professionals responsible for implementing these supports. Therefore, we recommend that FERPA consent forms be considered as a routine tool used to facilitate rapid collaboration between “primary” RTL team members (campus health center and disability offices). Healthcare providers are also encouraged to provide these accommodation details to the student in the form of a discharge note (Figure 1B). This affords students the option to disseminate accommodation details to faculty (Figure 1C), should a student incur a delay in receiving disability assistance. Supplementary Figure 1 presents a sample note that has garnered favor from college faculty (19). Figure 2 Sample FERPA consent form. Adapted from the Office of Legal Affairs at the Rochester Institute of Technology, 2023, https://www.rit.edu/fa/legalaffairs, Copyright Rochester Institute of Technology. Students who have received medical care and are supported by disability staff would be able to expect “secondary” members of the RTL team to receive notification (Figure 1C). Faculty and academic advisors comprise this group, bearing a supportive role (5). In this way, faculty are instructed to adhere to the prescribed classroom supports, whereas academic advisors are apprised to the student's temporary struggle. Follow-up Because students may receive follow-up care, and are anticipated to shed academic restrictions and accommodations as they progress through a RTL progression, routine communication will allow all to remain informed to any changes, and or when provided supports can be discontinued. Figure 1D illustrates this process. Rationale for the progression The present protocol emphasizes receiving medical evaluation and the acquisition of campus disability support for two reasons. First, encouraging all students to receive medical attention follows recommendations to evaluate suspected concussion injury (18), and two, it produces a formal record of the injury to which campus disability offices use to justify “formal” accommodations (i.e., classroom supports that are justified by documentation) (20). Others would oppose our approach, stating that “informal” adjustments (i.e., classroom supports that are not supported by documentation) should be implemented quickly, and through student-faculty collaboration (11). We appreciate that a subset of students could secure informal adjustments through dialog with empathetic faculty (5); yet, this does not appear to be the majority of instructors (5, 6), nor is it assured that these informal adjustments will be entertained throughout the student's entire recovery. Furthermore, a growing body of literature indicates that faculty support is conditional (5–8), and requires disability and or medical documentation prior to implementing the supports that are routinely prescribed post-concussion (i.e., excused absences, extended time on tests/assignments, altered due dates, etc.) (5, 6, 19). In fact, nearly two-thirds (62.4%) of sampled faculty hold this opinion (n = 255) (6). Therefore, and in an effort to ensure equity with academic support, our approach urges the utilization of disability offices to guarantee formal accommodations in and outside of the classroom, and abate any biases faculty may have toward students (21–23). While disability services take a range of time to implement services and most students will RTL within 2 weeks, seeking academic accommodations is worthwhile given the lack of effectiveness of informal adjustments and that some students will require support for more than 35 days. We further encourage students to present instructors with their medical discharge note (Figure 1C), which half (48.6%) of faculty require (6), and respect similarly to disability notes (5). This could formally initiate accommodations prior to disability services reaching out to faculty (Figure 1C), and simultaneously provide an avenue for students who forego FERPA consent or disability offices altogether. Overall, efforts should focus on expediting the provision of formal accommodations, and promote education for students aimed at seeking concussion care. Should a student's recovery not follow a “typical” timeframe, medical records at the time of injury can be used to award a “medical withdrawal” or “incomplete” to a student. Generally speaking, the former allows a student to withdraw from the semester while recuperating a portion of their tuition dollars, while the latter provides the student the ability to complete course requirements the following semester. Both options are expedited by proof of medical necessity, and function as a safeguard of sorts. Subsequent stages of the protocol sought to capitalize on known university procedures to establish a “point-person” and create channels of communication. A point-person has long been discussed as a critical component in a return-to-school effort for pediatric models, with the school nurse often identified for this role (24–26). RTL for college students will inherently look different, as universities do not have a single healthcare professional through which all medical information funnels (i.e., school nurse). Instead, disability offices are perhaps the ideal choice to assume point-person duties, as they currently interact with a student's medical and academic information, and are trained in producing academic orders from medical input. Additionally, they possess the infrastructure to electronically connect with academic team members, including students who are lost to follow-up. They also provide services that can directly assist faculty, such as administering examinations to a concussed student who requires additional time (27, 28). Conclusion The present manuscript discusses a reasonable approach to establishing a RTL team in the college setting. The suggested plan discusses how medical evaluation and FERPA consent can kick start a RTL procedure, unencumbered. RTL authors cite that implementing a RTL protocol could take a mean 17 years if barriers and contextual factors are not addressed beforehand (29, 30). The suggested approach minimizes some of these concerns, utilizing campus infrastructure (health centers, athletic trainers, and disability offices) and processes (disability accommodations and FERPA consent) that have long been entrenched in higher education. Additionally, the medical records that result from the initial evaluation provide a multitude of uses, from justifying disability accommodations, to ameliorating faculty concerns, to safeguarding a student's investment in the semester. Overall, we encourage stakeholders (i.e., National Collegiate Athletic Association, National Athletic Trainers' Association, consensus protocols) to consider the practical and high-yield characteristics of the proposed procedure. Author contributions ZB was responsible for the inception and drafting of the manuscript. ZB and JM contributed to the editing and refinement of the manuscript. Both authors contributed to the article and approved the submitted version. Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher's note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Supplementary material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpubh.2023.1188741/full#supplementary-material Click here for additional data file. ==== Refs References 1. Bevilacqua ZW Kerby ME Fletcher D Chen Z Merritt B Huibregtse ME . Preliminary evidence-based recommendations for return to learn: A novel pilot study tracking concussed college students. Concussion. (2019) 4 :4. 10.2217/cnc-2019-0004 31608152 2. Mokris Kessler A Williams K Ranney J Webster J Stauffer KR . Assessing concussion knowledge and awareness in faculty and staff in a collegiate setting. J Concussion. (2019) 3 :2059700219870920. 10.1177/2059700219870920 3. Finn C . Effects of a faculty training program on knowledge and awareness of student concussion and academic performance. J Am Coll Heal. (2020) 2020 :1–8. 10.1080/07448481.2020.1754220 32407248 4. Holmes A Chen Z Yahng L Fletcher D Kawata K . Return to learn: Academic effects of concussion in high school and college student-athletes. Front Pediatr. (2020) 8 :57. 10.3389/fped.2020.00057 32195210 5. Bevilacqua Z Cothran DJ Rettke DJ Koceja DM Nelson-Laird TF Kawata K . Educator perspectives on concussion management in the college classroom: A grounded theory introduction to collegiate return-to-learn. Br Med J Open. (2021) 11 :e044487. 10.1136/bmjopen-2020-044487 33879486 6. Weber Rawlins ML Johnson RS Schmidt JD Lynall RC O'Brien KH Welch Bacon CE . University educators' perceptions of academic adjustments following a concussion for student-athletes and non-student-athletes. J Am Coll Health. (2022) 2022 :2108324. 10.1080/07448481.2022.2108324 35943951 7. O'brien KH Wallace T Kemp A . Student perspectives on the role of peer support following concussion: Development of the SUCCESS Peer Mentoring Program. Am J Speech Lang Pathol. (2021) 30 :933–48. 10.1044/2020_AJSLP-20-00076 33556263 8. Bowman TG Singe SM Lacy AMP Register-Mihalik JK Pike Lacy AM Register-Mihalik JK . Challenges faced by collegiate athletic trainers, part II: Treating concussed student-athletes. J Athl Train. (2020) 55 :312–8. 10.4085/1062-6050-85-19 31986101 9. Sherry N Trbovich A Holland C Eagle S Bitzer H Kontos AP . Predictors of poor reading performance in student-athletes following sport-related concussion. Appl Neuropsychol Child. (2022) 11 :364–72. 10.1080/21622965.2020.1850449 33428451 10. Broggi M Ready RE . Academic skills, self-perceptions, and grades in university students with a history of multiple concussions: The mediating roles of processing speed and psychological symptoms. Clin Neuropsychol. (2022) 36 :2188–204. 10.1080/13854046.2021.1958924 34402398 11. Memmini AK Popovich MJ Schuyten KH Herring SA Scott KL Clugston JR . Achieving consensus through a modified delphi technique to create the post-concussion collegiate return-to-learn protocol. Sports Med. (2022) 22 :8. 10.1007/s40279-022-01788-8 36396900 12. Services USD of H&H. Joint Guidance on the Application of the Family Educational Rights and Privacy Act (FERPA) And the Health Insurance Portability and Accountability Act of 1996 (HIPAA) To Student Health Records. Washington, DC (2008). p. 6–7. 13. McAllister TW Broglio SP Katz BP Perkins SM LaPradd M Zhou W . Characteristics and outcomes of athletes with slow recovery from sport-related concussion: A CARE Consortium Study. Neurology. (2023) 2023 :206853. 10.1212/WNL.0000000000206853 36653178 14. Lumba-Brown A Teramoto M Zhang R Aukerman DF Bohr AD Harmon K . Multicentre evaluation of anxiety and mood among collegiate student athletes with concussion. Br Med J Open Sport Exerc Med. (2023) 9 :1446. 10.1136/bmjsem-2022-001446 36756287 15. Broglio SP McAllister T Katz BP LaPradd M Zhou W McCrea MA . The natural history of sport-related concussion in collegiate athletes: Findings from the NCAA-DoD CARE Consortium. Sports Med. (2022) 52 :403–15. 10.1007/S40279-021-01541-7 34427877 16. Bretzin AC Esopenko C D'Alonzo BA Wiebe DJ . Clinical recovery timelines following sport-related concussion in men's and women's collegiate sports. J Athl Train. (2021) 2021 :20. 10.4085/601-20 33626145 17. America OHEC of the US of. Patient Protection and Affordable Care Act. Washington, DC (2010). 18. McCrory P Meeuwisse W Dvorak J Aubry M Bailes J Broglio S . Consensus statement on concussion in sport-the 5(th) international conference on concussion in sport held in Berlin, October 2016. Br J Sport Med. (2017) 51 :838–47. 10.1136/bjsports-2017-097699 28446457 19. Bevilacqua ZW Cothran D Rettke D Koceja D Nelson-Laird T Kawata K . Return to learn: Preferences of college educators when receiving concussion medical notes. Neurotrauma Rep. (2022) 3 :185–9. 10.1089/neur.2022.0012 35558728 20. Division USD of JCRi. Final Rule Implementing the ADA Amendments Act of 2008. (2016). Available online at: http://www.ada.gov/regs2016/final_rule_adaaa.html (accessed May 23, 2023). 21. Stein K . Experiences of college students with psychological disabilities: The impact of perceptions of faculty characteristics on academic achievement. Int J Teach Learn High Educ. (2014) 26 :55–65. 22. Mamboleo G Meyer L Georgieva Z Curtis R Dong S Stender LM . Students with disabilities' self-report on perceptions toward disclosing disability and faculty's willingness to provide accommodations. Rehabil Couns Educ J. (2015) 8 :8.31008459 23. Marshak Van Wieren T Raeke Ferrell D Swiss L Dugan CL . Exploring barriers to college student use of disability services and accommodations. J Postsecond Educ Disabil. (2010) 2010 :22.32990478 24. Halstead ME McAvoy K Devore CD Carl R Lee M Logan K . Returning to learning following a concussion. Pediatrics. (2013) 132 :948–57. 10.1542/peds.2013-2867 24163302 25. Fetta J Starkweather A Huggins R Van Hoof T Casa D Gill J . Implementation of return to learn protocols for student athletes with sport and recreation related concussion: An integrative review of perceptions, challenges and successes. J Sch Nurs. (2023) 39 :18–36. 10.1177/10598405211056646 34816759 26. Bradley-Klug KL Garofano J Lynn C DeLoatche KJ Lam GYH . Returning to school after a concussion: Facilitating problem solving through effective communication. Sch Psychol Forum. (2015) 9 :184–98. 27. Affairs IUBD of S. Schedule an Exam. (2023). Available online at: https://studentaffairs.indiana.edu/student-support/disability-services/schedule-exams.html (accessed May 23, 2023). 28. Services SBSA,. Services. (2023). Available online at: https://www2.brockport.edu/life/accessibility-services/services/ (accessed May 23, 2023). 29. Grant J Liz G Mason B . Basic research and health: A reassessment of the scientific basis for the support of biomedical science. Res Eval. (2003) 12 :217–24. 10.3152/147154403781776618 30. Memmini AK Kinnett-Hopkins DL Hasson RE Rifat SF . Considerations for implementing the post-concussion collegiate return-to-learn protocol in the national collegiate athletic association power 5 conferences. J Head Trauma Rehabil. (2023) 2023 :862. 10.1097/HTR.0000000000000862 36854099

PMC010xxxxxx/PMC10307856.txt

==== Front Nat Commun Nat Commun Nature Communications 2041-1723 Nature Publishing Group UK London 37380648 39590 10.1038/s41467-023-39590-3 Article Acute stress induces long-term metabolic, functional, and structural remodeling of the heart http://orcid.org/0000-0003-0525-0407 Yoganathan Thulaciga thulacigayoganathan@calicolabs.com 1 http://orcid.org/0000-0002-4985-247X Perez-Liva Mailyn 12 http://orcid.org/0000-0003-3802-2224 Balvay Daniel 13 http://orcid.org/0000-0002-4935-7065 Le Gall Morgane 4 http://orcid.org/0000-0002-4954-1751 Lallemand Alice 1 Certain Anais 1 http://orcid.org/0000-0001-7335-350X Autret Gwennhael 13 http://orcid.org/0000-0002-1064-3881 Mokrani Yasmine 1 http://orcid.org/0000-0003-1484-4696 Guillonneau François 5 http://orcid.org/0000-0002-1522-9604 Bruce Johanna 4 http://orcid.org/0000-0002-6695-7943 Nguyen Vincent 6 Gencer Umit 7 http://orcid.org/0000-0002-8011-2277 Schmitt Alain 8 http://orcid.org/0000-0001-7748-9754 Lager Franck 9 http://orcid.org/0000-0001-5069-0730 Guilbert Thomas 10 Bruneval Patrick 1 http://orcid.org/0000-0002-9136-7726 Vilar Jose 1 http://orcid.org/0000-0002-7133-7220 Maissa Nawal 1 http://orcid.org/0000-0002-8076-1445 Mousseaux Elie 7 http://orcid.org/0000-0002-0756-9166 Viel Thomas 13 http://orcid.org/0000-0003-2273-1229 Renault Gilles 9 http://orcid.org/0000-0002-5269-3246 Kachenoura Nadjia 6 http://orcid.org/0000-0002-5349-8194 Tavitian Bertrand bertrand.tavitian@inserm.fr 137 1 grid.462416.3 0000 0004 0495 1460 Université Paris Cité, Inserm, PARCC, F-75015 Paris, France 2 grid.4795.f 0000 0001 2157 7667 Nuclear Physics Group and IPARCOS, Department of Structure of Matter, Thermal Physics and Electronics, CEI Moncloa, Universidad Complutense de Madrid, 28040 Madrid, Spain 3 grid.462416.3 0000 0004 0495 1460 Université Paris Cité, Plateforme d’Imageries du Vivant, PARCC, F-75015 Paris, France 4 grid.462098.1 0000 0004 0643 431X Université Paris Cité, P53 proteom’IC facility, Institut Cochin, INSERM, CNRS, F-75014 Paris, France 5 grid.418191.4 0000 0000 9437 3027 Institut de Cancérologie de l’Ouest, CNRS UMR6075 INSERM U1307, 15 rue André Boquel, F-49055 Angers, France 6 grid.462844.8 0000 0001 2308 1657 Sorbonne Université, Laboratoire d’Imagerie Biomédicale, Inserm, CNRS, F-75006 Paris, France 7 grid.414093.b 0000 0001 2183 5849 Service de Radiologie, AP-HP, hôpital européen Georges Pompidou, F-75015 Paris, France 8 grid.462098.1 0000 0004 0643 431X Université Paris Cité, Cochin Imaging, Electron microscopy, Institut Cochin, INSERM, CNRS, F-75014 Paris, France 9 grid.462098.1 0000 0004 0643 431X Université Paris Cité, Plateforme d’Imageries du Vivant, Institut Cochin, Inserm-CNRS, F-75014 Paris, France 10 grid.462098.1 0000 0004 0643 431X Université Paris Cité, Cochin Imaging Photonic, IMAG’IC, Institut Cochin, Inserm, CNRS, F-75014 Paris, France 28 6 2023 28 6 2023 2023 14 383517 5 2022 21 6 2023 © The Author(s) 2023, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Takotsubo cardiomyopathy is a stress-induced cardiovascular disease with symptoms comparable to those of an acute coronary syndrome but without coronary obstruction. Takotsubo was initially considered spontaneously reversible, but epidemiological studies revealed significant long-term morbidity and mortality, the reason for which is unknown. Here, we show in a female rodent model that a single pharmacological challenge creates a stress-induced cardiomyopathy similar to Takotsubo. The acute response involves changes in blood and tissue biomarkers and in cardiac in vivo imaging acquired with ultrasound, magnetic resonance and positron emission tomography. Longitudinal follow up using in vivo imaging, histochemistry, protein and proteomics analyses evidences a continued metabolic reprogramming of the heart towards metabolic malfunction, eventually leading to irreversible damage in cardiac function and structure. The results combat the supposed reversibility of Takotsubo, point to dysregulation of glucose metabolic pathways as a main cause of long-term cardiac disease and support early therapeutic management of Takotsubo. Takotsubo disease, a stress induced cardiomyopathy mimicking acute coronary syndrome, increases the risk of heart failure and cardiac death. The authors show here that heart function and structure keep on deteriorating continuously after a single acute stress, this snowball effect being triggered by abnormalities incardiac metabolism. Subject terms Cardiovascular biology Imaging Cardiovascular diseases https://doi.org/10.13039/100010661 EC | Horizon 2020 Framework Programme (EU Framework Programme for Research and Innovation H2020) MAESTRIA #965286 Tavitian Bertrand https://doi.org/10.13039/100010665 EC | EU Framework Programme for Research and Innovation H2020 | H2020 Priority Excellent Science | H2020 Marie Skłodowska-Curie Actions (H2020 Excellent Science - Marie Skłodowska-Curie Actions) 101030046 Tavitian Bertrand https://doi.org/10.13039/501100001665 Agence Nationale de la Recherche (French National Research Agency) ANR-11INBS-0006 PACIFIC ANR-18-CE14-0032 Tavitian Bertrand Infrastructures Biologie Santé IBiSa AVIESAN ASCP2001SSCP Région Ile de France SESAME Région Ile de France DIM Thérapie génique LABEX GR-Ex SIRIC CARPEMissue-copyright-statement© Springer Nature Limited 2023 ==== Body pmcIntroduction Stress is an independent risk factor for cardiovascular diseases in patients with pre-existing cardiac diseases or cardiovascular risk factors, as well as in persons without a known cardiac condition1,2. A spectacular consequence of acute stress on the heart is Takotsubo cardiomyopathy (TTC), a condition mimicking an acute coronary syndrome but in which coronarography appears normal3. In the absence of other signs of associated or underlying cardiac conditions, TTC patients are dismissed from hospitals because the acute cardiac symptoms (pain, electrocardiography, heart rate, blood pressure…) are reversible and because there is no evidence of coronary obstruction. However, recent long-term follow-up studies have reported similar cardiovascular annual death rates for TTC and for proven acute coronary obstruction4,5, suggesting that TTC often induces severe cardiac sequelae. Although there is converging evidence that the stress-induced catecholamine rush is the cause of Takotsubo6, the mechanism by which a single acute stressing event can have long-term deleterious effects on the heart remains mysterious. Catecholamines activate adrenoreceptors, in particular the beta-1 adrenoreceptors that are predominant in the heart. The activation of beta1 and beta2 receptors has positive inotropic and chronotropic effects, increases cardiac output, myocardial oxygen consumption and coronary flow7, while the activation of beta 3 receptors has a negative inotropic effect. Studies have reported that at high catecholamine concentrations, such as those found during acute stress, the activation of beta 3 receptors counteracts the activation of beta 1 and 2 receptors8, supporting the view that, mechanistically, the heart is well adapted to stress. Another role of stress-induced catecholamine release is to secure energy substrates for the “fight or flight” reaction, by opposing the action of insulin, namely by inducing hyperglycemia and glycogenolysis. The mammalian heart is omnivorous: in basal conditions, it feeds essentially on fatty acids (ca. 80%) and modestly on glucose (ca. 20%) and other substrates9. In conditions of high-energy demand, e.g., exercise, the supplementary energy requirements are provided by an increase of aerobic and anaerobic glucose breakdown. This is also the case under high glycaemia in diabetes, or in conditions with reduced cardiac muscle microperfusion such as under anti-angiogenic treatment10,11. The adaptation of the cardiac energy balance through the regulation of glucose utilization is advantageous for the challenged heart because it can be switched on very rapidly and can maintain ATP production in low oxygen conditions. The mechanisms involve increased intracellular transport of glucose through Glut1, the universal glucose membrane transporter, and of Glut4 that is expressed in the heart, skeletal muscle, and insulino-dependant tissues. Glut4 translocation to the plasmatic membrane is under the control of insulin growth factor (IGF) and other effectors, including the stress hormones catecholamines and cortisol. Although glucose metabolism is a major factor of regulation of cardiac activity in health and disease12, molecular imaging of metabolism is not an indication in patients with Takotsubo. One of the reasons may be that the metabolic responses of the heart to a sudden and dramatic rise of circulating catecholamines are not easily predictable and depend on the respective affinities and sensitivities of different adrenergic receptors and on their cardiac densities, as well as on the physiological and metabolic state of the heart, which is highly variable among individuals13. Moreover, such studies are difficult to conduct in human patients because the neurohumoral response to stress is highly variable and unpredictable among individuals, and because cardiac and metabolic comorbidities are frequent in Takotsubo patients14. In contrast, studies of TTC in animal models are easy to run in homogeneous populations using uniform stress triggers. Several Takotsubo-like models have been proposed in rodents, the more robust and reproducible being based on the administration of catecholamines in different routes, frequency and dose15,16. Considering that little is known about the regulation of glucose metabolism of the myocardium during the early and late phases of Takotsubo, and, more importantly, that even less is known on the influence of metabolism on the functional and structural integrity of the myocardium in the long term after Takotsubo, we performed a comprehensive set of non-invasive explorations for close longitudinal follow-up in the same animals before stress (baseline), during acute stress (2 h) and at early (7 days) and late (1 and 3 months) time points after stress. Our objective was to decipher the role of energy metabolism on long-term tissue and vascular remodeling17,18 in a preclinical model recapitulating the clinical signs of TTC over the whole course of the disease. Reasoning that acute stress was the cause of Takotsubo, at least in its initial description6, and that there is a large consensus that stress-induced catecholamine release is the main causal factor of Takotsubo, we mimicked in rats the catecholamine rush described in patients19, with a single administration of the adrenergic agonist isoproterenol, a drug used to increase heart rate in case of bradyarrhythmias15 that can provoke Takotsubo20. Here, we show that the increased glucose uptake in the myocardium induced by an isoprenaline stress is not used to increase energy production, but is diverted into alternative anabolic pathways of glucose. We provide evidence that this glucose diversion induces immediate and long-term tissue, metabolic and functional changes that may explain the increased risk of heart failure in Takotsubo patients. Results Acute pharmacological challenge recapitulates Takotsubo In young adult female rats, one single intraperitoneal injection of the catecholaminergic drug isoprenaline (ISO; 50 mg/kg, a non-lethal dose) induced typical signs of acute stress. Glycaemia, cardiac and respiratory rates were significantly increased 2 h post-ISO (Supplementary Fig. 1). As expected, biomarkers of the acute catecholamine surge were present13,21, including increased plasmatic concentrations of glycerol, triglyceride, non-esterified fatty acids and lactate (Supplementary Fig. 1). Glycogen deposits significantly decreased in both apical (−64%) and basal regions (−57%) (Supplementary Fig. 2), while lipid deposits significantly increased in the apex (threefold) and in the base (11-fold) of the heart (Supplementary Fig. 2). The plasmatic concentrations of Alanine-Amino Transferase (ALAT), Aspartate-Amino Transferase (ASAT) and creatinine kinase (CK) all increased 2 h post-ISO, reflecting an acute myocardial damage. The plasmatic concentration of brain natriuretic peptide (BNP) was unchanged during the acute stress phase (Supplementary Fig. 1). During the acute phase, electrocardiograms (ECG) of the heart were characteristic of impaired ventricular contraction and myocardial dysfunction: significant shortening of the RR interval concomitant with an increased heart rate, a substantial prolongation of the QRS interval, and a highly variable S wave amplitude with prolongation of the ST segment (Supplementary Fig. 1). High-resolution ultrasound showed increased left ventricular (LV) ejection fraction and fractional shortening along with notably reduced end-diastolic and end-systolic LV volumes. Systolic arterial blood pressure and cardiac output were low, and systemic arterial resistances were high, although not statistically different from pre-stress values (Supplementary Fig. 3). The original description of TTC in patients 30 years ago3 reported a spectacular apical ballooning of the LV shaping the heart into a Japanese octopus fishing pot (called a tako-tsubo in Japanese). Nowadays it is recognized that not all TTC hearts show apical ballooning and that some patients may present medial, basal or global cardiac ballooning, or even no visible ballooning of the LV14. In rodents with fast-beating hearts, LV ballooning is difficult to evidence. However, in our rat model a quantitative analysis of cardiac magnetic resonance (CMR) cine sequences enabled global and segmental measurement of strains in the three myocardial fiber orientations (Fig. 1). Two hours post-ISO, longitudinal strain (LS) increased in the LV and decreased in the right ventricle (RV). As both ventricles share the same interventricular septum, these observations suggest hypercontractility of the LV wall and especially a reduction of the deformation of the RV free wall22 (Fig. 1). In the right atrium (RA), the strain decreased, reflecting RA dysfunction (Fig. 1). Interestingly, in the LV both the longitudinal strain (LS) and the circumferential strain (CS) varied regionally: both significantly increased in the basal segment of the LV whereas they remained unchanged or decreased in the apical segment (Fig. 1). Taken together, myocardial deformation during the acute phase reflects simultaneous hypercontractility of the basal/mid segments of LV and hypocontractility of the apex, in line with the transitory LV apical ballooning observed in TTC patients23,24.Fig. 1 Cardiac strains in LV, RV, LA, and RA at baseline and 2 h, 7 d, 1 mo, and 3 mo post ISO. Boxplots showing median, 25 and 75 percentiles, and extremes of values of n = 4 to 6 animals per time point. a Longitudinal strain is increased in the LV at 2 h and returns to normal in the remaining time-points (0 vs. 2 h: p = 0.0428, 2 h vs. 7d: p = 0.0379, 2 h vs. 1mo: p = 0.0175, and 2 h vs. 3mo: p = 0.0385). In contrast, longitudinal strain in the RV decreases at 2 h and then returns progressively to baseline values (0 vs. 2 h: p = 0.0002). b Longitudinal and circumferential strains in the apex and in the basal region. Hyperkinesia, the normal reaction to a catecholaminergic surge is observed at 2 h in the basal region highlighted by the increase in the basal longitudinal (0 vs. 2 h: p = 0.0081) and circumferential (0 vs. 2 h: p = 0.0107, and 2 h vs. 3mo: p = 0.0185) strains, while at the same time no substantial changes were found in the longitudinal and circumferential strains in the apex. c LA and RA strains and LA surface area. LA strain decreases progressively after the initial rise at 2 h, suggesting an irreversible LA dysfunction (0 vs. 3mo: p = 0.0098, 2 h vs. 7d: p = 0.0038, 2 h vs. 3mo: p < 0.0001, and 1mo vs. 3mo: p = 0.0203). It goes along with the increase of LA minimum (0 vs. 2 h: p = 0.0447, and 2 h vs. 3mo: p = 0.0357) and maximum (2 h vs. 7d: p = 0.0038 and 2 h vs. 3mo: p = 0.0031) surface areas over time. The strain that corresponds to the reservoir function decreases in the RA at 2 h (2 h vs. 1mo: p = 0.0310, and 2 h vs. 3mo: p = 0.0376). Unpaired comparison tests: *p < 0.05, **p < 0.01 and ***p < 0.001. All values are individually normalized to the baseline value for each animal. Statistical significance (p < 0.05) for each variable was estimated by one-way or two-way ANOVA when group variances were equal (Bartlett test); if not the non-parametric Kruskall–Wallis test, and the Holm multiple comparisons test was used to execute simultaneous t-tests. Source data are provided as a « SourceData_Figure1 » file. The main sign for differential diagnosis between TTC and acute coronary syndrome is the absence of coronary obstruction visible on coronarography in TTC patients. In rodents, coronary obstruction can be assessed by measuring the difference in the coronary flow (CF) under basal and hyperemic conditions (CF during hyperemia – CF pre-hyperemia)/ CF pre-hyperemia), which corresponds to the coronary flow reserve (CFR, Supplementary Fig. 4). Two hours post-ISO, the peak velocity of coronary flow was increased with respect to pre-ISO (baseline). However, at that time, it was no further increased by the inhalation of 5% isoflurane (hyperemic enhancement) in contrast to the pre-stress response to hyperemia (Supplementary Fig. 4). The CFR was lower during the acute post-stress phase and recovered its baseline level at 7 days post-ISO. Therefore, the vasodilatation induced by isoprenaline was already maximal at 2 h post-ISO, which confirms the absence of coronary obstruction during the acute post-stress phase. Ultra-structurally, transmission electron microscopy revealed discrete signs of LV tissue remodeling at 2 h post-ISO reflecting the mechanical stress induced by the hypercontractility of the LV wall: mainly signs of disruption of intercalated discs and loss of cell–cell adhesion (Supplementary Fig. 5). The Takotsubo cardiomyopathy is partly reversible All the signs of acute cardiac stress observed at 2 h post-ISO were re-analyzed at 7d post-ISO. Glycaemia, cardiac and respiratory rates, blood pressure, global and segmental cardiac strains, and strain rates of the LV and RV returned to pre-ISO (baseline) levels (Supplementary Figs. 1 and 3, Fig. 1). The ECG and the plasmatic levels of ALAT, ASAT and CK returned to baseline; however, the plasmatic concentration of brain natriuretic peptide (BNP) was normal at 2 h post-ISO and increased at 7d post-ISO, suggesting a delayed LV dysfunction (Supplementary Fig. 1). Parameters derived from cardiac ultrasound images were normal at 7d post-ISO (Supplementary Figs. 3 and 4). In short, the major parameters of cardiac function and physiology had reversed to normal values, as typically described in TTC patients. Ultrastructural signs reflecting mechanical stress were not found at 7 days post-ISO, however fibrosis appeared in the LV apex (Fig. 2, Supplementary Figs. 5 and 6) but not in the LV base or in atria (Fig. 2). Taken together, these results on animal model strongly corroborate the clinical signs of Takotsubo, including the reversibility of acute signs of cardiac stress, although it cannot be excluded that diffuse myocardial fibrosis is a potential primer of cardiac dysfunction.Fig. 2 Progression of cardiac fibrosis after ISO stress. a–d Representative section of Sirius red staining in the LV apex and base, right atrium and left atrium of the heart, at baseline (scale bars of a length of 100 μm) and 2 h (scale bars of a length of 500 μm in LV images and of 100 μm in LA and RA images), 7d (scale bars of a length of 100 μm) and 1-month (scale bars of a length of 500 μm in LV apical and 250 μm in LV basal images, and of 100 μm in LA and RA images,) and 3-months (scale bars of a length of 500 μm in LV images and of 100 μm in LA and RA images) post-ISO. e–h Quantitative analysis of staining using FIBER-ML71 in the LV apex and base, right atrium and left atrium of the heart at the corresponding post-ISO time points for the indicated number of animals represented as boxplots showing median, 25 and 75 percentiles, and extremes of values. The diffuse fibrosis that first appears in the LV apex at 7d post-ISO augments and extends into the LV base at 1 and 3 months (in the apex, 0 vs. 7d: p = 0.0353, 0 vs. 1mo: p = 0.0011, 0 vs. 3mo: p = 0.0001, 2 h vs. 1mo: p = 0.0073, and 2 h vs. 3mo: p = 0.0012; in the base, 0 vs. 1mo: p = 0.0059, 0 vs. 3mo: p = 0.0015, 2 h vs. 1mo: p = 0.0038, 2 h vs. 3mo: p = 0.0009, 7d vs. 1mo: p = 0.0043, and 7d vs. 3mo: p = 0.0005); unpaired comparison tests: *p < 0.05, **p < 0.01 and ***p < 0.001. Statistical significance (p < 0.05) for each variable was estimated by one-way or two-way ANOVA when group variances were equal (Bartlett test); if not the non-parametric Kruskall–Wallis test, and the Holm multiple comparisons test was used to execute simultaneous t-tests. Diffuse fibrosis is also apparent in the LA at 1- and 3-months post-ISO: ANOVA test, p = 0.0294. Source data are provided as a « SourceData_Figure5 » file. Acute stress induces immediate metabolic remodeling Two hours post the acute administration of ISO, positron emission tomography (PET) imaging showed a 47% increase in myocardial uptake of the glucose analogue 2’-[18F]-fluoro-2’-deoxy-D-glucose (FDG) relatively to baseline (Fig. 3). A two-tissue compartmental analysis of FDG uptake kinetics25 showed that K1, the rate constant for FDG passage from blood to tissue, increased by 70% at 2 h post-ISO. However, k3, the constant that reflects phosphorylation of FDG by hexokinase, was reduced. Staining for Glut1 significantly increased 2 h post-ISO, with a slight predominance in the apical over the basal region of the heart, while staining for Glut4 was unchanged (Fig. 4).Fig. 3 Longitudinal FDG PET before, during and after ISO stress. a Representative images of FDG PET registered UUI (PETRUS) at the indicated time points post-ISO from diastolic phase to systolic phase. Images acquired 30 min after FDG injection of one section along the long axis of the LV. Color scale depicts the Standard Uptake Value (SUV) from 0 to 10. Note the increase in FDG uptake at 2 h and 7d respective to baseline. b Data are presented as mean values +/− SD of n = 9 animals per time point. Quantitative analysis of FDG uptake in the LV: SUV mean in whole LV, apex (7d vs. 1mo: p = 0.0337) and basal LV, normalized to baseline values; rate constant K1 reflecting the exchange of FDG from blood to tissue; rate constant k3 reflecting FDG phosphorylation, calculated using two-compartmental analysis. Note the global increase of SUV at 2 h and 7d post-ISO, the modest increase of K1 that is not statistically significant from baseline at any time point, and the significant decrease in k3 at 2 h followed by an increase at 7d post-ISO (in the apex, 2 h vs. 7d: p = 0.0004, 7d vs. 1mo: p = 0.0005, and 7d vs. 3mo: p = 0.0029; in the base, 0 vs. 2 h: p = 0.0312, 2 h vs. 7d: p = 0.0096, and 7d vs. 1mo: p = 0.0312). Paired comparison tests: *p < 0.05, **p < 0.01 and ***p < 0.001. Statistical significance (p < 0.05) for each variable was estimated by one-way or two-way ANOVA when group variances were equal (Bartlett test); if not the non-parametric Kruskall–Wallis test, and the Holm multiple comparisons test was used to execute simultaneous t-tests. Source data are provided as a « SourceData_Figure2 » file. Fig. 4 Immunohistochemistry of Glut1, Glut4 and Cd68 expression at baseline, 2 h and 7d post-ISO. a–f Representative fields of view of 4 µm cardiac sections stained for Glut1, Glut4 and Cd68. a–c apex, d–f base; the three stainings were performed in each section (OPAL® technology). g–l Quantification of staining densities (medians, 25 and 75 percentiles, and extremes of values of the percentage of section surface staining for each protein) of 4 to 7 animals for Glut1, Glut4 and Cd68. g, h apex, j-l: base. Note the significant increase in Glut1 expression at 2 h post-ISO followed by a return to baseline values in the apex (0 vs. 2 h: p = 0.0307) and the base (0 vs. 2 h: p = 0.0020, and 2 h vs. 7d: p = 0.0050). In contrast, Glut4 expression is unchanged from baseline at 2 h post-ISO but increases at 7d in the apex (0 vs. 7d: p = 0.0196, and 2 h vs. 7d: p = 0.0362). Disperse Cd68+ inflammatory cells are observed at 2 h (apex + basal, in the base: 0 vs. 2 h: p = 0.0063) and 7d (apex). Note that Glut1 and Glut4 stain essentially myocardial cells. Unpaired comparison tests: *p < 0.05, **p < 0.01 and ***p < 0.001. Statistical significance (p < 0.05) for each variable was estimated by one-way or two-way ANOVA when group variances were equal (Bartlett test); if not the non-parametric Kruskall–Wallis test, and the Holm multiple comparisons test was used to execute simultaneous t-tests. Indicated scale bars in the images correspond to a length of 40 μm. Source data are provided as a « SourceData_Figure3 » file. This increase of the entry of glucose into myocardial cells occurred simultaneously with a decrease in the rate of glucose phosphorylation, with lipid deposits and with low glycogenolysis demonstrated using oil-red and periodic acid Shiff stainings, respectively. Similar observations were reported26,27 and correspond to an immediate myocardial metabolic response28 that goes along with the catecholamine-induced heart transient dysfunction15. A metabolic paradox: more glucose-6-phosphate with lower glycolysis During the recovery stage (7d post -ISO), Glut1 staining was high in the apex but returned to baseline in the basal region, as well as Glut4 staining. However, a potential increase in Glut4 expression level was observed in the LV base at 7d compared to the baseline (+107%). FDG uptake and K1 remained at their acute phase levels, but remarkably, k3, which was below baseline at 2 h, rose to higher than baseline levels at 7d, in the apex: +34%, and in the base +44% (Fig. 3). In line with this observation, the expression of hexokinase 2 (Hk2) increased in the apex at 7d post-ISO (Fig. 5). Of note, the levels of the inflammatory cell biomarker Cd68 at 7d post-ISO were modest (0.1% in the apex and 0.04% in the base), and not significantly different from the 2-hour acute phase value, and Glut1 and 4 stained essentially cardiomyocytes (Fig. 4), which is not in favor of a major contribution of inflammatory cells to cardiac FDG uptake (Fig. 4). Taken together, these results indicate a switch in glucose metabolism during recovery from stress, that leads to myocardial accumulation of glucose and of its phosphorylated form, glucose-6-phosphate (G6P), mainly in the LV apex.Fig. 5 Expression of major proteins of the different pathways of glucose metabolism in the heart apex. a Hierarchical clustering with all proteins of interest involved in different pathways of glucose metabolism in the heart apex, based on Pearson correlation on z-scored quantification values. Heatmap indicating proteins involved in glycolysis, glycogenesis, gluconeogenesis, oxidative phosphorylation, hexosamine biosynthetic pathway, and polyol pathway in the LV apical segment. Colors depict the z-scored changes in protein expression at 2 h and 7d post-ISO respective to baseline expression levels: red, overexpression higher than 1.3-fold, green: under-expression lower than −1.3-fold. All values are statistically significant with p < 0.05 using two-tailed Student’s t-test, n = 5 per time point. b Western blot analysis of Gfat1 (0 vs. 7d: p = 0.0168, and 2 h vs. 7d: p = 0.0238), Gfat2 (0 vs. 2 h: p = 0.0486, and 0 vs. 7d: p = 0.0112) represented in boxplots showing median, 25 and 75 percentiles, and extremes of values (a.u.: arbitrary units) depicted a significant increase of their expression at 2 h and 7d post ISO. The levels of O-GlcNAcylated proteins (i.e., the effect of the HBP overactivation), were significantly increased at 7d post-ISO compared to the control groups (p = 0.0118); unpaired comparison tests: *p < 0.05, **p < 0.01 and ***p < 0.001. Statistical significance (p < 0.05) for each variable was estimated by one-way or two-way ANOVA when group variances were equal (Bartlett test); if not the non-parametric Kruskall–Wallis test, and the Holm multiple comparisons test was used to execute simultaneous t-tests. Source data are provided as a « SourceData_Figure4_WB » file. At 7d post-ISO, proteomic analyses indicate inactivation of glycolysis and oxidative phosphorylation and over-activation of glucose alternative pathways, namely the hexosamines biosynthetic (HBP) and polyol pathways, implying a diversion of glucose to anabolic pathways that involves only 3–5% of glucose under normal physiological conditions29. Indeed, the main proteins involved in glycolysis and oxidative phosphorylation were downregulated and large amounts of myocardial G6P were diverted to alternative pathways (Fig. 5, Supplementary Table 1). Contrasting with increased glucose phosphorylation (increased HK2 and k3), proteomic analysis at 2 h post-ISO revealed reduced expression of Gapdh, Pfkm, and Pgam2 (Fig. 5). At the same time, there was an activation of the HBP (overexpression of Pgm3, Uap1l1, Ogt, and Oga), and of the polyol pathway (overexpression of the rate-limiting enzyme aldose reductase, and of sorbitol dehydrogenase) (Fig. 5). Interestingly, the expression of Gfat1, the limiting enzyme of the HBP, also increased (+74%) at 7d post-ISO in the apical region, as compared to baseline (Fig. 5). Gfat2 has recently been reported to mediate cardiac hypertrophy in mice subjected to chronic ISO stress (one-week infusion at 15 mg.kg−1.day−1)30. In rats, Gfat1 is the primary cardiomyocyte isoform responsible for stress-induced protein O-GlcNAcylation while Gfat2 is only present in cardiac fibroblasts31. These observations were confirmed by western blot quantifications of the rate-limiting enzymes of the HBP, Gfat1 and Gfat2, which were highly expressed (respectively +74% and +469%) at 7d post-ISO in the apex compared to the control group and by increased O-GlcNAc levels (Fig. 5b), suggesting an overactivation of the HBP in both cardiomyocytes and myofibroblasts. The switch in myocardial glucose metabolism may have important consequences for myocardial structure and function. In parallel, proteomic analyses point to the overactivation of certain signaling pathways involved in myocardial tissue and vascular remodeling. Accordingly, we observed dysregulation of regulatory cascades known to be induced by O-GlcNAcylation32–35, such as phosphatidylinositol 3-kinase (Pi3k)/Akt, Pkc, Ampk, Mapk, p38, Nf-kB, and insulin receptor substrate (Irs-1, Irs-2) (Supplementary Table 1) with, in parallel, activation or inactivation of canonical pathways of tissue remodeling, in particular the inactivation of Hippo signaling (Dlg1, Ppp1cb, Ppp1cc, Ppp2r3a, Pp2r5e, Skp1, Ywhae, Ywhaq) (Supplementary Table 2). Hippo signaling induces the transition of cardiomyocytes into myofibroblasts, is essential in organ size control and tissue homeostasis and is directly regulated by HBP in response to high glucose uptake36,37. In addition, Hippo signaling maintains cardiac fibroblasts in a resting state, and its inactivation switches fibroblasts to active myofibroblasts that trigger fibrosis38. We also observed increased Rho signaling, a regulator of actin-based motility, increased sphingosine-1-phosphate signaling that modulates vascular tone, endothelial function and integrity, as well as lymphocyte trafficking; and increased Vegf signaling that triggers neo-angiogenesis (Supplementary Table 3). Accordingly, immunofluorescent staining for the endothelial proliferation marker Cd31 was significantly increased, as well as that of the smooth muscle cell activation marker alpha Sma. In the LV, Cd31 and alpha-Sma remained in proximity as the capillary density increased, suggesting coordinated vascular growth (Fig. 6). Taken together, these results support that HBP hyperactivation induced by an acute stress is a driver of cardiac fibrosis, dysfunction, and angiogenesis.Fig. 6 Post-stress vascular remodeling of the LV. a Representative 4 µm-section of the LV apex co-immunostained for Cd31 and alpha-Sma. The scale bar in the main image represents a length of 50 μm, while the scale bar in the two zoom-in images corresponds to a length of 20 μm. Quantitative analysis of Cd31 and alpha-Sma expression’s rate represented as boxplots showing median, 25 and 75 percentiles, and extremes of values. No dissociation between the Cd31-stained endothelial layer and the alpha-Sma-stained smooth muscle cell layer was found. b Increase in Cd31 expression (in the LV, 0 vs. 7d: p = 0.0013, and 2 h vs. 7d: p = 0.0318; in the apex, 0 vs. 7d: p = 0.0318, and 2 h vs. 7d: p = 0.0300; in the base, 0 vs. 7d: p = 0.0011, and 2 h vs. 7d: p = 0.0162) and in the ratio of capillaries per cardiomyocyte (2 h vs. 7d: p = 0.0003) and c in alpha-Sma (in the LV, 0 vs. 7d: p = 0.0004, and 2 h vs. 7d: p = 0.0209; in the apex, 0 vs. 7d: p = 0.0221, and 2 h vs. 7d: p = 0.0369) at 7d post-ISO indicates endothelial proliferation and angiogenesis. Unpaired comparison tests: *p < 0.05, **p < 0.01 and ***p < 0.001. Statistical significance (p < 0.05) for each variable was estimated by one-way or two-way ANOVA when group variances were equal (Bartlett test); if not the non-parametric Kruskall–Wallis test, and the Holm multiple comparisons test was used to execute simultaneous t-tests. Source data are provided as a « SourceData_Figure6 » file. In parallel with hyperactivation of the polyol pathway in the apex at 7d post-ISO, proteomic analysis showed a decrease in the antioxidant glutathione pathways and an increase in the generation of nitrite oxide and reactive oxygen species (Fig. 5, Supplementary Table 2). The activation of these two pathways has been regarded as a consequence of the hyperactivation of the polyol pathway, generating fructose whose accumulation leads to the production of polyols and harmful metabolites such as advanced-glycation end products (AGEs)39,40. Long-term TTC heart remodeling: metabolism, structure, and function Long-term observations at 1- and 3-months post-ISO confirmed the observations at 7d post-ISO (Supplementary Fig. 3). The most striking finding was the reinforcement of fibrosis and its extension to other regions of the heart. At 7d post-ISO, diffuse Sirius red staining concerned only the LV apex, but at 1mo it was found in the basal LV and the LA, where it dramatically increased at 3mo post-ISO (Fig. 2). This was in line with the gradual decrease of the longitudinal and circumferential strains in the apex, and with decreased LA strain and increased minimum and maximum surfaces of the LA (Fig. 1). Taken together, these results indicate a continuous tissue remodeling accompanying a gradual extension of fibrosis in the LA during the delayed post-stress phase, leading to irreversible degradation of the LA reservoir function and to LA enlargement, underlying biomarkers of diastolic dysfunction of the heart41,42. Similarly, PAS staining and FDG PET imaging indicate persistent metabolic remodeling in the long term (Fig. 3 and Supplementary Fig. 2a). Considering that isoprenaline has an elimination half-life of 3–7 h, its plasmatic levels are negligible after one day. Thus, acute effects of isoprenaline inducing the metabolic, functional, vascular and tissue remodeling during the acute and early phases (2 h and 7d post-ISO) are crucial for later aggravation of the disease. Discussion Takotsubo cardiomyopathy, induced by a variety of stress factors, is multiform in its clinical manifestations and cardiac localizations, and mostly concerns patients in the second half of their life, at a time when known or unknown comorbidities are often present14. While resorting to a Takotsubo-like small animal model is a reductionist approach, it offers fast and easily reproducible explorations in a homogeneous population for repeated longitudinal examinations in the same individual of the disease’s natural history. This opens a unique window on the time course of progression. In the acute phase (2 h post-ISO), all the animals showed typical signs of stress-induced cardiomyopathy, with ECG abnormalities and decreased blood pressure, heterogeneity in regional deformations of the LV, and glucose metabolism remodeling. Interestingly, quantitative compartmental analysis of FDG cardiac kinetics suggested that, during the acute phase, the global increase in cardiac uptake of FDG was the consequence of increased FDG entry, with increased rate constant K1 and increased Glut1 expression. While enhanced myocardial glucose uptake is expected in ISO-induced hyperglycemic conditions, surprisingly myocardial insulin-dependent Glut4 presented no changes in terms of its expression rate and translocation to the surface membrane. This observation goes along with the decreased level of PAS staining for glycogens, absence of fatty acids transporters, Fabp, expression’s changes, and lipid toxicity shown by the oil-red staining, indicating that compensation of energetic needs by fatty acid beta-oxidation, essentially a mitochondrial process, is highly unlikely. After acute ISO challenge, the energetic status of the heart resembles myocardial glucose intolerance, reported in diabetic hearts by Stratmann et al., in which lipid toxicity is caused by the myocardial decrease of glucose uptake and causes insulin resistance with decreased Glut4 activity28. Therefore, the heart under acute stress is in a situation of high glucose plasmatic concentrations that do not increase the capacity to use glucose as energy source through glycolysis. This paradoxical unbalanced situation leads to consider whether G6P could be diverted into alternative pathways of glucose metabolism. Most of the signs and symptoms typical of myocardial stress present during the acute phase revert during the recovery phase 7d post-ISO. Functional signs and, except BNP, plasmatic biomarkers of heart suffering return to normal at 7d post-ISO, and plasmatic concentrations of energetic substrates are not different from those prior to stress. Overall, the general state of the animals returns to normal, as in patients during the recovery phase23, although minor functional signs such as strain of the LA remain abnormal. It is plausible that in a clinical setting, these parameters would be disregarded as being minor sequelae of the transient Takotsubo cardiomyopathy. Cardiac metabolism has rarely been explored in TTC patients. Here, we document a high uptake of FDG, which has been considered a sign of cardiac suffering11, high expression of Glut1 and Glut4 and increased glucose phosphorylation (k3) in the apex, together with paradoxically reduced energetic breakdown of glucose in glycolysis and oxidative phosphorylation. Overall, isoprenaline acting on ß1 adrenoreceptors creates a situation in which glycolysis is uncoupled from myocardial glucose levels43. In a rat model of TTC, Godsman et al. have described a dysregulation of glucose and lipids metabolic pathways, as well as inflammation and upregulation of remodeling pathways and fibrosis43. Our present results show the deviation of excess glucose and phosphorylated glucose to alternative and anabolic pathways, which under normal physiological conditions would contribute but marginally to glucose metabolism29,44–48. Indeed, the cardiac apex shows a dramatic activation of the alternative pathways of glucose utilization: sorbitol dehydrogenase (SDH), a NADH-dependent enzyme of the polyol pathway is overexpressed at 2 h and 7d post-ISO and aldose reductase (AR), the rate-limiting enzyme of this pathway overexpressed at 7d post-ISO; the expression of Gfat1 and Gfat2, the rate limiting enzymes of the HBP in cardiomyocytes and myofibroblasts, respectively, increase dramatically at 2 h post-ISO and remains high at 7d post-ISO. The O-GlcNAc levels increased significantly at 7d post-ISO in the LV apex. In rats, Gfat1 is the primary cardiomyocyte isoform responsible for stress-induced protein O-GlcNAcylation while Gfat2 is found only in cardiac fibroblasts31. This increase of both isoforms may regulate cardiac myofilaments, induce cardiomyocyte dysfunction, and generate fibrosis31,46–48. Following an acute ISO surge in mice, Liao et al. observed increased cardiac infiltration of pro-inflammatory monocytes and a significant reduction of anti-inflammatory (Tim4+/Lyve1+) cardiac resident macrophages49. In their interesting study, the blockade of monocyte infiltration or pro-inflammatory activation alleviated ISO-induced cardiac dysfunction. Hence, they suggested that monocyte infiltration was a major mechanism of TTC pathogenesis, although the contribution of other cells such as neutrophils, dendritic cells, lymphocytes, etc., was not excluded49. In the present study, we observed an immediate inflammatory response with the presence of Cd68+ macrophages early during the acute phase at 2 h post-ISO, i.e., at an earlier time point than those explored by Liao et al.49, and also at 7d post-ISO with the activation of proinflammatory cytokines. Our results confirm that ISO induces an inflammatory response, the question being whether this is the main causal mechanism inducing Takotsubo-like ISO-induced pathology, or one of several mechanisms with contribution from other cells, including the cardiomyocytes themselves. In favor of a direct myocardial reaction to ISO, we show here that the metabolic remodeling concerns cardiomyocytes and does not colocalize with macrophages during the acute and early recovery phases post-ISO (see the cardiac section co-stained for Glut1, Glut4 and Cd68 in Fig. 4). On the other hand, we found as early as 2 h an increase in the expression of Gfat2, which has an anti-inflammatory role in macrophages50. In addition to Liao et al.’s study49, several studies have explored other avenues of treatment in preclinical TTC-like models. Tsikas et al. suggested that GAA, a non-protein guanidino amino acid acting as analogue of Lys, could induce drastic changes in the heart and kidneys compensating energy requirements in case of insufficient creatine supply51. According to Anwar et al., Entresto®, an approved drug that prevents natriuretic peptide degradation, increases BNP levels, and decreases mortality in heart failure patients with reduced EF, decreased cardiac sympathetic activity, attenuated ISO-induced myocardial hypoperfusion, decreasing mortality52. On the other hand, a study by Ellison et al. focusing on the effects of β-adrenergic overload on cardiomyocytes and cardiac stem cells (CSCs), proposed that CSC activation observed at 3 and 6 d post-ISO contributed significantly to the rapid clinical recovery of the TTC heart53. Although they differ in terms of species, sex and time points, these unrelated observations, including ours, strongly suggest that a combination of myocardial metabolic remodeling, inflammation, and other mechanisms involving interactions between cardiac muscle, vessels and multiple resident and infiltrating cell types is responsible for late sequelae of Takotsubo. Further studies are needed to integrate these various mechanisms into a global operational scheme to drive treatment. Simultaneously with the metabolic and inflammatory abnormalities, we observed extensive tissue and vascular remodeling in the apex of the LV already at 7d post-ISO: diffuse fibrosis and increase of endothelial and smooth muscle cells’ vascular biomarkers. It thus appears that a single acute ISO stress is responsible for long term functional, structural, vascular, and molecular changes. The coincidence between tissue and vascular remodeling, and the massive glucose entry and phosphorylation, as well as the hyperactivation of the hexosamine biosynthetic and polyol pathways suggest an intricate link between these events, as reported in other studies. Indeed, overactivation of the HBP promotes cardiac hypertrophy in vitro, and significantly increases the size of cardiac cells, protein synthesis, and the expression of hypertrophy markers46. It was also showed that overactivation of Gfat, the rate-limiting enzyme of the HBP, resulted in increased heart size and fibrosis in Gfat1 transgenic mice50. These results and those from other studies support the view that this overactivation is responsible for the remodeling of the cytoskeleton, the ECM, and the cardiac vessels50,54,55. On another hand, the HBP generates uridine-diphosphate-N-acetylglucosamine (UDP-GlcNAc), the limiting substrate for O-GlcNAcylation, a post-translational modification playing an important regulatory role for O-linked β-Nacetylglucosamine (O-GlcNAc) proteins54–59. A significant increase in O-GlcNAcylation was notably observed in the hypertensive60,61, diabetic32,62, chronically hypertrophied heart63, and in heart failure60 and thought to contribute to contractile and mitochondrial dysfunction63. Fülöp et al. have shown that O-glycosylation of specific proteins contributes to the impairment of cardiomyocyte function in diabetes64. Accordingly, here myocardial O-Glycosylated proteins were significantly more abundant at 7d post-ISO in the apex (Fig. 5b), reflecting a high level of O-GlcNAcylation, in line with the overactivation of the HBP observed in proteomic and western blot analyses of Gfat1/2 (Fig. 5b). The fibrotic conversion of the cardiac tissue that we have shown here definitively excludes the reversibility of stress-induced cardiac damage: fibrosis was present in the LV apex at 7d post-ISO and spread to the entire LV and to the LA at 1 and 3mo post-ISO. The change in the LA reservoir strain, along with the concomitant increase of the LA end-diastolic and end-systolic surfaces, indicate irreversible LA dysfunction and volume enlargement, which are known predictors of diastolic dysfunction41.This supports a long-term cardiac embrittlement following Takotsubo, and an evolution towards diastolic dysfunction predictive of a serious cardiac condition with reserved prognosis of Takotsubo patients41,42. Overall, longitudinal observations in this animal model Takotsubo-type cardiomyopathy triggered by a single injection of ISO, clearly reflect irreversible tissue, vascular, and metabolic sequelae that are prone to weaken the heart and render it susceptible to recurrent cardiac disease5. The present work points to alternative pathways of glucose metabolism as the critical mechanism by which the one-time stressed heart engages in a continuous degradation of structure and function. These pathways may represent attractive targets to prevent the progression of cardiac damage following an acute stress. The rapid initial changes in metabolic control and early appearance of diffuse apical fibrosis suggest that fibrosis progression, vascular network alteration, and metabolic remodeling are initiated early after stress. Therefore, preventing deleterious evolution of TTC may require treatment to be administered during or immediately after the acute phase of TTC. This calls for careful attention to the management of TTC patients, whose incidence has been reported to increase recently5,23,65,66. Our study also highlights the importance of myocardial strain measurements and FDG PET imaging to assess and detect both regional and global functional, tissue, and metabolic remodeling in the TTC heart. The outcomes reported here in an animal model are in favor of clinical trials using these two in vivo and non-invasive imaging modalities is crucial for a better management of TTC patients, and the optimization of the diagnosis and prognosis of Takotsubo. Clinically, Takotsubo is multifactorial and complex and often associated with different comorbidities, The present study used a simple animal model to investigate the longitudinal functional, anatomical, metabolic, tissue, and vascular modifications of the heart in a reproducible Takotsubo-type animal model, without the various comorbidities encountered in patients, such as the cardiovascular diseases associated with aging, hypertension, diabetes, obesity, etc. A unique injection of beta-adrenergic catecholamines led to a cascade of reproducible activation and inactivation of different metabolic and anabolic pathways of glucose and to short- and long-term tissue, fibrotic and vascular sequelae of the heart. In short, the stress induced a transient ventricular dysfunction during the acute phase, and an irreversible tissue and functional impairment of the left atrium in the long term. The proteomic analysis at 2 h and 7 days were based on 5 replicates per condition, which may lead some samples to yield less protein raw identification than the average across all samples. However, in the present study we used the proteomics data to build hypotheses regarding energetic pathway changes and interpreted these in the light of the in vivo and ex vivo imaging, and of the physiological and biochemical analysis. This reasonable use of proteomics-derived data does suggest clues/leads about pathophysiological mechanisms in TTC, which naturally remain to be examined in the light of other proposed mechanisms such as, e.g., direct lipidotoxicity27, impairment of the mitochondrial respiratory chain67 innate immune reactions46, modifications of calcium signaling62,68, and others. Importantly, as with any reductionist approach, transposition into the clinical realm should be weighted for its relevance with respect to clinical translation. Among the limitations of the model used here, we used young adult female rats, while TTC is most frequently observed in post-menopausal female patients. However, TTC, has also been described in young females and in males, therefore our study may benefit from its extension to male and to older animals. Secondly, the animal model used here showcases the consequences of a single triggering factor of TTC, a sudden catecholaminergic rush, while Takotsubo has been associated with mental and physical stress, surgical stress, neurological disorders, pheochromocytoma, etc.14. No single model can encompass the whole TTC spectrum, the variable severity of the disease and of its progression, and the role of inducing stressors5,14. Comparing the results from the present animal model, with different forms of TTC in individual patients and with the original description of emotional stress-induced TTC14,65,66,69, is certainly a challenge of future TTC studies. Methods Animals Experiments were approved by the French Animal Ethics Committee (agreement number 19064). One hundred and twenty 12-week-old female Wistar rats were obtained from Janvier (Le Genest-St-Isle, France) and housed in our facility on a 12:12 light-dark cycle, controlled temperature, 25 °C, 60% air humidity, and free access to food (M-bricks normal chow from TAPVEI, Kiili Fajumaa, Estonia) and water. Our longitudinal study of myocardial remodeling focused on the female TTC-like animal model for several reasons. The main reason is that TTC predominantly affects female patients (about 90% of TTC cases)14. In addition, male rodents depict more variability in functional parameters following ISO injection than female rodents70. Kneale et al. speculate that this is the consequence of a different sensitivity to adrenergic stimulation between males and females70,71. The purpose of our study was to explore metabolic, functional, tissue, and vascular remodeling cascades following a viable dose of ISO in small groups of animals, therefore we prioritized females for a better homogeneity between individuals and for the closest match with the clinical incidence of TTC. Induction of Takotsubo-like syndrome in animals Imaging, physiological parameters, blood chemistry, histology, western blots and proteomics were obtained before stress induction (day 0, baseline). A unique intraperitoneal injection of 50 mg/kg isoprenaline (ISO, Isoproterenol hydrochloride, Sigma-Aldrich, Germany) was performed on day 1. Imaging, physiological parameters, blood chemistry, histology, western blots and proteomics were obtained 2 h and 7 days post-ISO and then repeated at 1 month and 3 months post-ISO in the same animals. Two rats died during MRI imaging sessions, one before ISO injection from anesthesia, and one rat died 2 h post-ISO (1% mortality rate). FDG positron emission tomography (PET, Supplementary Fig. 7) On days 0, 1, 7 and at 1 and 3 months, non-fasted rats were anesthetized (isoflurane 4% induction and 2% maintenance in air), weighted and glycemia and animal temperature were recorded. The animal was placed supine in a nanoScan PET-CT scanner (Mediso Medical Imaging Systems, Hungary) with respiratory and cardiac monitoring. A commercial ultrasound probe (SuperLinear™ SLH20-6, Supersonic Imagine, France, central frequency 15 MHz) connected to a small animal ultrasound device (Aixplorer, Supersonic Imagine, France) was positioned on the depilated chest of the animal to obtain a view of the full long-axis of the beating heart. The animal was then moved in the PET gantry and a whole-body X-ray tomodensitometry (CT) was acquired using the following acquisition parameters: semi-circular mode, 70 kV tension, 720 projections full scan, 300 ms per projection, binning 1:4. Images were reconstructed by filtered retro-projection (filter: Cosine; Cutoff: 100%) using Nucline version 3.00.010.0000 (Mediso Medical Imaging Systems, Hungary). Immediately after CT acquisition, a 30 min dynamic PET scan and 30 s after starting the acquisition, 32.9 ± 1.4 MBq of 2’-deoxy-2’-[18F]fluoro-D-glucose (FDG; Advanced Applied Applications, France) in 0.4 mL saline were injected in the lateral tail vein. At the end of the first dynamic PET scan, a static 30 min PET scan was acquired with ECG and respiratory gating. PET data was collected in list mode and binned using a 5 ns time window, with a 400-600 keV energy window and a 1:5 coincidence mode. Data was reconstructed using the Tera-Tomo reconstruction software (3D-OSEM based manufactured customized algorithm) with expectation maximization iterations, scatter, and attenuation correction. The 30 min dynamic PET exam was reconstructed in 23 frames as follows: 30 s, 6 × 5 s; 4 × 10 s; 6 × 30 s; 3 × 120 s; 4 × 300 s. The ECG-gated cardiac PET was reconstructed in a single frame of 15 min, 45–60 min post FDG injection. Using the PET/CT fusion slices, volumes-of-interest (VOI) were delineated for the left ventricle (LV) using PMOD software (PMOD Technologies Ltd, version 3.8, Zürich, Switzerland). FDG uptake was quantified as Standard Uptake Value. The Peak SUV was calculated as the maximum average SUV within a 1-cm3 spherical VOI, and the LV volume was automatically segmented at 40% of this value. Compartmental kinetic assessment of FDG uptake was based on the 2-tissue compartment model of the PMOD kinetics package with a lump constant set to 1. Positron emission tomography registered ultrafast sonography The high temporal sampling of ultrafast ultrasound imaging (UUI) and its unique spatial resolution (~0.1 mm) were used as a priori anatomical information conducive to correct movement and partial volume effect of cardiac PET images using Super-Resolution (SR) methods72,73. B-mode images were acquired during the acquisition of the dynamic PET scan, over 1 s, using Aixplorer® (SuperSonic Imagine, France) in order to sample the whole cardiac cycle. Respiratory and cardiac motions were monitored during the acquisitions. Cardiac magnetic resonance imaging Cardiac magnetic resonance (CMR) imaging acquisitions were performed in a preclinical 4.7 T MRI system (Bruker BioSpec 47/40 USR, Ettlingen, Germany) with a 7 cm inner diameter resonator for emission and a phase array surface coil for reception under isoflurane anaesthesia as above. Respiration, heart rate, rectal temperature and isoflurane delivery were monitored constantly and maintained stable during the acquisitions. LV function was measured by cine T1-weighted cine sequences including short axis stack of 6–7 slices covering the left and right ventricles from base to apex, along with conventional 2-chamber and a 4-chamber views, using the following scan parameters: TR = 7.2 ms; TE = 2.7 ms; flip angle=18.0°; field of view (FOV) = 60 × 60 mm; matrix size = 256 × 256, planar resolution = 235 µm, oversampling = 100, number of frames = 16 per heart cycle, slice thickness = 2 mm. Total scan time was 1min32s per slice. The cine intra-gate sequence was triggered using averaged heart and respiration rates in a retrospective reconstruction. Acquisitions were performed using Paravision software 6.0.1 (Bruker, Ettlingen, Germany) and T1 cine images were analyzed manually using Circle civ42 software (version 5.13.5, Circle Cardiovascular Imaging Inc., Canada) to estimate global heart chamber volumes. Cardiac strain and strain rate analysis Cardiac strain defines the level of the wall deformation of a cardiac cavity along the cardiac cycle. It is a non-volumetric parameter, complementary to the conventional functional parameters, allowing to study different spatial components of the myocardial contraction function. Strain can be measured by the speckle tracking method using ultrasound70, or by MRI that offers the advantage of an observer-independent complete coverage of the heart74. In our study, we opted for strain measurements based on T1 cine sequences using a semi-automatic homemade software, CardioTrack (Sorbonne University, Paris, France)75,76, based on feature tracking77 and written in Matlab® (The MathWorks, Natick, MA, USA). The strain measurements of the left and right ventricles and of the atria were performed using the 4 chamber views for reliability and repeatability reasons. We also determined on this view the minimum and maximum atrial surfaces, as surrogates of the atrial volumes. Missing data correspond only to the strain values that were incalculable for technical reasons, e.g., incorrect orientation or improper ECG gating, but no data was removed to ameliorate statistical significance. Sample preparation for histology The LV and atria were excised and fixed during 24 h in 4% formaldehyde, then transferred to 70% EtOH and embedded in paraffin. For oil-red staining, they were fixed in an isopentane bath placed in liquid nitrogen. Hearts were cut in 4 µm thick section with a microtome (RM2145, Leica, Germany) for paraffin embedded tissue, or at 6 µm in a cryostat (CM3050 S, Leica, Germany) for frozen tissue. For each heart, 6 sections were taken at 5 positions spaced 100 µm apart in the apex, 6 sections at 3 positions spaced 200 µm apart in the basal region, and 6 sections at 4 positions spaced 50 µm apart in the atria. Immunohistochemistry (IHC) Each cardiac section was stained in triplicate for red Sirius. Sirius red staining was performed in an automate (ST5020, Leica, Germany). Stained sections were scanned in their entirety with a NanoZoomer HT 2.0 (Hamamatsu) at a magnification of x20. Immunohistofluorescence (IHF) Sections were deparaffinized, rehydrated and incubated in blocking buffer (5% Bovine Serum Albumine (BSA) in Tris buffered saline-tween® 20, Sigma-Aldrich, USA) during 30 min. Sections were incubated with the primary antibody in 3% BSA overnight before adding the fluorophore coupled-secondary antibody in 3% BSA during 2 h. Immuno-staining for Glut1 (1:200, rabbit monoclonal, #ab115730, Abcam), Glut4 (1:500, rabbit polyclonal, #ab33780, Abcam) and Cd68 (1:400, mouse monoclonal, #MCA341R, Bio-rad) were performed using Opal™ Multiplex immunohistochemistry (IHC) kits (AKOYA Biosciences, Marlborough, MA and Menlo Park, CA, USA) that allow simultaneous IHC detection of the three antigens. Immuno-staining for Cd31 (1:200, goat polyclonal, #AF3628, R&D systems) and cyanine 3 coupled-alpha SMA (mouse monoclonal, #C6198, Sigma-Aldrich) were performed simultaneously. The immune-stained sections were digitalized using Vectra® Polaris™ (AKOYA Biosciences, Marlborough, MA and Menlo Park, CA, USA) with the entire section in a single field of view. Staining quantification and analysis A homemade machine learning based software, Fiber-ML, written in Matlab® (The MathWorks, Natick, MA, USA)77, was used to quantify IHC and IHF staining in the entirety of each section. The capillary density in the LV was calculated as the number of Cd31 stained capillaries divided by the number of cardiomyocytes in a given field of the IHF sections using an ImageJ® macro. Western blot LV samples were divided into apex and basal segments and frozen. Two protocols were performed in order to detect and quantify the immunoblots of interest. Protocol 1: samples lysed in 1X SDS sample buffer and sonicated for 10–15 s. Protein concentrations were determined using the Pierce™ BCA Protein Assay kit (#23225 and #23227, Thermoscientific). A volume of 20 µL of each lysate was loaded onto a SDS-PAGE gel (10 cm × 10 cm, mini-protein TGX gel, #5678024, BioRad), run at 100 volts for 1 h under constant amperage at 30 mA, and the gel was electro-transferred to nitrocellulose membranes (Trans-Blot Turbo Midi 0.2 µm, #1704159, BioRad). The membranes were blocked in skimmed milk and incubated overnight with the following primary antibodies: anti-Gfat1 (1:1000, #D12F4, Cell Signaling Technology) and anti-Gapdh (1:1000, Abcam). Blots were then incubated during 2 h with HRP-linked anti-rabbit IgG (1:5000, goat, #4030-05, SouthernBiotech) and anti-mouse IgG (1:5000, light chain binding protein, #sc-516102, Santa Cruz Biotechnology). Protocol 2: Ripa buffer (150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 1% Sodium Deoxycholate, 0.1% SDS, 50mMTrisHCl pH7.4,) with protease cocktail inhibitor (cOmplete Mini, Roche) was added to the samples and they were first mechanically homogenized (30 s; UltraTurrax IKA) and then stored on ice for 30 min to complete the lysis. Samples were centrifuged (4500 g 5 min 4 °C) and the pellet discarded. Protein concentrations were determined using the DC™ Protein Assay kit (Bio-Rad). 30 µg total proteins of each lysate were loaded onto a SDS-PAGE gel and then electro-transferred to nitrocellulose membrane (Nitrocellulose 0.2 µm, #1620112, BioRad). The membranes were blocked 1 h in Blotto 3% (TBS buffer with 0.05% Tween20 and 3% skimmed milk) and incubated 1 h with the following primary antibodies: anti-Gfat2 (1:1000, #ab190966, Abcam) or O-linked N-acetylglucosamine (1:2000, #RL2, Invitrogen) and anti-Gapdh (1:1000, #ab8245, Abcam). Blots were then incubated 1 h with HRP-linked donkey anti-rabbit IgG (1:5000, Jackson ImmunoReasearch) or HRP-linked donkey anti-mouse IgG (1:5000, Jackson ImmunoReasearch). Antibody binding was revealed using the chemiluminescent substrate kit (Clarity Western ECL substrate, BioRad) and SuperSignal West Femto Maximum sensitivity substrate (#34095, ThermoFisher). Proteomic analysis Frozen hearts samples were individually ground under liquid nitrogen to yield a fine powder using a pestle and mortar. The tissue powder was quickly weighted and solubilized in 95 °C lysis buffer (4% SDS, 100 mM Tris-HCl, pH 8.0). Protein extracts were clarified by centrifugation at 21,000 × g for 1 h at 4 °C. Protein concentration of the supernatant was estimated and normalized using image intensity integration of the Coomassie blue G250-colored SDS PAGE, loaded with the same volume of each lysate. Peptides were prepared by the Strap technique (ProtiFi, NY, USA) and desalted on C18 StageTips78. After speed-vacuum drying, peptides were solubilized in 2% trifluoroacetic acid (TFA) and fractionated by strong cationic exchange (SCX) StageTips. Each SCX fraction of each sample was matched with the same fraction and the two adjacent fractions (1 with 1 and 2; 2 with 1, 2 and 3; 3 with 2, 3 and 4…). LC-MS analyses were performed on an U3000 RSLC nano-LC (Dionex) system coupled to a TIMS-TOF Pro mass spectrometer (Bruker Daltonik GmbH, Germany). After drying, peptides from SCX StageTip, the fractions were solubilized in 10 μL of 0.1% TFA containing 2% acetonitrile (ACN). one μL was loaded, concentrated, and washed for 3 min on a C18 reverse phase precolumn (3 μm particle size, 100 Å pore size, 75 μm inner diameter, 2 cm length, from Thermo Fisher Scientific). Peptides were separated on an Aurora C18 reverse phase resin (1.6 μm particle size, 100 Å pore size, 75 μm inner diameter, 25 cm length mounted to the CSI module, from IonOpticks; Australia) with a 120 min run time and a gradient ranging from 99% of solvent A containing 0.1% formic acid in milliQ-grade H2O, to 40% of solvent B containing 80% acetonitrile, 0.085% formic acid in mQH2O. The mass spectrometer acquired data throughout the elution process and operated in DDA PASEF mode with a 1.89 second/cycle, with Timed Ion Mobility Spectrometry (TIMS) mode enabled and a data-dependent scheme with full MS scans in PASEF mode. This enabled a recurrent loop analysis of a maximum of the 120 most intense nLC-eluting peptides which were CID-fragmented between each full scan every 1.89 s. Ion accumulation and ramp time in the dual TIMS analyzer were set to 166 ms each and the ion mobility range was set from 1/K0 = 0.6 Vs cm−2 to 1.6 Vs cm−2. Precursor ions for MS/MS analysis were isolated in positive polarity in the 100–1.700 m/z range by synchronizing quadrupole switching events with the precursor elution profile from the TIMS device. The cycle duty time was set to 100%, accommodating as many MSMS in the PASEF frame as possible. Singly charged precursor ions were excluded from the TIMS stage by tuning the TIMS using the Otof control software (Bruker Daltonik GmbH). Precursors for MS/MS were picked from an intensity threshold of 1000 arbitrary units (a.u.) and resequenced until reaching a ‘target value’ of 20,000 a.u, considering a dynamic exclusion gap of 0.40 min. The mass spectrometry data were analyzed using MaxQuant version 1.6.17 (Max Planck institute of Biochemistry, Germany). The database used was a concatenation of Rattus norvegicus sequences from the Swissprot database (release 2020-10) and a list of contaminant sequences from MaxQuant. The enzyme specificity was trypsin. The precursor and fragment mass tolerance were set to 20ppm. Carbamido-methylation of cysteins was set as permanent modification and acetylation of protein N-terminus and oxidation of methionines were set as variable modifications. Second peptide search was allowed, and minimal length of peptides was set at 7 amino acids. False discovery rate (FDR) was kept below 1% on both peptides and proteins. Label-free protein quantification (LFQ) was done using both unique and razor peptides. At least 2 such peptides were required for LFQ. The “match between runs” (MBR) option was allowed with a match time 0.7 min window and an alignment time window of 20 min. For differential analysis, LFQ results from MaxQuant were imported into the Perseus software (version 1.6.14). After excluding reverse and contaminant proteins from analysis, 4257 proteins were identified, among which 3203 proteins had at least one LFQ value across all samples. Before performing statistics, the data was transformed to a logarithmic scale (log2) and further filtered: only 2618 proteins with at least 3 out of 5 valid LFQ values in at least one experimental group were used for comparison between groups. Of these, 971 proteins were identified with a value in each sample (Supplementary Fig. 8a). Five samples/group post-ISO were performed. In the five 7d post-ISO LV apical group, two samples were outliers. However, in order to maintain homogeneous sample comparison between groups, we did not eliminate these 2 outlier samples since (i) they showed no series effect and (ii) their inclusion did not flaw the statistical confidence of this group’s comparisons (Supplementary Fig. 8b). Significant differential proteins were identified with a two-tailed Student’s t-test. Only proteins with p-value < 0.05 and |fold change|>1.3 were kept for over-representation analysis using Ingenuity Pathway Analysis (QIAGEN Inc, version 60467501). Proteins with missing values in all samples of one group (0/5) and at least 3 values in another group (3/5) were considered as appearance / disappearance and included in the dataset of 2618 proteins (supplementary fig. 8a). Statistical analysis of six group-to-group comparisons, i.e., in the apex: 0 vs. 2 h, 7d vs. 0 and 7d vs. 2 h; in the basal region: 0 vs. 2 h, 7d vs. 0 and 7d vs. 2 h, showed significant differences in 1110 proteins (743 proteins in the apex and 614 proteins in the basal region). Significantly over-represented terms (canonical pathways, functions, upstream regulators) were identified with a right-tailed Fisher’s Exact Test that calculates an overlap p‐value. The z-score was also calculated to assess the activation (positive z-score) or repression (negative one) of each term. Statistical analysis For each experimental series, data are presented as means ± standard deviation. For data from longitudinal follow-up of in vivo imaging, each animal is its own control, and values were normalized to the animal’s baseline values. Statistical analysis was performed with GraphPad Prism 9.2.0 (GraphPad Software, San Diego, CA, USA). Statistical significance (p < 0.05) for each variable was estimated by one-way or two-way ANOVA when group variances were equal (Bartlett test); if not the non-parametric Kruskall–Wallis test, and the Holm multiple comparisons test was used to execute simultaneous t-tests. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Peer Review File Reporting Summary Source data Source Data Supplementary information The online version contains supplementary material available at 10.1038/s41467-023-39590-3. Acknowledgements We are grateful to Corinne Lesaffre from the PARCC Histology Core Facility, Florence Marliot from the HEGP Histology Core Facility, Dr Hana Manceau and Nicolas Sorhaindo from the Bichat Biochemistry Core Facility, to Maryline Favier and Rachel Onifarasoaniaina from the Cochin Histology Core Facility, and to Omar Zenteno for Supplementary Fig. 7a. In vivo imaging was performed at the Imaging Facility (Plateforme d’Imageries du Vivant) of the University Paris Cité supported by France Life Imaging (ANR-11INBS-0006), by Infrastructures Biologie-Santé (IBiSa), by Aviesan grants #ASC20001SSP, ASC16025KSA and ASC20031KSA, and by the Région Ile-de-France SESAME funding program. V.N. was funded by H2020 MAESTRIA #965286. We warmly thank Philippe Chafey for his contribution to sample preparation, proteomics data processing, and for his long career of contributions to protein analysis for the scientific community. Many thanks to Tony Lefebvre from University of Lille for his expert advice on O- and N- acetylglycosylation. This work was supported by the DIM Thérapie Génique Paris Ile-de-France Region, by IBiSA, by the Labex GR-Ex, and by a grant from ANR PACIFIC (ANR-18-CE14-0032). T.Y. was supported by the French Ministry of Research and Higher Education, and by the ANR PACIFIC grant. M.P.-L. received funding from the European Union’s Horizon 2020 research and innovation Program under the Marie Sklodowska-Curie Grant Agreement no.101030046, and by the Programme Ramón y Cajal RYC2021-032739-I, funded by MCIN/AEI/10.13039/501100011033 and the European Union “NextGenerationEU”/PRTR. Part of the technology developed for this study was supported by SIRIC CARPEM grants to B.T. Author contributions T.Y. carried out all the experimental work and data analysis. T.Y. and B.T. conceived the experiments, discussed the results, and wrote the manuscript. M.P.L. developed cardiac PETRUS technology. D.B. created the Fiber-ML software. M.L.G., F.G., and J.B. conducted and analyzed the proteomics data. A.C. performed the OPAL experiments and histological analysis using Fiber-ML, supervised by D.B. and B.T. P.B. provided expert assistance for the optimization and interpretation of histological results. Y.M. analyzed all the OPAL experiment under the supervision of T.Y. G.A. optimized cardiac MRI sequences and trained T.Y. in their usage. U.G., V.N., and N.K. developed the Cardiotrack software. A.L. analyzed the strain measurements results under the supervision of T.Y. and B.T. A.S. performed the electron microscopy. T.G. supervised the SHG imaging and analysis performed by T.Y. F.L. and G.R. performed the cardiac ultrasound imaging, G.R. overseeing the analysis conducted by T.Y. J.V. performed and analyzed western blot with N.M. E.M. provided expert assistance in myocardial MRI and ultrasound imaging. T.V. supervised PET FDG acquisitions made by T.Y. All authors reviewed and approved the final manuscript. Peer review Peer review information Nature Communications thanks Richard Wilson and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Data availability Source data are provided with this paper. The mass spectrometry proteomics data that support the findings of this study are available in the ProteomeXchange Consortium via the PRIDE79 partner repository with the dataset identifier PXD032667. Source data are provided with this paper. Code availability The software developed for this study has been archived in open repositories: Fiber-ML is available at https://gitlab.com/balvayda/fiber-ml. CardioTrack is available at https://gitlab.com/LIB_ICV/cardiotrack. Competing interests The authors declare no competing interests. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Change history 7/12/2023 A Correction to this paper has been published: 10.1038/s41467-023-39910-7 ==== Refs References 1. Nyberg ST Job strain as a risk factor for type 2 diabetes: a pooled analysis of 124,808 men and women Diabetes Care 2014 37 2268 2275 10.2337/dc13-2936 25061139 2. Hackett RA Steptoe A Type 2 diabetes mellitus and psychological stress—a modifiable risk factor Nat. Rev. Endocrinol. 2017 13 547 560 10.1038/nrendo.2017.64 28664919 3. Sato, H. Tako-tsubo-like left ventricular dysfunction due to multivessel coronary spasm. Clin. Asp. Myocard. Inj. Ischemia Heart Fail. 56–64 (Kagakuhyoronsha Publishing Co., Tokyo, 1990) (in Japanese). 4. Lyon AR Current state of knowledge on Takotsubo syndrome: a Position Statement from the Taskforce on Takotsubo Syndrome of the Heart Failure Association of the European Society of Cardiology Eur. J. Heart Fail. 2016 18 8 27 10.1002/ejhf.424 26548803 5. Ghadri JR Long-term prognosis of patients with Takotsubo Syndrome J. Am. Coll. Cardiol. 2018 72 874 882 10.1016/j.jacc.2018.06.016 30115226 6. Wittstein IS Neurohumoral features of myocardial stunning due to sudden emotional stress N. Engl. J. Med. 2005 352 539 548 10.1056/NEJMoa043046 15703419 7. Motiejunaite J Amar L Vidal-Petiot E Adrenergic receptors and cardiovascular effects of catecholamines Ann. Endocrinol. 2021 82 193 197 10.1016/j.ando.2020.03.012 8. Moens AL Yang R Watts VL Barouch LA Beta 3-adrenoreceptor regulation of nitric oxide in the cardiovascular system J. Mol. Cell. Cardiol. 2010 48 1088 1095 10.1016/j.yjmcc.2010.02.011 20184889 9. Jones DP Shan X Park Y Coordinated multisite regulation of cellular energy metabolism Annu. Rev. Nutr. 1992 12 327 343 10.1146/annurev.nu.12.070192.001551 1503809 10. Sourdon J Sunitinib-induced cardiac hypertrophy and the endothelin axis Theranostics 2021 11 3830 3838 10.7150/thno.49837 33664864 11. Sourdon J Cardiac metabolic deregulation induced by the tyrosine kinase receptor inhibitor sunitinib is rescued by endothelin receptor antagonism Theranostics 2017 7 2757 2774 10.7150/thno.19551 28824714 12. Stanley WC Recchia FA Lopaschuk GD Myocardial substrate metabolism in the normal and failing heart Physiol. Rev. 2005 85 1093 1129 10.1152/physrev.00006.2004 15987803 13. Barth E Glucose metabolism and catecholamines Crit. Care Med. 2007 35 S508 10.1097/01.CCM.0000278047.06965.20 17713401 14. Ghadri J-R International Expert Consensus Document on Takotsubo Syndrome (Part I): Clinical Characteristics, Diagnostic Criteria, and Pathophysiology Eur. Heart J. 2018 39 2032 2046 10.1093/eurheartj/ehy076 29850871 15. Shao Y Novel rat model reveals important roles of β-adrenoreceptors in stress-induced cardiomyopathy Int. J. Cardiol. 2013 168 1943 1950 10.1016/j.ijcard.2012.12.092 23357048 16. Sachdeva J Dai W Kloner RA Functional and histological assessment of an experimental model of Takotsubo’s cardiomyopathy J. Am. Heart Assoc. 2014 3 e000921 10.1161/JAHA.114.000921 24958782 17. McClain DA Hexosamines as mediators of nutrient sensing and regulation in diabetes J. Diabetes Complications 2002 16 72 80 10.1016/S1056-8727(01)00188-X 11872372 18. Veerababu G Overexpression of glutamine: fructose-6-phosphate amidotransferase in the liver of transgenic mice results in enhanced glycogen storage, hyperlipidemia, obesity, and impaired glucose tolerance Diabetes 2000 49 2070 2078 10.2337/diabetes.49.12.2070 11118009 19. Paur H High levels of circulating epinephrine trigger apical cardiodepression in a β2-adrenergic receptor/Gi-dependent manner: a new model of Takotsubo cardiomyopathy Circulation 2012 126 697 706 10.1161/CIRCULATIONAHA.112.111591 22732314 20. Collen J Bimson W Devine P A variant of Takotsubo cardiomyopathy: a rare complication in the electrophysiology lab J. Invasive Cardiol. 2008 20 E310 E313 18987407 21. Grippo AJ Johnson AK Stress, depression and cardiovascular dysregulation: a review of neurobiological mechanisms and the integration of research from preclinical disease models Stress 2009 12 1 21 10.1080/10253890802046281 19116888 22. Citro R Multimodality imaging in takotsubo syndrome: a joint consensus document of the European Association of Cardiovascular Imaging (EACVI) and the Japanese Society of Echocardiography (JSE) J. Echocardiogr. 2020 18 199 224 10.1007/s12574-020-00480-y 32886290 23. Templin C Clinical features and outcomes of Takotsubo (stress) cardiomyopathy N. Engl. J. Med. 2015 373 929 938 10.1056/NEJMoa1406761 26332547 24. Prasad A Lerman A Rihal CS Apical ballooning syndrome (Tako-Tsubo or stresscardiomyopathy): a mimic of acute myocardial infarction Am. Heart J. 2008 155 408 417 10.1016/j.ahj.2007.11.008 18294473 25. Phelps ME Tomographic measurement of local cerebral glucose metabolic rate in humans with (F-18)2-fluoro-2-deoxy-D-glucose: validation of method Ann. Neurol. 1979 6 371 388 10.1002/ana.410060502 117743 26. Nef HM Tako- Tsubo cardiomyopathy: intraindividual structural analysis in the acute phase and after functional recovery Eur. Heart J. 2007 28 2456 2464 10.1093/eurheartj/ehl570 17395683 27. Shao Y A mouse model reveals an important role for catecholamine‐induced lipotoxicity in the pathogenesis of stress‐induced cardiomyopathy Eur. J. Heart Fail. 2013 15 9 22 10.1093/eurjhf/hfs161 23099354 28. Stratmann B Tschoepe D The diabetic heart: sweet, fatty and stressed Expert Rev. Cardiovasc. Ther. 2011 9 1093 1096 10.1586/erc.11.109 21932950 29. Copeland RJ Bullen JW Hart GW Cross-talk between GlcNAcylation and phosphorylation: roles in insulin resistance and glucose toxicity Am. J. Physiol. Endocrinol. Metab. 2008 295 E17 E28 10.1152/ajpendo.90281.2008 18445751 30. Ishikita A GFAT2 mediates cardiac hypertrophy through HBP-O-GlcNAcylation-Akt pathway iScience 2021 24 103517 10.1016/j.isci.2021.103517 34934932 31. Nabeebaccus AA Cardiomyocyte protein O-GlcNAcylation is regulated by GFAT1 not GFAT2 Biochem. Biophys. Res. Commun. 2021 583 121 127 10.1016/j.bbrc.2021.10.056 34735873 32. Fülöp N Marchase RB Chatham JC Role of protein O-linked N-acetyl-glucosamine in mediating cell function and survival in the cardiovascular system Cardiovasc. Res. 2007 73 288 297 10.1016/j.cardiores.2006.07.018 16970929 33. Huang J Increased glucose disposal and AMP-dependent kinase signaling in a mouse model of hemochromatosis J. Biol. Chem. 2007 282 37501 37507 10.1074/jbc.M703625200 17971451 34. Kolm-Litty V Sauer U Nerlich A Lehmann R Schleicher ED High glucose-induced transforming growth factor beta1 production is mediated by the hexosamine pathway in porcine glomerular mesangial cells J. Clin. Invest. 1998 101 160 169 10.1172/JCI119875 9421478 35. Yang X Phosphoinositide signalling links O-GlcNAc transferase to insulin resistance Nature 2008 451 964 969 10.1038/nature06668 18288188 36. Peng C Regulation of the Hippo-YAP Pathway by Glucose Sensor O-GlcNAcylation Mol. Cell 2017 68 591 604.e5 10.1016/j.molcel.2017.10.010 29100056 37. Kim E O-GlcNAcylation on LATS2 disrupts the Hippo pathway by inhibiting its activity Proc. Natl Acad. Sci. USA 2020 117 14259 14269 10.1073/pnas.1913469117 32513743 38. Xiao Y Hippo pathway deletion in adult resting cardiac fibroblasts initiates a cell state transition with spontaneous and self-sustaining fibrosis Genes Dev. 2019 33 1491 1505 10.1101/gad.329763.119 31558567 39. Lorenzi M The polyol pathway as a mechanism for diabetic retinopathy: attractive, elusive, and resilient Exp. Diabetes Res. 2007 2007 61038 10.1155/2007/61038 18224243 40. Tang WH Martin KA Hwa J Aldose reductase, oxidative stress, and diabetic mellitus Front. Pharmacol. 2012 3 87 10.3389/fphar.2012.00087 22582044 41. Pritchett AM Diastolic dysfunction and left atrial volume J. Am. Coll. Cardiol. 2005 45 87 92 10.1016/j.jacc.2004.09.054 15629380 42. Aouar LMME Relationship between left atrial volume and diastolic dysfunction in 500 Brazilian patients Arq. Bras. Cardiol. 2013 101 52 58 23702813 43. Godsman N Metabolic alterations in a rat model of takotsubo syndrome Cardiovasc. Res. 2022 118 1932 1946 10.1093/cvr/cvab081 33711093 44. Srivastava SK Ramana KV Bhatnagar A Role of aldose reductase and oxidative damage in diabetes and the consequent potential for therapeutic options Endocr. Rev. 2005 26 380 392 10.1210/er.2004-0028 15814847 45. González RG Barnett P Aguayo J Cheng HM Chylack LT Direct measurement of polyol pathway activity in the ocular lens Diabetes 1984 33 196 199 10.2337/diab.33.2.196 6692996 46. Tran DH Chronic activation of hexosamine biosynthesis in the heart triggers pathological cardiac remodeling Nat. Commun. 2020 11 1771 10.1038/s41467-020-15640-y 32286306 47. Ramirez-Correa GA Removal of abnormal myofilament O-GlcNacylation restores Ca2+ sensitivity in diabetic cardiac muscle Diabetes 2015 64 3573 3587 10.2337/db14-1107 26109417 48. Ramirez-Correa GA O-linked GlcNAc modification of cardiac myofilament proteins: a novel regulator of myocardial contractile function Circ. Res. 2008 103 1354 1358 10.1161/CIRCRESAHA.108.184978 18988896 49. Liao X Cardiac macrophages regulate isoproterenol-induced Takotsubo-like cardiomyopathy JCI Insight 2022 7 e156236 10.1172/jci.insight.156236 35132957 50. Al-Mukh H Lipopolysaccharide induces GFAT2 expression to promote O-linked β-N-acetylglucosaminylation and attenuate inflammation in macrophages J. Immunol. 2020 205 2499 2510 10.4049/jimmunol.2000345 32978282 51. Tsikas D Redfors B Pilot study on acute effects of pharmacological intraperitoneal L-homoarginine on homeostasis of lysine and other amino acids in a rat model of isoprenaline-induced takotsubo cardiomyopathy Int. J. Mol. Sci. 2022 23 4734 10.3390/ijms23094734 35563125 52. Ali A Sacubitril/valsartan decreases mortality in the rat model of the isoprenaline‐induced takotsubo‐like syndrome ESC Heart Fail. 2021 8 4130 4138 10.1002/ehf2.13530 34463049 53. Ellison GM Acute β-adrenergic overload produces myocyte damage through calcium leakage from the ryanodine receptor 2 but spares cardiac stem cells J. Biol. Chem. 2007 282 11397 11409 10.1074/jbc.M607391200 17237229 54. Qin CX Insights into the role of maladaptive hexosamine biosynthesis and O-GlcNAcylation in development of diabetic cardiac complications Pharmacol. Res. 2017 116 45 56 10.1016/j.phrs.2016.12.016 27988387 55. Hebert LF Overexpression of glutamine:fructose-6-phosphate amidotransferase in transgenic mice leads to insulin resistance J. Clin. Invest. 1996 98 930 936 10.1172/JCI118876 8770864 56. Akella NM Ciraku L Reginato MJ Fueling the fire: emerging role of the hexosamine biosynthetic pathway in cancer BMC Biol. 2019 17 52 10.1186/s12915-019-0671-3 31272438 57. Fisi V Miseta A Nagy T The role of stress-induced O-GlcNAc protein modification in the regulation of membrane transport Oxid. Med. Cell. Longev. 2017 2017 1308692 10.1155/2017/1308692 29456783 58. Yang X Qian K Protein O-GlcNAcylation: emerging mechanisms and functions Nat. Rev. Mol. Cell Biol. 2017 18 452 465 10.1038/nrm.2017.22 28488703 59. Biwi J Biot C Guerardel Y Vercoutter-Edouart A-S Lefebvre T The many ways by which O-GlcNacylation may orchestrate the diversity of complex glycosylations Mol. Basel Switz. 2018 23 E2858 60. Lunde IG Cardiac O-GlcNAc signaling is increased in hypertrophy and heart failure Physiol. Genomics 2012 44 162 172 10.1152/physiolgenomics.00016.2011 22128088 61. Lima VV Rigsby CS Hardy DM Webb RC Tostes RC O-GlcNAcylation: a novel post-translational mechanism to alter vascular cellular signaling in health and disease: focus on hypertension J. Am. Soc. Hypertens. 2009 3 374 387 10.1016/j.jash.2009.09.004 20409980 62. Clark RJ Diabetes and the accompanying hyperglycemia impairs cardiomyocyte calcium cycling through increased nuclear O-GlcNAcylation J. Biol. Chem. 2003 278 44230 44237 10.1074/jbc.M303810200 12941958 63. Mailleux F Gélinas R Beauloye C Horman S Bertrand L O-GlcNAcylation, enemy or ally during cardiac hypertrophy development? Biochim. Biophys. Acta 2016 1862 2232 2243 10.1016/j.bbadis.2016.08.012 27544701 64. Fülöp N Impact of Type 2 diabetes and aging on cardiomyocyte function and O-linked N-acetylglucosamine levels in the heart Am. J. Physiol. Cell Physiol. 2007 292 C1370 C1378 10.1152/ajpcell.00422.2006 17135297 65. Gianni M Apical ballooning syndrome or takotsubo cardiomyopathy: a systematic review Eur. Heart J. 2006 27 1523 1529 10.1093/eurheartj/ehl032 16720686 66. Dias A Takotsubo syndrome: state-of-the-art review by an expert panel – Part 1 Cardiovasc. Revasc. Med. 2019 20 70 79 10.1016/j.carrev.2018.11.015 30528096 67. Scally C Persistent long-term structural, functional, and metabolic changes after stress-induced (Takotsubo) cardiomyopathy Circulation 2018 137 1039 1048 10.1161/CIRCULATIONAHA.117.031841 29128863 68. Singh T Manganese-enhanced magnetic resonance imaging in Takotsubo Syndrome Circulation 2022 146 1823 1835 10.1161/CIRCULATIONAHA.122.060375 36317524 69. Stiermaier T Transient left atrial dysfunction is a feature of Takotsubo syndrome J. Cardiovasc. Magn. Reson. 2017 19 15 10.1186/s12968-017-0328-8 28162089 70. Walsh-Wilkinson E Segmental analysis by speckle-tracking echocardiography of the left ventricle response to isoproterenol in male and female mice PeerJ 2021 9 e11085 10.7717/peerj.11085 33763310 71. Kneale BJ Gender differences in sensitivity to adrenergic agonists of forearm resistance vasculature J. Am. Coll. Cardiol. 2000 36 1233 1238 10.1016/S0735-1097(00)00849-4 11028476 72. Provost J Simultaneous positron emission tomography and ultrafast ultrasound for hybrid molecular, anatomical and functional imaging Nat. Biomed. Eng. 2018 2 85 94 10.1038/s41551-018-0188-z 31015628 73. Perez-Liva, M. et al. Ultrafast ultrasound imaging for super-resolution preclinical cardiac PET. Mol. Imaging Biol. 10.1007/s11307-020-01512-w (2020). 74. Lamy J Scan-rescan reproducibility of ventricular and atrial MRI feature tracking strain Comput. Biol. Med. 2018 92 197 203 10.1016/j.compbiomed.2017.11.015 29227821 75. Huber AT Cardiac MR strain: a noninvasive biomarker of fibrofatty remodeling of the left atrial myocardium Radiology 2018 286 83 92 10.1148/radiol.2017162787 28813234 76. Voigt J-U Cvijic M 2- and 3-Dimensional myocardial strain in cardiac health and disease JACC Cardiovasc. Imaging 2019 12 1849 1863 10.1016/j.jcmg.2019.01.044 31488253 77. Facchin C FIBER-ML, an open-source supervised machine learning tool for quantification of fibrosis in tissue sections Am. J. Pathol. 2022 192 783 793 10.1016/j.ajpath.2022.01.013 35183511 78. Rappsilber J Mann M Ishihama Y Protocol for micro-purification, enrichment, pre-fractionation and storage of peptides for proteomics using StageTips Nat. Protoc. 2007 2 1896 1906 10.1038/nprot.2007.261 17703201 79. Perez-Riverol Y The PRIDE database resources in 2022: a hub for mass spectrometry-based proteomics evidences Nucleic Acids Res. 2022 50 D543 D552 10.1093/nar/gkab1038 34723319

PMC010xxxxxx/PMC10312499.txt

==== Front bioRxiv BIORXIV bioRxiv Cold Spring Harbor Laboratory 37398029 10.1101/2023.05.30.542963 preprint 2 Article mRNA condensation fluidizes the cytoplasm Xie Ying Liu Tiewei Gresham David http://orcid.org/0000-0002-4028-0364 Holt Liam J http://orcid.org/0000-0002-4002-0861 15 7 2023 2023.05.30.542963http://biorxiv.org/lookup/doi/10.1101/2023.05.30.542963 nihpp-2023.05.30.542963.pdf The intracellular environment is packed with macromolecules of mesoscale size, and this crowded milieu significantly influences cell physiology. When exposed to stress, mRNAs released after translational arrest condense with RNA binding proteins, resulting in the formation of membraneless RNA protein (RNP) condensates known as processing bodies (P-bodies) and stress granules (SGs). However, the impact of the assembly of these condensates on the biophysical properties of the crowded cytoplasmic environment remains unclear. Here, we find that upon exposure to stress, polysome collapse and condensation of mRNAs increases mesoscale particle diffusivity in the cytoplasm. Increased mesoscale diffusivity is required for the efficient formation of Q-bodies, membraneless organelles that coordinate degradation of misfolded peptides that accumulate during stress. Additionally, we demonstrate that polysome collapse and stress granule formation has a similar effect in mammalian cells, fluidizing the cytoplasm at the mesoscale. We find that synthetic, light-induced RNA condensation is sufficient to fluidize the cytoplasm, demonstrating a causal effect of RNA condensation. Together, our work reveals a new functional role for stress-induced translation inhibition and formation of RNP condensates in modulating the physical properties of the cytoplasm to effectively respond to stressful conditions. ==== Body pmc

PMC010xxxxxx/PMC10312914.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398471 10.21203/rs.3.rs-2981903/v1 10.21203/rs.3.rs-2981903 preprint 1 Article A computational approach to demonstrate the control of gene expression via chromosomal access in colorectal cancer Pecka Caleb J. 1 Thapa Ishwor 1 Singh Amar 2 Bastola Dhundy 1* 1 College of IS&T, University of Nebraska at Omaha, Omaha, NE, United States of America 2 Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha, NE, United States of America Authors’ contributions CP was the primary programmer and author for the work presented in this manuscript. IT and DB guided the direction of the project, helped interpret the results, and offered advice in the development of the pipeline. AS was a major contributor in proposing project goals and motivation. All authors read and approved the final manuscript. * Correspondence: dkbastola@unomaha.edu, 1College of IS&T, University of Nebraska at Omaha, Omaha, NE, United States of America 01 6 2023 rs.3.rs-2981903https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs2981903v1.pdf Improved technologies for chromatin accessibility sequencing such as ATAC-seq have increased our understanding of gene regulation mechanisms, particularly in disease conditions such as cancer. This study introduces a computational tool that quantifies and establishes connections between chromatin accessibility, transcription factor binding, transcription factor mutations, and gene expression using publicly available colorectal cancer data. The tool has been packaged using a workflow management system to allow biologists and researchers to reproduce the results of this study. Through the application of this pipeline, we present compelling evidence linking chromatin accessibility to gene expression, with particular emphasis on SNP mutations and the accessibility of transcription factor genes. Furthermore, we have identified significant upregulation of key transcription factor interactions in colon cancer patients, including the apoptotic regulation facilitated by E2F1, MYC, and MYCN, as well as activation of the BCL-2 protein family facilitated by TP73. The code for this project is openly available on GitHub at the following address: https://github.com/CalebPecka/ATAC-Seq-Pipeline/. colorectal cancer epigenetics chromatin accessibility snakemake Fund for Undergraduate Scholarly Experience (FUSE)VA-meritBX002761 National Institute of Health RO1DK124095 ==== Body pmcIntroduction The regulatory mechanisms of gene expression play a critical role in cell differentiation and development, especially in disease conditions such as cancer. Transcription factors (TFs) have been shown to direct the regulation of genes by recognizing transcription factor binding sites (TFBSs) to initiate transcription of downstream genes [1]. TFs are incapable of initiating transcription if their binding site is condensed around a histone octamer structure called a nucleosome. Proteins in the cell can more easily interact with uncondensed chromatin, otherwise called accessible regions of DNA. In this paper, our goal was to design and develop a computational tool that demonstrated the interactions between chromatin access, TF binding, TF mutations, and gene expression using publicly available colorectal cancer (CRC) data. Chromatin accessibility assessment can be accomplished using a variety of protocols including DNase-seq [2], and ATAC-seq [3]. ChIP-seq [4], a protocol that analyzes TF-DNA interactions, can require hundreds of millions of cells as input [3], while the ATAC-seq protocol only requires a standard input of 50,000 cells, making the technique appropriate for research with precious cell types, including cancer cells [3]. The development of chromatin sequencing technology has made it more feasible for researchers to incorporate chromatin accessibility with the analysis of other gene regulation mechanisms. For example, the interaction between chromatin access and transcription has improved predictive models of gene expression based on HiChIP throughput data [5]. Studies have also demonstrated that there is a correlation between chromatin accessibility and gene expression. Pearson correlations have shown that accessible chromatin of promoter regions had similar correlation patterns with gene expression for both healthy and cancerous tissues [6]. Furthermore, it has been shown that inferred TF binding interactions are capable of predicting and differentiating cell types [6]. As researchers improve our ability to assess gene regulation mechanisms, there is an increased need for computational tools that can perform integrative analysis in an accurate and streamlined manner. There is a lack of tools that can perform analysis on the results of chromatin accessibility data in a user-friendly manner. Using a functional programming approach, we designed a reproducible workflow environment that requires minimal systems administration knowledge. Our workflow uses chromatin accessibility data from The Cancer Genome Atlas (TCGA) to predict TFBSs based on motif sequences found in accessible chromatin regions [6]. These results were validated using a database of known TF motifs from JASPAR [7] as well as gene expression profiling data. In addition to statistical validation, we have incorporated a dynamic track-based visualization system that clearly shows the interaction between chromatin accessibility, TF motif sequences, and the genes they regulate. The outputs from our workflow are designed to be compatible with common file formats used in other track-based visualization applications, including the UCSC Genome Browser. Code for this project is publicly available on GitHub (https://github.com/CalebPecka/ATAC-Seq-Pipeline/). Materials and methods Project overview and reproducibility A high-level overview of the project can be seen in Fig 1, including Data Preprocessing, Peak Calling (to identify accessible chromatin regions), Motif Identification (to identify putative TFBSs), Motif Comparison (to compare putative TFBSs against validated databases of TFBSs), Site Matching, Statistical Analysis, and Track-Based Visualization. The high-level overview is color-coded to correspond with individual scripts, inputs, and outputs reflected in the low-level overview given in Fig 2. Our pipeline requires the user to input a set of binary alignment map (BAM) files for analysis. All other input files are either provided in the GitHub repository or automatically installed by the pipeline. For example, the hg38 human reference genome is automatically installed in a Snakemake script. The pipeline outputs files for the genomic location of accessible chromatin regions (Upstream Peaks), motif sequences identified by BCrank (BC Rank Consensus Sequences), and layered genomic visualizations (tracks.png). For full descriptions of these files and all available fields, please refer to our GitHub documentation. We employed Snakemake [8] to automatically detect the progress of the workflow and run necessary code based on a configuration file that can be modified by the user, making it possible for most users to ignore the technical intricacies in our low-level overview. Snakemake automatically installs conda environments required for software dependencies. We have also enabled a parameter to configure the conda environment to perfectly replicate the dependency build used in this study. The perfectly reproducible configuration may require an adaptable installation script provided with the workflow. To the best of our knowledge, there are no other methodologies or tools currently available that can provide a meaningful comparison with our pipeline. Data preprocessing and indexing ATAC-seq, RNAseq, and SNP mutation data for 41 CRC patients were preprocessed by TCGA [6]. The hg38 human reference genome was used as a reference for the Bowtie2 alignment tool [6]. Samtools sorted the mapped reads and Picard removed duplicates, resulting in a set of 41 binary alignment map (BAM) files for each of the patient samples [6]. For patients with multiple ATAC-seq BAM files, our mutation data only contained one instance of each patient ID. Seven more patients were missing mutation data in TCGA, as shown in Table 1. A supplementary file of TCGA barcodes is provided as a Table in Additional File 1. Preprocessing procedures were carried out by TCGA study and are not included in the GitHub pipeline. The pipeline requires the user to input a BAM file for each sample. Peak calling MACS2 was used to identify DNA read fragment pileups, also called chromatin peaks [9]. Chromatin peaks are indicative of DNA regions where chromatin structure is accessible. Each peak contains a summit value, indicating the base-pair location where fragment pileup is highest [9]. The first step of our pipeline compiles a list of all peak summits located 100–1000 base-pairs upstream of all genes at the ENSG level. Motif identification and site matching The 50 base-pair region centered around each peak was extracted as a FASTA sequence using Biostrings and the hg38 reference genome for Homo sapiens [10]. For each of the 41 CRC samples, BCrank created a list of 1200 motifs from the upstream FASTA sequences [11]. BCRank requires the input FASTA sequences to be ordered according to confidence level. Our script automatically sorts the sequences according to q-score, the confidence level provided by MACS2. BCrank was also used to map those motifs onto the upstream FASTA sequences in order to obtain the location of putative TFBSs [11]. We would like to note that this method is limited due to an excess of false positives, a common problem with motif prediction tools. Motif comparison For each sample, the list of 1200 motifs was reformatted to be compatible with MEME Suite tools [12]. A collection of known TFs from JASPAR was converted into a similar format [7]. TomTom searched for pattern matches between the position weighted matrices of our 1200 putative TFBSs against the known JASPAR collection [13]. The FDR p-value correction method was globally employed across the TomTom results for all samples, and non-significant results were removed from the merged list of all patients (padjust ≤ 0.05). Determination of differentially expressed genes The Data Driven Referencing (DDR) method [14] was employed to create a list of differentially expressed (DE) genes for the 31 non-duplicated TCGA CRC barcodes discussed in ”Data preprocessing and indexing”. The DDR method normalizes gene expression levels into 5 tiers based on the relative expression of housekeeping genes in each sample [14]. We chose the DDR method over other gene expression analysis tools because the DDR normalization process is better suited for accounting for non-biological variabilities [14]. Fisher’s Exact Tests were used to determine which genes are enriched in cancer samples versus healthy samples [14]. DDR outputted a list of differentially upregulated and downregulated genes after performing the FDR p-value correction method and subsetting the results (padjust ≤ 0.05). The resulting DE genes as well as fold changes and p-values are provided as a supplementary Table in Additional File 2. Creation of genome tracks Track-based visualizations were created using pyGenomeTracks [15] [16]. Bigwig files were created to visualize chromatin accessibility using the bamCoverage program provided by deepTools [17]. Known TF motifs were collected from the JASPAR database [7]. Tracks containing known gene locations were imported from the hg38 reference genome. Results and discussion Chromatin accessibility across the CRC genome In the Peak Calling step, MACS2 returns a score that quantifies chromatin peak accessibility by comparing the fragment pileup relative to various background regions of fragment pileup at a maximum of 10,000 base pairs away [9]. In the original TCGA study by Corces et al., the researchers noted that the MACS score was problematic due to its variability across different datasets [6]. We hypothesized that this issue may not be relevant in our study because we focused on a specific cohort of CRC patients. The experiments performed in this subsection were intended to investigate this hypothesis. We used a one-tailed wilcoxon rank sum test to compare the mean accessibility scores across all samples in housekeeping genes versus non-housekeeping genes. Our findings revealed that the mean accessibility score for housekeeping genes is significantly higher than non-housekeeping genes (p ≤ 0.05). In some situations, a gene will not have a mean accessibility score because MACS2 did not identify a statistically significant peak upstream of the gene in any of the 38 non-duplicated patient IDs (see Table 1). In these situations, we can quantify gene accessibility based on a second accessibility metric, whether or not a gene has an accessible upstream promoter region in each patient sample. Out of 3805 housekeeping genes [18], 128 were not identified as significant by MACS2. Our search list included 58,387 unique gene symbols, of which a total of 26,274 were not identified as significant by MACS2. Using a Fisher’s Exact Test, we concluded that the variance in chromatin accessibility can be partially explained by whether or not a gene is a housekeeping gene (p ≤ 0.05). The raw data for these statistical tests is provided as a Table in Additional File 3. By definition, we expect housekeeping genes to be expressed in the cell at all times, as they are necessary for basic cellular functions. Therefore, we expect housekeeping genes to also have accessible promoter regions, as they constantly need to be transcribed. Our statistics support these assumptions and help to verify that the MACS2 accessibility score is a useful metric for quantifying chromatin accessibility. Therefore, in the following section (Correlation between chromatin access and gene expression), we use the MACS2 accessibility score to correlate chromatin access with gene expression. In all other cases, we quantify accessibility based on our second accessibility metric. Correlation between chromatin access and gene expression For each gene, Pearson correlations were used to identify a relationship between MACS2 chromatin accessibility scores and the raw HTseq gene expression across all 38 patient samples. A two-tailed Wilcoxon rank sum test was used to compare the correlation coefficients for housekeeping and non-housekeeping genes, and we concluded that these correlations were not significantly different from each other. This specific test is not supportive evidence for the regulation of non-housekeeping genes via chromosomal access. Instead, this result supports the claim that the underlying mechanisms of chromosomal access are not specific to housekeeping or non-housekeeping genes. Similar tests were performed to compare the correlation coefficients for CRC biomarker genes (determined using the DDR method, a tool previously developed in our lab) and non-biomarker genes [14]. A one-tailed Wilcoxon test suggests that correlation coefficients between chromatin access and gene expression are expected to be higher in our list of biomarkers acquired from DDR (p ≤ 0.05). We can interpret this result to explain the mechanisms of differential gene expression. Genes which have differential gene expression patterns can be closely tied to a respective increase or decrease in chromatin access. Motif comparison The 1200 motifs produced by BCrank were a highly conservative estimate of the number of motifs necessary to describe the global pattern of TF binding in a patient. For each patient, BCrank calculated 100 motifs to describe a global optimum set. This procedure was repeated 12 times with a different seed generation each time. TomTom is a commonly used tool that identifies pattern matches between two sets of motif sequences [13]. Using TomTom, we calculated that, on average, 68% of our predicted motifs had a statistically significant pattern match with known JASPAR motifs [13]. Using BCrank, these pattern matches can be mapped back to the hg38 reference genome as predicted locations of TFBSs. An example of these predictions is show-cased in Fig 3. These visualizations clearly illustrate the predictive capabilities of the pipeline we developed for matching known and putative motifs within accessible chromatin regions of promoter regions that regulate gene expression. Integrated mutation data and interaction of gene regulation mechanisms To model the interaction of gene regulation mechanisms, we employed Sankey diagrams as shown in Fig 4, 5, 6, and 7. These models showcase the epigenetic interactions in a subset of TFs (accessible and nonaccessible, for example). Each TF is classified in terms of their DE pattern, presence or absence of SNP mutations, and DE patterns in the TF’s target genes. Raw data for these measurements is provided as a supplementary table in Additional File 4. SNP mutations lead to structural changes in the mutated TF. In many cases, this behavior prevents TF binding to the promoter DNA sequence, causing downregulation in that TF’s target genes. Verified target genes for TF regulation were identified using the Harmonizome CHEA Transcription Factor Targets data set [19]. The regulation of target genes was estimated by subtracting the percentage of differentially downregulated target genes in the DDR list from the percentage of differentially upregulated target genes. If the TF target genes were not found in the list of DE genes from DDR, they were not included in the final column of the Sankey diagram. A list of TF genes and their targets used in this analysis is provided as a table in Additional File 5. From these results, we observed that accessible TF genes (Fig 4) were more likely to regulate DE genes than unaccessible TF genes (Fig 5). Furthermore, we observed that accessible TF genes were more likely to produce DE TFs than unaccessible TF genes. Similar diagrams were produced to compare nonmutated TFs (Fig 6) and mutated TFs (Fig 7). In general, mutated genes were more likely to produce differentially downregulated targets. We also observe that upregulated and accessible TFs are more likely to produce downregulated target genes if the TF was mutated, as compared to the nonmutated group. We expected this result because conformational changes to the TF structure prevent it from correctly binding to the TFBS of promoter regions in the TF’s target genes. These expected behaviors help us verify that our pipeline is accurately explaining gene regulation mechanisms and has the potential to be applied to other data sources to understand the origin of disease conditions. Deviations in the relationship between gene expression of TF accessibility could be explained by gene expression mechanisms unrelated to TFs. For example, gene expression is globally increased in larger cell volumes, rather than on a gene-to-gene basis [20]. RNA polymerase II holoenzyme expression scales with cell volume [20], possibly explaining the mechanism of global gene expression increase, even in genes with downregulated transcription factors. Indeed, experimental models of distributions of gene expression profiles have been improved using exponentially scaling of cell volume [21], as well as other non-TF related gene expression mechanisms including dosage compensation, the exponential rate of mRNA maturation, and the first-order kinetics rate of mRNA decay [21]. Depletion of the cohesin complex and CTCF has been experimentally shown to both upregulate and downregulate gene expression, explained by a variety of mechanisms such as CTCF’s direct binding to the gene’s promoter region [22]. It is also important to recognize that mutations in TFs do not guarantee a loss in gene expression. For example, mutations in the promoter binding site region of the TF are far more likely to compromise the TF’s functionality, as well as regions that complex to other transcription mediators. To further complicate matters, it has been shown that TF families with similar binding site regions are able to substitute mutated TFs, supplementing and reducing the mutation’s impact on overall gene expression [23]. In the future, the overall effectiveness of our tool could greatly be improved by thoroughly investigating and categorizing the impact of mutations on TFs. Gene regulation mechanisms in CRC To better understand gene regulation mechanisms in CRC, we mapped accessible, upregulated TFs to their target genes using the TRRUSTv2 transcription factor database [24]. TRRUSTv2 was chosen because it is a manually curated list that includes metadata for whether the target genes are activated or repressed [24]. The data was further subset to only include TFs that activate differentially upregulated genes or TFs that repress differentially downregulated genes. As seen in Table 2, a total of 39 transcription factors were identified including CEBPB, the E2F family (E2F1, E2F3, E2F6), the FOX family (FOXA2, FOXL1), MYC, MYCN, and TP73. Transcriptional regulation of BIRC5 via E2F1 has been shown to contribute to the pathogenesis of colorectal cancer [25]. Our data similarly observes that the availability of E2F1 in CRC patients is contributing to the activation of BIRC5, which is also upregulated in CRC. Notably, several upregulated TFs in CRC are known to interact with apoptosis-regulating genes. E2F1 and MYC both upregulate TP73 (see Table 2). Supporting our observation, TP73 has been shown to block transactivation of TP53, further preventing mechanisms of apoptosis [26]. Additionally, E2F1 and MYCN have repressive interactions with the downregulated TP53 gene. TP73, BBC3, and PMAIP1 were all found to be differentially upregulated in the CRC data. TP73 functions as a transcription factor that activates BBC3 and PMAIP1, both members of the BCL-2 protein family. Many members of the BCL-2 protein family have been targeted for CRC treatment due to their implications in apoptosis [27]. These observations showcase the potential of our pipeline to easily identify mechanisms of gene regulation in disease. In this case, upregulation of E2F1 and MYC in colon cancer leads to the activation of TP73, further leading to the activation of apoptosis regulators. Conclusion In this paper, we presented a computational approach to quantify gene regulation mechanisms, including TF binding and chromatin accessibility. The advancement of chromatin accessibility technologies like ATAC-seq presents an exciting opportunity to increase our understanding of gene regulation mechanisms in various disease conditions. We validated our theoretical model by showing that there is a statistical relationship between chromatin access and gene expression data in CRC, especially in genes that encode TFs. We believe that this model has great potential to be applied to additional data sets to improve our understanding of the underlying mechanisms behind differential gene expression in other disease conditions, especially in the context of genetic mutations. Funding This work was supported by the Fund for Undergraduate Scholarly Experience (FUSE) grant provided by The University of Nebraska at Omaha Office of Research and Creative Activity (https://www.unomaha.edu/office-of-research-and-creativeactivity/students/fuse.php). No grant numbers are assigned to this funding. Additionally, this work was supported in part by the funds from VA-merit award (BX002761) and National Institute of Health RO1 grant funding (DK124095; to ABS) (https://www.nih.gov/grants-funding). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Availability of data and materials Publicly available data from The Cancer Genome Atlas (TCGA) was used for this study. TCGA Barcode sequences for all used samples can be found in Additional File 1: TCGA Barcodes. Figure 1 High-level overview of the project pipeline. Data from Raw FASTQ Reads, TCGA reference genome, JASPAR Transfac motifs, and Data Driven Referencing (DDR) Expression are processed by a series of scripts that lead to a Track-Based Visualization and Statistical Analysis. Figure 2 Low-level overview of the project pipeline. Inputs and commands are color-coded to reflect the high-level overview shown in Fig 1. Figure 3 Example track-based visualization of the SLC04A1 gene. The x-axis displays the chromosomal position of our data. The top track (Peak Chromatin Accessibility) represents read fragment pileup as a method of quantifying chromatin accessibility. The blue box track beneath it (Narrow Peaks) represents MACS2’s interpretation of peak regions, as well as each peak’s summit location (black tick marks in each box). The GTF genes track shows the full reference genome labeling, as well as a small directional arrow to showcase whether a gene is transcribed on the positive or negative strand. The BCRANK Motifs track showcases our predicted motifs mapped onto the reference genome. The Known JASPAR TFBS track showcases how closely mapped our predicted motifs line up with validated TFBS resources. Figure 4 Sankey diagram for subset of TFs with accessible chromatin promoter regions. A subset of this TF list is categorized as differentially and non-differentially expressed (Non DE). This data is further categorized into mutated vs non-mutated TFs. Finally, their target genes are categorized in the right-most column as generally upregulated vs downregulated. Figure 5 Sankey diagram for subset of TFs with unaccessible chromatin promoter regions. These TFs are more likely be Non DE and produce fewer DE target genes than the accessible TFs in Fig 4. Figure 6 Sankey diagram for subset of TFs without mutations. TFs that are upregulated have a roughly equal likelihood of producing a differentially upregulated target gene or a downregulated target gene. Figure 7 Sankey diagram for subset of TFs with mutations. The subset of TFs that is upregulated and accessible is far more likely to produce a downregulated target when the structural features of the TF are compromised by gene mutation. Table 1 List of data subsets and the respective number of samples in each group. Data subset Number of samples ATAC-seq BAM files 41 Non-duplicated patient IDs 38 Non-duplicated patient IDs with corresponding mutation data 31 Table 2 Notable TF-target gene interactions. TF Target genes Interaction Type Target ED Pval adjust E2F1 BIRC5 Activation 1.307 1.46e-16 CEBPB CXCL8 Activation 1.881 2.12e-20 CEBPB GDF15 Activation 2.050 2.47e-31 FOXA2 MMP7 Activation 2.056 4.00e-30 FOXA2 ABCA1 Repression −1.386 3.32e-18 FOXL1 BMP4 Activation 1.381 1.14e-15 E2F1 TP73 Activation 0.355 1.66e-06 E2F1 TP53 Repression −0.207 3.28e-06 MYC TP73 Activation 0.355 1.66e-06 MYCN TP53 Repression −0.207 3.28e-06 TP73 BBC3 Activation 0.805 7.35e-10 TP73 PMAIP1 Activation 0.978 6.02e-27 List of DE target genes regulated by TFs that were found to be differentially upregulated and accessible in our data set. Target ED is a logarithmic adjustment of fold change for the target gene versus wild type, taken from DDR. Pval adjust is the p-value associated with the Target ED. Ethics approval and consent to participate Not applicable Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. AdditionalFile1.csv AdditionalFile2.csv AdditionalFile3.csv AdditionalFile4.csv AdditionalFile5.csv AdditionalFileLegends.pdf ==== Refs References 1. Todeschini A.-L. , Georges A. , Veitia R.A. : Transcription factors: specific dna binding and specific gene regulation. Trends in genetics 30 (6 ), 211–219 (2014)24774859 2. Song L. , Crawford G.E. : Dnase-seq: a high-resolution technique for mapping active gene regulatory elements across the genome from mammalian cells. Cold Spring Harbor Protocols 2010 (2 ), 5384 (2010) 3. Buenrostro J.D. , Wu B. , Chang H.Y. , Greenleaf W.J. : Atac-seq: a method for assaying chromatin accessibility genome-wide. Current protocols in molecular biology 109 (1 ), 21–29 (2015) 4. Landt S.G. , Marinov G.K. , Kundaje A. , Kheradpour P. , Pauli F. , Batzoglou S. , Bernstein B.E. , Bickel P. , Brown J.B. , Cayting P. , : Chip-seq guidelines and practices of the encode and modencode consortia. Genome research 22 (9 ), 1813–1831 (2012)22955991 5. Schmidt F. , Kern F. , Schulz M.H. : Integrative prediction of gene expression with chromatin accessibility and conformation data. Epigenetics & chromatin 13 (1 ), 1–17 (2020)31918747 6. Corces M.R. , Granja J.M. , Shams S. , Louie B.H. , Seoane J.A. , Zhou W. , Silva T.C. , Groeneveld C. , Wong C.K. , Cho S.W. , : The chromatin accessibility landscape of primary human cancers. Science 362 (6413 ) (2018) 7. Fornes O. , Castro-Mondragon J.A. , Khan A. , Van der Lee R. , Zhang X. , Richmond P.A. , Modi B.P. , Correard S. , Gheorghe M. , Baranašić D. , : Jaspar 2020: update of the open-access database of transcription factor binding profiles. Nucleic acids research 48 (D1 ), 87–92 (2020) 8. older F. , Jablonski K.P. , Letcher B. , Hall M.B. , Tomkins-Tinch C.H. , Sochat V. , Forster J. , Lee S. , Twardziok S.O. , Kanitz A. , : Sustainable data analysis with snakemake. F1000Research 10 (2021) 9. Zhang Y. , Liu T. , Meyer C.A. , Eeckhoute J. , Johnson D.S. , Bernstein B.E. , Nusbaum C. , Myers R.M. , Brown M. , Li W. , : Model-based analysis of chip-seq (macs). Genome biology 9 (9 ), 1–9 (2008) 10. Pages H. , Aboyoun P. , Gentleman R. , DebRoy S. , Pages M.H. , IRanges L. , BSgenome S. , XStringSet-class R. , MaskedXString-class R. , XKeySortedData-class R. : Package ‘Biostrings’ (2019) 11. Ameur A. : BCRANK: predicting binding site consensus from ranked DNA sequences (2019) 12. Bailey T.L. , Johnson J. , Grant C.E. , Noble W.S. : The meme suite. Nucleic acids research 43 (W1 ), 39–49 (2015) 13. Gupta S. , Stamatoyannopoulos J.A. , Bailey T.L. , Noble W.S. : Quantifying similarity between motifs. Genome biology 8 (2 ), 1–9 (2007) 14. Zhang L. , Thapa I. , Haas C. , Bastola D. : Multiplatform biomarker identification using a data-driven approach enables single-sample classification. BMC bioinformatics 20 (1 ), 1–13 (2019)30606105 15. Ramírez F. , Bhardwaj V. , Arrigoni L. , Lam K.C. , Grüning B.A. , Villaveces J. , Habermann B. , Akhtar A. , Manke T. : High-resolution tads reveal dna sequences underlying genome organization in flies. Nature communications 9 (1 ), 1–15 (2018) 16. Lopez-Delisle L. , Rabbani L. , Wolff J. , Bhardwaj V. , Backofen R. , Grüning B. , Ramírez F. , Manke T. : pygenometracks: reproducible plots for multivariate genomic datasets. Bioinformatics 37 (3 ), 422 (2021)32745185 17. Ramírez F. , Ryan D.P. , Grüning B. , Bhardwaj V. , Kilpert F. , Richter A.S. , Heyne S. , Dündar F. , Manke T. : deeptools2: a next generation web server for deep-sequencing data analysis. Nucleic acids research 44 (W1 ), 160–165 (2016) 18. Eisenberg E. , Levanon E.Y. : Human housekeeping genes, revisited. TRENDS in Genetics 29 (10 ), 569–574 (2013)23810203 19. Lachmann A. , Xu H. , Krishnan J. , Berger S.I. , Mazloom A.R. , Ma’ayan A. : Chea: transcription factor regulation inferred from integrating genome-wide chip-x experiments. Bioinformatics 26 (19 ), 2438–2444 (2010)20709693 20. Padovan-Merhar O. , Nair G.P. , Biaesch A.G. , Mayer A. , Scarfone S. , Foley S.W. , Wu A.R. , Churchman L.S. , Singh A. , Raj A. : Single mammalian cells compensate for differences in cellular volume and dna copy number through independent global transcriptional mechanisms. Molecular cell 58 (2 ), 339–352 (2015)25866248 21. Cao Z. , Grima R. : Analytical distributions for detailed models of stochastic gene expression in eukaryotic cells. Proceedings of the National Academy of Sciences 117 (9 ), 4682–4692 (2020) 22. Zuin J. , Dixon J.R. , van der Reijden M.I. , Ye Z. , Kolovos P. , Brouwer R.W. , van de Corput M.P. , van de Werken H.J. , Knoch T.A. , van IJcken W.F. , : Cohesin and ctcf differentially affect chromatin architecture and gene expression in human cells. Proceedings of the National Academy of Sciences 111 (3 ), 996–1001 (2014) 23. Lu R. , Rogan P.K. : Transcription factor binding site clusters identify target genes with similar tissue-wide expression and buffer against mutations. F1000Research 7 (2018) 24. Han H. , Cho J.-W. , Lee S. , Yun A. , Kim H. , Bae D. , Yang S. , Kim C.Y. , Lee M. , Kim E. , : Trrust v2: an expanded reference database of human and mouse transcriptional regulatory interactions. Nucleic acids research 46 (D1 ), 380–386 (2018) 25. Xu F. , Xiao Z. , Fan L. , Ruan G. , Cheng Y. , Tian Y. , Chen M. , Chen D. , Wei Y. : Rfwd3 participates in the occurrence and development of colorectal cancer via e2f1 transcriptional regulation of birc5. Frontiers in cell and developmental biology 9 (2021) 26. Grob T. , Novak U. , Maisse C. , Barcaroli D. , thi A. , Pirnia F. , Hügli B. , Graber H. , De Laurenzi V. , Fey M. , : Human δnp73 regulates a dominant negative feedback loop for tap73 and p53. Cell Death & Differentiation 8 (12 ), 1213–1223 (2001)11753569 27. Ramesh P. , Medema J.P. : Bcl-2 family deregulation in colorectal cancer: potential for bh3 mimetics in therapy. Apoptosis 25 (5 ), 305–320 (2020)32335811

PMC010xxxxxx/PMC10312916.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398464 10.21203/rs.3.rs-3001450/v1 10.21203/rs.3.rs-3001450 preprint 1 Article Economic Burden of Sanfilippo Syndrome in the United States http://orcid.org/0000-0002-6975-8990 Ashby Frederick University of Florida College of Medicine Park Haesuk University of Florida College of Pharmacy Svensson Mikael University of Florida College of Pharmacy Heldermon Coy University of Florida College of Medicine Authors’ contributions FA conceived the study, designed the methods, performed analysis, interpreted the data, created the initial draft, and revised drafts. HP designed methods, assisted in data interpretation, and revised drafts. KS designed methods, assisted in data interpretation, and revised drafts. CH designed methods, assisted in data interpretation, and revised drafts. ✉ ricky.ashby@ufl.edu 12 6 2023 rs.3.rs-3001450https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs3001450v1.pdf Introduction: Sanfilippo syndrome is a rare disease and fatal genetic disorder in the United States with no FDA-approved treatment, and no comprehensive assessment of economic disease burden is available. Objective To develop a model to estimate the economic burden associated with Sanfilippo syndrome in the United States (US) using valued intangibles (disability-adjusted life years lost) and indirect burden (lost caregiver productivity) from 2023 onward. Design and Setting: A multistage comorbidity model was generated using publicly available literature on Sanfilippo syndrome disability, and 14 disability weights from the 2010 Global Burden of Disease Study. Attributable increase in caregiver mental health burden and caregiver productivity loss were also estimated using data from the CDC National Comorbidity Survey, retrospective studies on caregiver burden in Sanfilippo syndrome, and the Bureau of Labor Statistics. Monetary valuations were adjusted to USD 2023 and given a 3% discount rate from 2023 onward. Main Outcomes and Measures: Year-over-year incidence and prevalence of Sanfilippo syndrome was calculated for each age group in each year, and year-over-year disability-adjust life years (DALYs) lost due to patient disability was calculated by comparing to health-adjusted life expectancy (HALE), considering years of life lost (YLLs) due to premature mortality and years lived with disability (YLDs). Intangibles were valued in USD 2023, adjusted for inflation and discounted to provide economic burden of disease. Results From 2023–2043, overall economic burden in the US attributable to Sanfilippo syndrome was estimated to be $1.55 billion USD with current standard of care. The burden to individual families exceeded $5.86 million present value from time of birth per child born with Sanfilippo syndrome. These figures are also a conservative estimate, since they do not consider direct cost associated with the disease, as extensive primary data on the direct healthcare cost of Sanfilippo syndrome does not currently exist in the literature. Conclusions and Relevance: Sanfilippo syndrome is a rare lysosomal storage disease, however the severe burden associated with the disease for individual families demonstrates a profound cumulative impact. Our model represents the first disease burden estimate associated with Sanfilippo syndrome. This underscores the substantial morbidity and mortality burden of Sanfilippo syndrome. NIH NINDSR01NS102624 ==== Body pmcIntroduction Sanfilippo syndrome (MPS III) is a rare mucopolysaccharidosis within the lysosomal storage disease category, which is characterized by an inability to break down heparan sulfate, a glycosaminoglycan, which leads to accumulation, inflammation, oxidative stress and ultimately the debilitating manifestations of the disease.[1] MPS III is categorized into four subtypes (A, B, C and D) based on which enzyme is affected in the heparan sulfate breakdown pathway,[2] and hence each would require different biological products to replace the defective enzyme. Each subtype has been described with different levels of severity and life-expectancy,[3] yet clinically they remain virtually indistinguishable without molecular testing.[4] Like many rare diseases, there is currently no cure for the MPS III subtypes, however multiple clinical trials have been initiated to attempt enzyme replacement or reduction of heparan sulfate storage.[2, 5, 6] While there are clinical trials for MPS III subtypes A and B, many have failed or been discontinued.[7] Reasons for discontinuation often stem not only from lack of efficacy, but also from perceived small market size, difficulty in designing a trial with a small patient population, lack of clear clinical end-points and challenges demonstrating statistical power and efficacy sufficient for FDA approval in a short time frame.[8–11] Thus, the potential risks to the company performing the trial have historically been seen as incommensurate with the potential financial return, yet trial funding trends in the last few decades have shifted this narrative and demonstrated potential market viability.[12, 13] Clinical trial costs have historically been a barrier to initiating or continuing clinical trials for rare childhood diseases, considering the risk of failure, and can make investment unappealing.[8–11] Government initiatives in the United States (US), such as the FDA Office of Office of Orphan Products Development, and the NIH Office of Rare Diseases have been formed out the recognition that collectively, these diseases carry great societal burden, and society benefits from public funding to stimulate research and development with the intention of creating marketable products eventually. Despite the perspective that researching orphan diseases is expensive and generally not viable in the market, we found few or no peer-reviewed investigations into what the actual cost of most rare diseases are, particularly MPS III. Thus, we aimed to estimate the economic burden of MPS III including: (1) the patient intangible health losses expressed through monetary estimates of disability-adjusted life years (DALYs) lost, (2) the caregiver intangible health losses, and (3) economic burden of disease when considering lost productivity of caregivers. Methods Model Overview and Assumptions Projected burden of disease estimates were calculated using a model accounting for comorbidity[14] on a multistage scheme for MPS III patients born from 1992–2100 in the US (Fig. 1), presuming only comorbidities associated with MPS III (Supplementary Table 1). A total of 1,073 MPS III patients were simulated with a lifelong time horizon. In the same timeframe, 2,146 parents with a time horizon of 25-year after onset of child’s symptoms. Burden of disease considered DALYs lost in the patient due to illness, both parents due to mental health, and lost productivity due to caregiver burden. The value of labor for parents was set to $75,424 annually ($52,378 in wages, $23,046 in social fees, averaged between male and female) based on the US Census and Bureau of Labor Statistics 2019 and it was presumed that the equivalent of one full-time person worth of labor would be lost fulltime during the symptomatic phase of the disease until death of the patient. We obtained the forecasted birth rate in the US, as calculated by the Institute for Health Metrics and Evaluation (IHME), through GHDx the dataset “Global Fertility, Mortality, Migration, and Population Forecasts 2017–2100.”[15] Due to the unpredicted decline in US fertility rate 2020–2022 compared to the IHME models, this analysis makes the presumption that US fertility rates will normalize to IHME predictions by 2025. The natural sex ratio of male to female live births of 105 to 100 was used for male and female estimate stratifications (UN World Population Prospects 2022). A live-birth prevalence of MPS III in the US of 0.27 per 100,000 live births was used, which was derived from National MPS Society between 1995–2005.[16] Disability Adjusted Life Year Modeling: Patients Health-adjusted life expectancy at birth for the US was presumed 65.2 for males and 67.0 for females based on the World Health Organization 2022 report.[17] Due to clinical similarity between subtypes,[18] our multi-stage model used the presumption that the average disease course will manifest similarly between subtypes in terms of stage onsets.[1, 4] Disability weight values were derived from the Global Burden of Disease Study 2010.[19] The presumed occurrence of each MPS III subtype was estimated based on the natural rate of each subtype observed in France, the United Kingdom, Greece, and Australia cumulatively: 64.2% MPS IIIA; 19.8% MPS IIIB; 9.9% MPS IIIC; and 6.2% MPS IIID (Fig. 2a).[20, 21] Multi-stage Comorbidity Disability Weight Modeling The average non-adjusted life expectancy or each subtype of MPS III was used: MPS IIIA 15.22 years; MPS IIIB 18.91 years; MPS IIIC 23.43 years.[22] Average life expectancy of MPS IIID was not available, so the weighted average of every other subtype was used (16.86 years) -- this was close to the average lifespan of two reported MPS IIID cases (14 years).[21] The comorbidity calculations for MPS III emulated a three-stage natural history previously described.[1, 4, 22] For stage 1 (onset age 1–4), the presumed average age of onset was the midpoint of the range of onset (2.5 years old), similar with stage 2 (4.0 years old) and stage 3 (11.5 years old). The first 2.5 years of life (Stage 0) were considered healthy for this model,[23, 24] and the final years of life were presumed stage 3 until death. The cumulative disability weight of Stage 0 was 0; Stage 1 was 0.149;[1, 3, 4, 25–28] Stage 2 was 0.357;[1, 4, 25, 27–31] and Stage 3 was 0.68[1, 4, 26–36] (Supplementary Table 1). Disability Adjusted Life Year Modeling: Caregivers Caregiver DALYs were factored into caregiver economic burden (Supplementary Table 2). For major depressive disorder (MDD) an odds ratio (OR) of 2.90 for mothers and 2.42 for fathers was used.[37] Post-traumatic stress disorder (PTSD) prevalence in parents of MPS III children was presumed to be 26.9% in mothers, and 15.8% in fathers. These statistics reference a retrospective study of MPS III parents in the Netherlands,[37] which was the only available public study on the prevalence of depression, anxiety, and PTSD in caregivers of MPS III patients. We utilized a simple and widely used logistical regression model to obtain estimated risk ratio (eRR)[38] instead of using the OR, which can overestimate the risk, [39–41] and US baseline prevalence of these mental health disorders. The baseline expected US prevalence of clinical depression and general anxiety disorder were presumed to be 10.4% and 19.0% for mothers, and 5.5% and 11.9% in fathers.[42, 43] The baseline expected prevalence of PTSD was presumed to be 5.2% in mothers and 1.8% in fathers based on the National Comorbidity Survey.[44] Because there is very limited data on how long depression, anxiety and PTSD persist in parents of MPS III children, the presumed length of disease was approximated to 25 years after onset of symptoms, regardless of MPS III subtype. Disease Burden Estimates A monetary value per each DALY of $114,339 was used, based on a comprehensive study investigating cost-effectiveness across all US healthcare services from 1996–2016.[45] We used this as our primary analysis since it represents status quo for US healthcare services. A 3% annual discount rate was used accruing from 2023 onward[46]. We also conducted a sensitivity analysis with similar parameters using upper limit of $150,000 and a lower limit of $69,499, since this is reported as an international cost-effectiveness threshold for very-high income countries.[47] A one-time 17% inflation factor was applied to convert USD 2019 to USD 2023 (Bureau of Labor Statistics, 2023). Results DALYs Lost per MPS III Birth with Current Standard of Care The multistage model used for this analysis demonstrated an average DALY loss of 55.80, 54.46, 53.10 and 55.08, for a male child born in the US with MPS IIIA, MPS IIIB, MPS IIIC and MPS IIID, respectively. The DALY loss for a female MPS III child born in the US was 58.18, 57.06, 55.70 and 57.68, respectively (Supplementary Table 3). Caretaker disease burden for each child born with MPS III, regardless of child sex at birth, was 2.08 DALYs lost for the child’s father and 4.40 DALYs for the child’s mother (Supplementary Table 3). Estimated Economic Burden of MPS III in the United States After applying the IHME projections for US birthrate, presuming 0.27 MPS III cases per 100,000 live-birth in the US, a year-over-year estimate of the total number of undiscounted DALYs lost was generated using a dynamic cohort starting at those born in 1992 and ending with those born in 2100 (Supplementary Fig. 1). Based on the US birthrate estimates and incidence of MPS III, there are an estimated average of 11 MPS III births each year, and due to the presumed declining US birth rate, this eventually reaches only 8 annual MPS III births by 2100. Based on the weighted average lifespan of MPS III patients, it was also derived that the estimated total US population of MPS III patients may be at a steady state of around 185–195 patients from 2020–2053, with a slow decline that correlates with a declining US birthrate. This timeframe was the focus of the analysis (Fig. 2b). The total estimated undiscounted societal economic burden on an annual basis were also extrapolated and visualized by subtype (Fig. 2c) and patient economic burden compared to caregiver economic burden (Fig. 2d). Burden of MPS III Cases 2023–2043 The analysis performed for all MPS subtypes demonstrated a cumulative discounted societal burden of $1.55 billion between 2023–2043 alone, with an upper limit of $1.97 billion and a lower limit of $1.02 billion. In total, 79.7% of this amount was from MPS III patient burden, and 21.3% representing caregiver burden. By subtype, burden composition was similar between subtypes: MPS IIIA 19.0% caregiver burden (Fig. 3a); MPS IIIB 22.2% caregiver burden (Fig. 3b); MPS IIIC 24.9% caregiver burden (Fig. 3c); MPS IIID 20.2% caregiver burden (Fig. 3d). Total burden from 2023–2043 by subtype was $973 million for MPS IIIA, $315 million for MPS IIIB, $159 million for MPS IIIC and $99 million for MPS IIID. The total discounted economic burden caused for each individual birth, is $6.50 million for MPS IIIA; $6.13 million for MPS IIIB; $5.86 million for MPS IIIC; and $6.43 million for MPS IIID (Fig. 3f). This was calculated by simulating a single hypothetical patient, discounting from year of birth, and taking into account caregiver burden. Economic burden composition was similar regardless of year, as this analysis presumes no improvement in standard of care for MPS III. For reference, the dynamic present value of damages each subtype of MPS III up until a given year between 2023–2053 is plotted for MPS IIIA (Fig. 4a), MPS IIIB (Fig. 4b), MPS IIIC (Fig. 4c), MPS IIID (Fig. 4d). Cumulative burden for all subtypes from 2023 to a given year until 2053 are graphed in Fig. 4e with an increased y-axis range. Discussion In this paper, publicly available data was analyzed, and our results suggest that MPS III will continuously remain with an economic burden in the US with an estimated $1.55 billion per year between 2023 and 2043 including children’s lives lost, caretaker mental health, and lost wages. Each family who bears a child afflicted with MPS III can expect to lose on average between $0.89 and $1.32 million in lost wages alone and lose between $4.54 and $5.61 million in collective DALYs lost – for both patient and caregiver. This analysis also suggests that the US MPS III patient population may currently be at a steady-state of around 185 patients, which represents 0.97–1.54% of the entire estimated MPS III patient population worldwide,[48] a higher share than we were expecting given this disease is usually thought of as heavily skewed to specific countries with high carrier prevalence.[49] One important finding to our analysis is that the caregiver economic burden was relatively small in comparison to the societal value of DALYs lost in the child (~ 25% of total). While this analysis focused on MPS III, a rare lysosomal storage disease in the US (0.27 per 100,000 live births), it is important to consider the litany of neglected rare diseases have been estimated to cost our society $966 billion per year according to the Government Accountability Office (GAO).[50] While no one can predict how soon an effective treatment can be developed before a randomized-control trial is done, understanding the value of the disease can help inform policy and funding decisions. For MPS III, the cumulative disease burden without considering direct cost is estimated to be around $100 million (USD 2023) per year for the next few decades based on our analysis. Investing in a cure would potentially be highly valuable considering the burden associated with MPS III, and the current lack of treatment or palliation. The severe burden to each family with an MPS III child is quite substantial to consider, as diseases such as diabetes have a total yearly cost of around $11,700-$13,241 when considering both direct and indirect costs,[51] despite getting substantially more research funding than MPS III. Currently in the US, there is not even newborn screening for MPS III, which by itself may be cost-effective in driving new research forward and prevent the stress of delayed diagnosis.[52–55] One of the biggest hurdles to MPS III clinical trials is simply acquiring enough patients for enrollment, and newborn screening would make this process an order of magnitude simpler. While our analysis provides a novel perspective on the societal burden of MPS III, there are obvious limitations that should be considered. As with any estimates, these analyses are based on previously reported disease birth-prevalence, and these may be underestimated due to missed or undiagnosed patients. This report also makes presumptions about a consistent disease birth-prevalence and a predictable birth-rate within the US. This analysis did not calculate direct costs associated with the disease, which would increase the overall economic burden and require extensive primary data collection from MPS III families for accurate assessment. Our analysis also did not consider future immigration/emigration from the US, which has the potential to profoundly skew the prevalence of child-bearing heterozygous disease carriers if involving high prevalence regions, such as Europe and the Middle East. We also took a conservative stance in our analysis about the value of lost productivity from the MPS III patients directly, so it is important to note that a functional cure would have implications for marginally increasing the workforce as well. Regarding our analysis, we emphasize that the combined economic burden of over $5.86 million per case of MPS III born is a representation of the total potential present value of curative treatment from the point of childbirth, not a price target. While complete resolution of the disease is typically the lofty goal of any investigational treatment, future treatments of MPS III will likely recover a fraction of this burden. In the case of MPS III, if treatment plans could recover just 25% of morbidity and mortality associated with the burden of disease for MPS III, it would potentially be collectively valued at tens of millions of dollars per year based on our estimates. Conclusions This estimate is currently, to our knowledge, the only estimate of disease burden associated with Sanfilippo syndrome in the United States, in DALYs as well as valuation of disease burden. We also illustrate novel methods for contextualizing rare disease burden in the US. Policy makers and advocates need to consider the disease burden associated with rare diseases in the context of cumulative damages caused rather than the disease prevalence in isolation, as the later can mask the true theoretical value of funding research in these diseases. Acknowledgements The National MPS Society for feedback and support for the project. Funding This work was funded by NIH NINDS R01NS102624 Availability of data and material All data utilized in the manuscript is publicly available as cited. The simulation was performed by Microsoft Excel, and the original work is available from the corresponding author upon reasonable request. Units of Measurement USD 2023 Represents United States dollar accounting for 2023 value. DALYs Disability adjusted life years, typically expressed as those lost due to disease. HALYs Health adjusted life years, a measure of life expectancy which also considers health. YLL Years of life lost, the health adjusted life years lost due to mortality. YLD Years lived with disability, the health adjusted life years lost due to morbidity. Abbreviations and Symbols BLS Bureau of Labor Statistics CDC Center for Disease Control and Prevention (United States) DALY Disability adjusted life year eRR Estimated risk ratio or estimated relative risk FDA Food and Drug Administration (United States) GAO Government Accountability Office GHDx Global health data exchange HALY Health adjusted life year ICER Institute for Clinical and Economic Review IHME Institute for Health Metrics and Evaluation MDD Major depressive disorder MPS Mucopolysaccharidosis OR Odds ratio PTSD Post-traumatic stress disorder RR Risk ratio or relative risk UN United Nations US United States USD United States Dollar YLL Years of life lost YLD Years lived with disability WHO World Health Organization $ US dollars Figure 1 Concept Diagram of Simulated Patient. A conceptual example of a single MPS III patient is shown for the purposes of illustration. Patient burden was calculated by intangibles (DALYs), which included comorbidity after age of onset for each stage of the disease and expected age of death (Year X). The first 2.5 years of life are presumed normal, with no change in disease burden. Economic burden of DALYs for the patient, parents as well as lost productivity for each respective year after onset of symptoms was assigned to their chronologically respective year so discounting rates for the dollar amount could be applied appropriately. Figure 2 Annual Undiscounted Economic Burden of MPS III in the US. Total undiscounted DALYs within the US are shown above (a) from year 2020–2053, and are based on the IHME projections for the US birthrate, with adjustments from 2020–2025 to account for unpredicted decline in birthrate. These estimates are broken down by MPS III subtype, and presume subtype distribution close to natural rates previously observed in the literature (b). Year-over-year economic burden is shown (c and d) subcategorized by MPS III subtype (c) and patient (DALYs lost) versus caregiver (DALYs lost and lost wages) burden (d). The above figures presume a DALY value of $114,339 in the US, which is the status quo of US healthcare 1996–2016. Figure 3 US Societal Burden of MPS III from 2023–2043. The above graphs visualize the presumed disparities between MPS III subtypes regarding discounted burden components (a-d). Overall, all subtypes had between 19.8%–26.1% of their burden composition associated caregivers. When parsing out differences between intangible value lost and lost productivity for parents (d), all subtypes cost families between $0.89m-$1.32m in lost wages alone, and between $4.54m-$5.61m in DALYs lost (both patient and parents). Figure 4 Cumulative Societal Burden of MPS III 2023–2053. The following figures demonstrate the cumulative societal burden of not finding a cure for MPS III by subtype (a-d), starting in 2023 as a zero point, and cumulatively increasing from there. The total cumulative burden for all subtype is also shown (e). The average US healthcare value per DALY of $114,339 was used as the primary analysis, with an upper limit of $150,000 (ICER) and a lower limit of $69,499 was used. An inflation factor of 17% was used to convert to USD 2023 (a-e), and a standard 3% discount rate was used. Competing interests Dr Coy D Heldermon Stock in Lacerta, a gene therapy company. Patent pending for an AAV gene therapy vector for Sanfilippo Type B. Declarations Ethics approval and consent to participate Not applicable Consent for publication Not applicable Birth Rate: CDC NVSS for past years. https://www.cdc.gov/nchs/nvss/births.htm IHME 2017–2100 projection for future years. https://ghdx.healthdata.org/record/ihme-data/global-population-forecasts-2017-2100 Live-birth prevalence of MPS III[16] MPS subtype composition estimation[20, 21] Average wage value for a parent (2019): https://www.census.gov/library/publications/2020/demo/p60-270.html#:~:text=The%202019%20real%20median%20earnings,2018%20ratio%20(Figure%205). Composition of social fees (benefits; 2019): https://www.bls.gov/news.release/pdf/ecec.pdf Disability Weights[19] ==== Refs References 1. Valstar MJ , Ruijter GJ , van Diggelen OP , Poorthuis BJ , Wijburg FA . Sanfilippo syndrome: a mini-review. J Inherit Metab Dis. 2008;31 (2 ):240–52.18392742 2. Pearse Y , Iacovino M . A Cure for Sanfilippo Syndrome? A Summary of Current Therapeutic Approaches and their Promise. Med Res Arch 2020, 8 (2 ). 3. van de Kamp JJ , Niermeijer MF , von Figura K , Giesberts MA . Genetic heterogeneity and clinical variability in the Sanfilippo syndrome (types A, B, and C). Clin Genet. 1981;20 (2 ):152–60.6796310 4. Muschol N , Giugliani R , Jones SA , Muenzer J , Smith NJC , Whitley CB , Donnell M , Drake E , Elvidge K , Melton L , Sanfilippo syndrome: consensus guidelines for clinical care. Orphanet J Rare Dis. 2022;17 (1 ):391.36303195 5. Seker Yilmaz B , Davison J , Jones SA , Baruteau J . Novel therapies for mucopolysaccharidosis type III. J Inherit Metab Dis. 2021;44 (1 ):129–47.32944950 6. Ashby F , Heldermon CD . The significance of triple-capsid-mutant AAV8 for treatment of Sanfilippo Syndrome Type B. Archives of Stem Cell and Therapy 2022, 3(1 ). 7. Ghosh A , Shapiro E , Rust S , Delaney K , Parker S , Shaywitz AJ , Morte A , Bubb G , Cleary M , Bo T , Recommendations on clinical trial design for treatment of Mucopolysaccharidosis Type III. Orphanet J Rare Dis. 2017;12 (1 ):117.28651568 8. Rees CA , Pica N , Monuteaux MC , Bourgeois FT . Noncompletion and nonpublication of trials studying rare diseases: A cross-sectional analysis. PLoS Med. 2019;16 (11 ):e1002966.31751330 9. Bell SA , Tudur Smith C . A comparison of interventional clinical trials in rare versus non-rare diseases: an analysis of ClinicalTrials.gov. Orphanet J Rare Dis. 2014;9 :170.25427578 10. Mulberg AE , Bucci-Rechtweg C , Giuliano J , Jacoby D , Johnson FK , Liu Q , Marsden D , McGoohan S , Nelson R , Patel N , Regulatory strategies for rare diseases under current global regulatory statutes: a discussion with stakeholders. Orphanet J Rare Dis. 2019;14 (1 ):36.30736861 11. Mellerio JE . The challenges of clinical trials in rare diseases. Br J Dermatol. 2022;187 (4 ):453–4.36192346 12. Yates N , Hinkel J . The economics of moonshots: Value in rare disease drug development. Clin Transl Sci. 2022;15 (4 ):809–12.35334152 13. Hughes DA , Poletti-Hughes J . Profitability and Market Value of Orphan Drug Companies: A Retrospective, Propensity-Matched Case-Control Study. PLoS ONE. 2016;11 (10 ):e0164681.27768685 14. Hilderink HB , Plasmans MH , Snijders BE , Boshuizen HC , Poos MJ , van Gool CH . Accounting for multimorbidity can affect the estimation of the Burden of Disease: a comparison of approaches. Arch Public Health. 2016;74 :37.27551405 15. Vollset SE , Goren E , Yuan CW , Cao J , Smith AE , Hsiao T , Bisignano C , Azhar GS , Castro E , Chalek J , Fertility, mortality, migration, and population scenarios for 195 countries and territories from 2017 to 2100: a forecasting analysis for the Global Burden of Disease Study. Lancet. 2020;396 (10258 ):1285–306.32679112 16. Puckett Y , Mallorga-Hernandez A , Montano AM . Epidemiology of mucopolysaccharidoses (MPS) in United States: challenges and opportunities. Orphanet J Rare Dis. 2021;16 (1 ):241.34051828 17. World health statistics. 2022: monitoring health for the SDGs, sustainable development goals. In : World Health Organization; 2022 : 125. 18. Cleary MA , Wraith JE . Management of mucopolysaccharidosis type III. Arch Dis Child. 1993;69 (3 ):403–6.8215557 19. Salomon JA , Vos T , Hogan DR , Gagnon M , Naghavi M , Mokdad A , Begum N , Shah R , Karyana M , Kosen S , Common values in assessing health outcomes from disease and injury: disability weights measurement study for the Global Burden of Disease Study 2010. Lancet. 2012;380 (9859 ):2129–43.23245605 20. Meikle PJ , Hopwood JJ , Clague AE , Carey WF . Prevalence of lysosomal storage disorders. JAMA. 1999;281 (3 ):249–54.9918480 21. Arfi A , Richard M , Gandolphe C , Bonnefont-Rousselot D , Therond P , Scherman D . Neuroinflammatory and oxidative stress phenomena in MPS IIIA mouse model: the positive effect of long-term aspirin treatment. Mol Genet Metab. 2011;103 (1 ):18–25.21353610 22. Lavery C , Hendriksz CJ , Jones SA . Mortality in patients with Sanfilippo syndrome. Orphanet J Rare Dis. 2017;12 (1 ):168.29061114 23. Kuiper GA , Meijer OLM , Langereis EJ , Wijburg FA . Failure to shorten the diagnostic delay in two ultra-orphan diseases (mucopolysaccharidosis types I and III): potential causes and implications. Orphanet J Rare Dis. 2018;13 (1 ):2.29310675 24. Wijburg FA , Wegrzyn G , Burton BK , Tylki-Szymanska A . Mucopolysaccharidosis type III (Sanfilippo syndrome) and misdiagnosis of idiopathic developmental delay, attention deficit/hyperactivity disorder or autism spectrum disorder. Acta Paediatr. 2013;102 (5 ):462–70.23336697 25. Ruijter GJ , Valstar MJ , van de Kamp JM , van der Helm RM , Durand S , van Diggelen OP , Wevers RA , Poorthuis BJ , Pshezhetsky AV , Wijburg FA . Clinical and genetic spectrum of Sanfilippo type C (MPS IIIC) disease in The Netherlands. Mol Genet Metab. 2008;93 (2 ):104–11.18024218 26. Valstar MJ , Marchal JP , Grootenhuis M , Colland V , Wijburg FA . Cognitive development in patients with Mucopolysaccharidosis type III (Sanfilippo syndrome). Orphanet J Rare Dis. 2011;6 :43.21689409 27. Bianchi PM , Gaini R , Vitale S . ENT and mucopolysaccharidoses. Ital J Pediatr. 2018;44 :57–66.29784060 28. Mesolella M , Cimmino M , Cantone E , Marino A , Cozzolino M , Della Casa R , Parenti G , Iengo M . Management of otolaryngological manifestations in mucopolysaccharidoses: our experience. Acta Otorhinolaryngol Ital. 2013;33 (4 ):267–72.24043915 29. Escolar ML , Jones SA , Shapiro EG , Horovitz DDG , Lampe C , Amartino H . Practical management of behavioral problems in mucopolysaccharidoses disorders. Mol Genet Metab. 2017;122S :35–40.29170079 30. Bax MC , Colville GA . Behaviour in mucopolysaccharide disorders. Arch Dis Child. 1995;73 (1 ):77–81.7639557 31. Fraser J , Wraith JE , Delatycki MB . Sleep disturbance in mucopolysaccharidosis type III (Sanfilippo syndrome): a survey of managing clinicians. Clin Genet. 2002;62 (5 ):418–21.12431260 32. Heron B , Mikaeloff Y , Froissart R , Caridade G , Maire I , Caillaud C , Levade T , Chabrol B , Feillet F , Ogier H , Incidence and natural history of mucopolysaccharidosis type III in France and comparison with United Kingdom and Greece. Am J Med Genet A. 2011;155A (1 ):58–68.21204211 33. Valstar MJ , Neijs S , Bruggenwirth HT , Olmer R , Ruijter GJ , Wevers RA , van Diggelen OP , Poorthuis BJ , Halley DJ , Wijburg FA . Mucopolysaccharidosis type IIIA: clinical spectrum and genotype-phenotype correlations. Ann Neurol. 2010;68 (6 ):876–87.21061399 34. Valstar MJ , Bruggenwirth HT , Olmer R , Wevers RA , Verheijen FW , Poorthuis BJ , Halley DJ , Wijburg FA . Mucopolysaccharidosis type IIIB may predominantly present with an attenuated clinical phenotype. J Inherit Metab Dis. 2010;33 (6 ):759–67.20852935 35. Gilkes JA , Heldermon CD . Mucopolysaccharidosis III (Sanfilippo Syndrome)-disease presentation and experimental therapies. Pediatr Endocrinol Rev. 2014;12 (Suppl 1 ):133–40.25345095 36. Scarpa M , Lourenco CM , Amartino H . Epilepsy in mucopolysaccharidosis disorders. Mol Genet Metab. 2017;122S :55–61.29170080 37. Conijn T , Nijmeijer SCM , van Oers HA , Wijburg FA , Haverman L . Psychosocial Functioning in Parents of MPS III Patients. JIMD Rep. 2019;44 :33–41.29980992 38. Zhang J , Yu KF . What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA. 1998;280 (19 ):1690–1.9832001 39. Knol MJ , Le Cessie S , Algra A , Vandenbroucke JP , Groenwold RH . Overestimation of risk ratios by odds ratios in trials and cohort studies: alternatives to logistic regression. CMAJ. 2012;184 (8 ):895–9.22158397 40. Davies HT , Crombie IK , Tavakoli M . When can odds ratios mislead? BMJ 1998, 316 (7136 ):989–91.9550961 41. Mantel N , Haenszel W . Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst. 1959;22 (4 ):719–48.13655060 42. Brody DJ , Pratt LA , Hughes JP . Prevalence of Depression Among Adults Aged 20 and Over: United States, 2013–2016. NCHS Data Brief 2018 (303 ):1–8. 43. Terlizzi EP , Villarroel MA . Symptoms of Generalized Anxiety Disorder Among Adults: United States, 2019. NCHS Data Brief 2020 (378 ):1–8. 44. Kessler RC , Berglund P , Chiu WT , Demler O , Heeringa S , Hiripi E , Jin R , Pennell BE , Walters EE , Zaslavsky A , The US National Comorbidity Survey Replication (NCS-R): design and field procedures. Int J Methods Psychiatr Res. 2004;13 (2 ):69–92.15297905 45. Weaver MR , Joffe J , Ciarametaro M , Dubois RW , Dunn A , Singh A , Sparks GW , Stafford L , Murray CJL , Dieleman JL . Health Care Spending Effectiveness: Estimates Suggest That Spending Improved US Health From 1996 To 2016. Health Aff (Millwood). 2022;41 (7 ):994–1004.35787086 46. Sanders GD , Neumann PJ , Basu A , Brock DW , Feeny D , Krahn M , Kuntz KM , Meltzer DO , Owens DK , Prosser LA , Recommendations for Conduct, Methodological Practices, and Reporting of Cost-effectiveness Analyses: Second Panel on Cost-Effectiveness in Health and Medicine. JAMA. 2016;316 (10 ):1093–103.27623463 47. Daroudi R , Akbari Sari A , Nahvijou A , Faramarzi A . Cost per DALY averted in low, middle- and high-income countries: evidence from the global burden of disease study to estimate the cost-effectiveness thresholds. Cost Eff Resour Alloc. 2021;19 (1 ):7.33541364 48. Kong W , Wu S , Zhang J , Lu C , Ding Y , Meng Y . Global epidemiology of mucopolysaccharidosis type III (Sanfilippo syndrome): an updated systematic review and meta-analysis. J Pediatr Endocrinol Metab. 2021;34 (10 ):1225–35.34271605 49. Zelei T , Csetneki K , Voko Z , Siffel C . Epidemiology of Sanfilippo syndrome: results of a systematic literature review. Orphanet J Rare Dis. 2018;13 (1 ):53.29631636 50. Rubin R. The Cost of Rare Diseases. JAMA. 2021;326 (20 ):2000. 51. American Diabetes A. Economic Costs of Diabetes in the U.S. in 2017. Diabetes Care. 2018;41 (5 ):917–28.29567642 52. Bodamer OA , Giugliani R , Wood T . The laboratory diagnosis of mucopolysaccharidosis III (Sanfilippo syndrome): A changing landscape. Mol Genet Metab. 2014;113 (1–2 ):34–41.25127543 53. NORD. : Barriers to Rare Disease Diagnosis, Care and Treatment in the US: A 30-Year Comparative Analysis. 2020. 54. Pollard S , Weymann D , Dunne J , Mayanloo F , Buckell J , Buchanan J , Wordsworth S , Friedman JM , Stockler-Ipsiroglu S , Dragojlovic N , Toward the diagnosis of rare childhood genetic diseases: what do parents value most? Eur J Hum Genet. 2021;29 (10 ):1491–501.33903739 55. Rehm HL . Time to make rare disease diagnosis accessible to all. Nat Med. 2022;28 (2 ):241–2.35132266

PMC010xxxxxx/PMC10312917.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37397998 10.21203/rs.3.rs-3017935/v1 10.21203/rs.3.rs-3017935 preprint 1 Article An all-ultrasound cranial imaging method to establish the relationship between cranial FUS incidence angle and transcranial attenuation in non-human primates in 3D Singh Aparna 1 Jiménez-Gambín Sergio 1 Konofagou Elisa E 12* 1. Department of Biomedical Engineering, Columbia University, New York, NY, USA 2. Department of Radiology, Columbia University, New York, NY, USA Contributions A.S., and S.J.G participated in data collection. A.S, S.J.G, and E.E.K assisted in preparation and interpretation of the data. A.S. prepared Figures 1–5 with input from S.J.G, and E.E.K. The first draft of the manuscript was written by A.S, S.J.G, and E.E.K. * ek2191@columbia.edu 13 6 2023 rs.3.rs-3017935https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs3017935v1.pdf Focused ultrasound (FUS) is a non-invasive and non-ionizing technique which deploys ultrasound waves to induce bio-effects. When paired with acoustically active particles such as microbubbles (MBs), it can open the blood brain barrier (BBB) to facilitate drug delivery inhibited due to the presence of BBB. One of the parameters that affects the FUS beam propagation is the beam incidence angle on the skull. Prior work by our group has shown that, as incidence angles deviate from 90°, FUS focal pressures attenuate and result to a smaller BBB opening volume. The incidence angles calculated in our prior studies were in 2D and used skull information from CT. The study presented herein develops methods to calculate incidence angle in 3D in non-human primate (NHP) skull fragments using harmonic ultrasound imaging without using ionizing radiation. Our results show that ultrasound harmonic imaging is capable of accurately depicting features such as sutures and eye-sockets of the skull. Furthermore, we were able to reproduce previously reported relationships between the incidence angle and FUS beam attenuation. We also show feasibility of performing ultrasound harmonic imaging in in-vivo non-human primates. The all-ultrasound method presented herein combined with our neuronavigation system stands to increase more widespread adoption of FUS and render it accessible by eliminating the need for CT cranial mapping. ==== Body pmcIntroduction In neurological diseases such as epilepsy, Parkinson’s, and Alzheimer’s, the first line of treatment is typically pharmacological. These medications, however, have not been shown efficacious over time due to developed tolerance10 and progression of disease. In those cases, brain stimulation methodologies, such as deep brain stimulation (DBS) can treat the symptoms especially in advanced stages of Parkinson’s disease where in one randomized trial, DBS of subthalamic nucleus caused greater improvements from baseline to six months in quality of life in Parkinson’s disease patients when compared to medication alone10. However, DBS has resulted in serious adverse events in 13 percent of the cases including a fatal intracerebral hemorrhage, possibly due to it being invasive in nature12. Another widely used brain stimulation method, transcranial magnetic stimulation (TMS), has shown reduction in seizures in patients with epilepsy with longer stimulation groups reporting fewer seizures than shorter stimulation groups19. However, TMS suffers from poor depth penetration and limited spatial resolution2,43. Focused ultrasound (FUS) is an exciting, relatively new alternative technique as it is not only inherently non-invasive, but also achieves greater depth penetration. Several clinical trials have shown the potential of FUS for treating essential tremors13,26, for providing pain relief5 and, for blood-brain barrier (BBB) opening applications1,29. For all FUS-guided therapies, the gold standard method for targeting is MRI1,6,13,26. MRI provides excellent tissue contrast with tissue temperature monitoring capabilities42 and helps with predicting therapeutic outcomes such as FUS ablation. However, it fails to monitor microbubble activity, which is paramount to safe and successful FUS BBB opening procedures. A second imaging modality that is widely used for FUS pre-planning is CT as it provides with important acoustic parameters of the skull and brain that are essential to predicting FUS beam path, attenuation of FUS pressures due to thickness and density of skull and predicting the incidence angle of FUS beams onto the skull20,21,45. However, CT is costly, ionizing, and without intra-monitoring capabilities. Another alternative for guidance and particle activity monitoring is ultrasound. An ultrasound imaging array, when set on receive mode, can monitor this microbubble activity at high frame rates and inform us of safety of neuromodulation procedures or blood-brain barrier (BBB) opening. Past research published by our group has developed methods to detect microbubble activity during BBB opening39,40. Other significant developments have been reported over the past few years where ultrasound imaging arrays have been used to identify anatomical sutures on mice skull and has been used to guide blood-brain barrier opening in small animals and deliver molecules across the BBB7,8. Additionally, advancements in B-mode 20image processing have made it possible for ultrasound imaging transducers to be used for brain vascular imaging11,14, perform transcranial imaging through the human skull31, and for detecting functional activity in brain in small animals3,28 and in newborns9. Overall, these recent developments of ultrasound-guided focused ultrasound technologies enabled the use of transcranial power Doppler image to guide BBB opening in rats36. Ultrasound monitoring has also played a huge role in other neuromodulation procedures that do not involve BBB opening. In one study by our group25, researchers showed that during neuromodulation of peripheral nervous system (PNS), ultrasound imaging transducer can be used to image real time displacement and cavitation, thereby informing us about the intricate interplay of cavitation and displacement in causing neuromodulation of PNS. In addition, other groups have also shown that ultrasound can be used to detect drug release from nanoparticles after FUS application24. Thus, ultrasound imaging guidance and monitoring can provide an efficient, reliable and promising alternative for FUS applications. In this study, we aimed to further explore the capability of real-time ultrasound harmonic B-Mode images for FUS applications. Studies published by our group has shown that incidence angle, angle between the normal vector to skull plane and the transducer plane, is critical for the reliable and reproducible BBB opening in non-human primates20and humans4,22,23. In this study, we developed methods to use an all-ultrasound device which can perform cranial imaging and predict FUS incidence angles on the skull using ultrasound imaging. We validated this technique in two ex-vivo NHP skulls and showed the feasibility of transcranial imaging in two in-vivo NHP experiments. Results: K-wave simulation predictions: Before performing k-wave simulations, we rotated the skull using the ‘imrotate3’ function of MATLAB. We, then, calculated the incidence angle by computing the angle between the normal vector fitted on to the skull plane and the normal vector corresponding to the transducer plane (Fig 2b). The results of k-Wave simulations (Fig 1) show that the presence of skull at different incidence angles attenuate pressure fields where the pressure attenuated by as much as 47% when the incidence was 85.7° in Fig 1b. The attenuation further increased to up to 60% when incidence angle was 67.4° in Fig 1c. We then compared how attenuation was affected at different FUS incidence angles to the skull in Fig 1f. We observed that as the incidence angle approximated to 90 degrees, the attenuation decreased and that there was a strong correlation (R2 = 0.84) between the attenuation and FUS incidence angle. We then used all our eight simulations and compared the focal shift with respect to the free field in Fig 1d and found that the average axial and lateral focal shift was 5.0±2.4mm and 0.38±0.27mm respectively. 3-D reconstruction of the cranial NHP skull map: To generate a skull map, we manually segmented the skull from each of the 500 reconstructed, harmonic B-Mode images with a resolution of 0.14mm × 0.14mm × 0.2mm (Fig 2b). We then assembled each of those planes to generate a skull mask using the volshow function of MATLAB. The 3D skull map reconstructed from 2D raster scans for skull number 1 in Fig 3a shows the skull sutures identified using the red arrows. These sutures are also present in the physical skull in Fig 3b. We then identified similar slices in CT (Fig 3c) and reconstructed harmonic B-Mode (Fig 3d) via comparing features (in yellow arrows) present in both CT and B-Mode slices. Thereafter, we calculated the thickness of the bone in those slices. We compared bone thickness in a total of 3 slices in Fig 3e and found that the thickness computation from B-Mode and CT slices were comparable. In slice 1, the measurements computed from CT vs B-mode were equal to 4.9mm vs 5.1mm, in slice 2 it was 3.62mm vs 3.95mm, and in slice 3 was 8.5mm vs 8.9mm. The 3D skull map for NHP skull # 2 in Fig 3f clearly shows eye sockets and features that are present in the orthogonal skull in Fig 3g. We compared the physical measurements of the eye sockets computed via calipers (Mitutoyo, Aurora, IL) with that of computed from B-Mode at 4 different positions shown via different color arrows in Fig 3h. We found that the measurements, of calipers vs B-Mode, for location 1 was 19.34mm vs 20.7mm, for location 2 was 22.57mm vs 24.7mm, for location 3 was 24mm vs 22.17mm, and for location 4 was 7.6mm vs 8.17mm. Ex-vivo incidence angle estimation using harmonic B-Mode imaging and FUS pressure attenuation: To calculate the incidence angle through two ex-vivo NHP skulls, we acquired and segmented 70 slices of harmonic B-Mode images, each 0.5mm apart using the imaging transducer that was co-axially aligned with our FUS transducer. Our imaging sequence comprised 256 diverging waves acquired at 2-MHz with phased array (P4–2, ATL, Philips). B-mode images were acquired at a depth of 110mm. To obtain a single harmonic B-Mode image, a 2-cycle diverging wave at 2MHz followed by another 2-cycle diverging wave with opposite polarity was delivered. The final reconstructed B-Mode image was then saved onto the computer and the transducer was then moved to the next plane 0.5mm away to ultimately acquire 70 slices. We used this partial 3D skull map in Fig. 4a for both skulls to estimate FUS incidence angle on the skull (Fig 2c). We then performed FUS sonications with our H-231 FUS single-element transducer (Sonic Concepts, Bothell, WA). Our FUS transducer (F0 = 250 kHz, OD = 110mm, ID = 44mm, and focal distance = 110mm), was coaxially aligned with the imaging transducer. FUS pulses of 15 cycles at a PRF of 100 Hz at 0.3 MPa were transmitted thereafter and we used bullet hydrophone (H0400, ONDA Corporation, Sunnyvale, CA) to record field pressures underneath the skull. For skull #1 in Fig 4b, we observe that the FUS incidence angle on the skull impacts the intensity of transcranial FUS pressure field recorded via hydrophone. As a result, similar to what was predicted in the simulation an incidence angle of 86.7 degrees shows lower attenuation than incidence angle of 44.3 degrees. When attenuation results are combined for both skulls in Fig 4c, the same linear relationship as that of the simulation is obtained whereas the incidence angle approached 90 degrees, the attenuation decreased. The dependence of attenuation on the incidence angle was found to be high (R2 = 0.81). We also calculated average axial and lateral focal shifts in Fig 4d from the beam path generated and found an average axial focal shift of 6.54±4.81mm and lateral focal shift of 2.31±1.19mm. When examining incidence angles vs attenuation, there were 3 incidence angles in Fig 4e that were comparable between simulation and experimental condition. An incidence angle of 85.7 degrees in simulation resulted in 39% attenuation whereas an incidence of 86.7 degrees in experimental condition resulted in 45% attenuation. Similarly, an incidence angle of 76.7 degrees in simulation resulted in 49% attenuation whereas an incidence of 75.4 degrees in experimental condition resulted in 48% attenuation. Finally, an incidence angle of 79.6 degrees in simulation resulted in 46% attenuation whereas an incidence of 79.4 degrees in experimental condition resulted in 49% attenuation. In-vivo harmonic cranial B-Mode imaging: We performed harmonic cranial B-Mode imaging in two in-vivo NHPs (Rhesus Macaques, Male, 8 years old) using our ultrasound guided FUS setup driven by Robotic Arm (Universal Robots, UR5E). In our first NHP imaging study, we collected coronal slices. In the original B-Mode image in Fig 5a, we can identify skin, muscle, and skull. We manually segmented out the skull from original B-Mode image (Fig 5a) and created 3D skull map in Fig 5b. In the 3D skull map rendered using the ‘volshow’ function of MATLAB, sutures inside the skull surface are depicted which is a characteristic of NHP skull. In our second in-vivo setting in Fig 5c, we performed imaging in the sagittal plane. Similar to our first in-vivo imaging case, we identified skin, muscle, and skull layers. We manually segmented out the skull and assembled all 500 slices to generate a 3D skull map in Fig 5d. In the sagittal plane, we were able to recover eye sockets. Discussion: FUS is a non-invasive and non-ionizing therapeutic technology that can treat neurological conditions by focusing ultrasound waves at desired target regions. Due to its non-invasive nature, it has been FDA approved for ablation for uterine fibroids and essential tremors. The clinical method for targeting FUS is using MRI thermometry26,32,46. To target accurately and calculate FUS attenuation for transcranial applications, CT images of the skull are utilized18,27. Such targeting methods have been used towards blood brain barrier opening and neuromodulation procedures. However, using MRI and CT collectively for targeting and trajectory planning can render FUS ionizing and less accessible. In this study, we demonstrated the feasibility of using transcranial ultrasound B-Mode imaging to compute the incidence angle during FUS for BBB opening in NHP using a clinical system35. Using an all-ultrasound system is advantageous when compared to other modalities such as MRI and CT. In addition to providing real-time guidance, ultrasound also provides tools for real-time monitoring thereby, enabling performing repeated FUS procedures reliably and cost-efficiently. Our neuronavigation clinical system, which has been used in this study, has successfully performed BBB opening and by developing methods to image skulls and accurately predict FUS incidence angles, we can enhance the applicability of our clinical system for FUS procedures. We showed that we can successfully use this clinical system to achieve our desired objective and that our experimental results are comparable to simulations Our simulation results showed a determinant coefficient of 0.85 between FUS angle incidence and attenuation. This is in line with published studies where a similar relationship and determinant coefficient was seen between volume of BBB opening and incidence angle at fixed input pressures 20,21. Cranial harmonic images of the skull accurately depicted skull landmarks: Cranial B-Mode images of ex-vivo skulls in Fig 3 were able delineate skull boundaries. This enabled segmentation of the skull and eventually helped with 3D skull map reconstruction. The 3D skull map was able to show sutures and eye sockets present on the skull. Furthermore, B-Mode images were able correctly infer the skull thickness and eye socket width measurements. These will be unprecedentedly advantageous for targeting purposes as these anatomical sutures and locations can be further used with our neuronavigation system for real-time registration and FUS guidance4,22,34,35,44. Comparison between simulation and experimental findings: Our ex-vivo pressure measurements were found to be in good agreement with simulations. A determinant coefficient of 0.81 was observed in experimental condition which was in line with our simulations. Ex-vivo pressure attenuation measurements closely resembled simulation pressure attenuation. When comparing attenuation values between simulation and physical measurements, the average error was within 6%. Furthermore, the average axial focal shifts were similar in magnitude when experimental conditions were compared to simulations. However, the lateral focal shift was found to be 2mm greater in experiments than in simulations. The average lateral shift in both cases was calculated by averaging contributions from all incidence angles. In experimental conditions, some of the incidence angles were lower than 45 degrees. On the other hand, in the simulations, lowest incidence angle was 69.8 degrees. A wider of incidence angle may have, thus, contributed to larger average lateral shift the in experimental condition. In-vivo harmonic imaging revealed certain features on the NHP skull: We further showed feasibility of acquiring harmonic B-Mode images in two different in-vivo cases. In both cases in Fig 5, the skin, muscle, and skull layers could be clearly distinguished. Furthermore, after segmentation and reconstruction of skull map, we were able to identify sutures and key anatomical landmarks such as eye sockets. These features will enable us to use harmonic B-Mode imaging derived landmarks to guide and open the BBB using our clinical system in NHP. While our method can perform cranial harmonic B-Mode imaging and predict the angle of incidence in transcranial FUS applications, it has some limitations. This method relies on slice-by-slice segmentation which can present some challenges. Manual segmentation can be time consuming and subjective and need to be done off-line. Thus, to render this procedure clinically relevant, segmentation needs to be performed automatically with real-time feedback such as through the use of machine learning approaches. Conclusion In this study, we developed cranial harmonic B-Mode imaging method to predict angle of incidence for FUS therapies, thereby eliminating the need to use ionizing methods, such as CT, for targeting purposes. We, first, performed cranial imaging using our clinical neuronavigation system, with ultrasound imaging transducer coaxially aligned with FUS transducer, on ex-vivo skulls. We used these cranial images to predict FUS beam incidence angle on the skull and compared it FUS pressure field recorded, via bullet hydrophone, at those respective incidence angles. The results showed a decreasing trend in attenuation as incidence angle approximated to normal. This agreed with our k-wave simulation results and previously published results Click or tap here to enter text. We, then, showed feasibility of performing harmonic cranial imaging in in-vivo NHPs. Our future work will incorporate in-vivo harmonic cranial imaging for targeting blood brain-barrier opening in large animals and clinical studies and will be used to calculate FUS incidence angles during brain therapies in real-time. Materials and Methods: Simulations using k-Wave: Numerical simulations to predict focused ultrasound pressure through NHP skull were modelled using k-Wave package 37,38. This method is selected as it provides low numerical dispersion as compared with finite-differences methods15. First, an ex-vivo NHP skull was degassed for 24 hours. Then, a CT scan of the skull fragment was acquired using a clinical CT scanner (Siemens BioGraph mCT 64 Slices Scanner, Siemens Healthcare), with a resolution of 0.24 × 0.24 × 0.6 mm3. The CT data was converted from Hounsfield Units to heterogeneous acoustic properties of sound speed and density using the linear-piecewise polynomials proposed previously30,41. The absorption of the NHP skull is assumed to be heterogenous where the maximum absorption was 0.68 dB/cm at 250 kHz with a frequency-power law exponent of two33. We used an isotropic grid with a spatial step of 0.4 mm, which corresponds to a spatial sampling of 15 points per wavelength (ppw) in water at the working frequency. Even though the simulation’s convergence, stability and accuracy can be reached working at 5–6 ppw in k-Wave16,17, we used 15 ppw to have a minimum of 8 grid points across the NHP skull thickness to capture microstructure and irregularities of the skull. The numerical temporal step was set to 26.7 ns and 54 ns for the simulations with and without the NHP skull, respectively, leading to a Courant-Friedrichs-Lewy number of 0.2 in both cases. The H-231 FUS single-element transducer (Sonic Concepts, Bothell, WA) was then modelled in k-Wave. A bowl was modelled with dimensions which were comparable to the actual transducer (f0 = 250 kHz, outer diameter (OD) = 110 mm, inner diameter (ID) = 44 mm, and focal distance = 110 mm). The geometric focus was placed 3 cm below the skull surface. The skull was rotated using imrotate3 MATLAB function such that it created different incidence angles. We, then, performed a GPU-accelerated 3D acoustic k-Wave simulation on a workstation PC (Dell) with a NVIDIA Quadro P6000 GPU (Nvidia, Santa Clara, CA). The pixel resolution was 0.4 × 0.4 × 0.4 mm3 with a 3D grid composed of total of 271 × 215 × 215 voxels. The CT data was resampled to fit the k-Wave simulation voxel size. The maximum pressure was recorded for every voxel in the simulation grid. A total of 8 simulations were performed at 8 different incidence angles. Additionally, another simulation was performed to mimic free-field. Resulting pressure fields were used to obtain values of focal shift, focus full width at half maximum (FWHM), and skull insertion loss. The relationship between attenuation and incidence angle was then established. Cranial B-Mode imaging: To obtain 3D transcranial skull map, B-mode images of two ex-vivo skulls were acquired at 2MHz with phased array (P4–2, ATL, Philips) using 256 diverging waves. B-mode images were acquired at a depth of 110mm. To obtain a single B-Mode image, a 2-cycle diverging wave at 2MHz followed by another 2-cycle diverging wave with opposite polarity was delivered to perform harmonic imaging. After acquisition of a B-Mode image of a plane, imaging transducer was moved 0.2mm to acquire image of the next plane. A total of 500 planes were acquired to reconstruct image of an entire skull. This process was repeated for the second skull. After obtaining B-Mode image of each plane, skull was segmented out using ‘roipoly’ function of MATLAB (Fig 2b). After segmenting skull from all planes, a 3D skull map was reconstructed. After raster scan, B-Mode slices of skull was compared to skull CT of skull number 1 to identify similar slices. Once identified, skull thickness from CT was compared to skull thickness obtained from B-Mode slices (Figs 3c and 3d). Skull thickness for skull number 1 was computed for 3 different slices. For skull number 2, calipers were used to measure distances at 4 different positions on the physical skull (Fig 3h). The measurements from calipers were compared to measurements extracted from B-Mode. Our initial 3D scan was aimed at reconstructing the whole skull map. Ex-vivo setup for calculating incidence angle, performing FUS, and collecting pressure field maps: For calculating incidence angle for FUS procedure, we acquired only 70 slices, 0.5mm apart using Robotic Arm (Universal Robots, UR5E). This helped us get 35mm of the NHP ex-vivo skull. We, then, moved the imaging transducer to centermost slice and sonicated the FUS transducer (F0 = 250 kHz, OD = 110mm, ID = 44mm, and focal distance = 110mm), which was coaxially aligned to the imaging transducer. FUS pulses of 15 cycles at a PRF of 100 Hz at 0.3 MPa were transmitted thereafter. The resulting pressure fields were recorded using bullet hydrophone (Fig 2a). A free field recording was also performed in absence of the skull and was compared to pressure fields obtained in the presence of skull. In order to create incidence angle, the skull was manually rotated. A total of 10 pressure fields, including one in free field, was obtained. After obtaining the pressure field and NHP ex-vivo B-Mode images, the images were segmented as previously mentioned. After segmenting the images, the 3D incidence angle was calculated between the central axis of the transducer (propagation direction) and the plane containing the sonicated skull surface (Fig 2c). We first obtained the grid points corresponding to the sonicated skull surface by evaluating the size of the beam covering the skull. Thereafter, we fitted the resulting surface to a plane. Finally, we calculated the incidence angle from the normal vector of this plane where incidence angle was computed to be 90°-α, where α is the angle between the normal vector of the skull fitted plane and the propagation direction vector of the transducer. In-vivo harmonic B-Mode imaging feasibility – To show the feasibility of transcranial imaging in NHPs, we performed in-vivo imaging in two NHPs (Male rhesus-macaques, 8 years old). All procedures were reviewed and approved by the Columbia University Institutional Animal Care and Use Committee and performed in accordance with the relevant guidelines and regulations for animal research. Additionally, our study followed the ARRIVE guidelines. In the first NHP, we imaged coronal slices. For our second NHP, we imaged sagittal slices. We mounted our neuronavigation4,34,35 system onto a robotic arm (Universal Robots, UR5E). Once mounted on the robotic arm, 2 cycle diverging waves at 2MHz followed by 2 cycle diverging waves at 2 MHz with opposite polarity were delivered to acquire harmonic images of the skull. A total of 4 frames were acquired, saved, and then robotic arm was moved to another plane 0.5mm away from the previous plane. This procedure was repeated until 500 planes of data were acquired for each NHP. The planes were then segmented off-line (Fig 5b) and were assembled to create a 3D skull map of in-vivo NHPs. Acknowledgments: This work was funded by National Institute of Health grants R01AG038961and R01EB009041. Data availability: The dataset generated and used in this study will be available from the corresponding author upon request. Fig 1: Effects of different FUS incidence angles on ex-vivo skull predicted by k-wave simulations. a) Lateral and axial pressure fields distributions, respectively, in free field propagations. b) Lateral and axial pressure fields distributions in presence of skull at the best incidence angle of 85.7 degree. c) Lateral and axial pressure fields distributions at an incidence angle of 67.4 degree. d) Average axial and lateral focal shift computed from pressure fields in presence of skull at all incidence angles. e) Graph showing a trend of increased attenuation away from normal incidence angle (90 degrees). Fig 2: Transcranial skull imaging using ultrasound imaging transducer P4–2. a) NHP skull fragment B-Mode images were acquired using P4–2 by performing raster scan of the entire skull. b) B-Mode image (left) was segmented (right) below to only contain the skull. c) Angle between normal vector fitted to the skull plane and normal vector corresponding to transducer plane was used to calculate incidence angle. Fig 3. a) 2D raster scans of skull #1 helped reconstruct 3D skull map. This 3D skull map shows sutures. b) The sutures visible on 3D skull map were also present on the physical skull. c) Skull thickness was computed using CT of the skull. Red line shows the region that was measured. d) B-Mode image of the skull consisted of the same anatomical region. Red line denotes the region where measurements were computed from. e) Skull thickness was computed from 3 different CT and B-Mode slices where anatomical regions could be matched. Comparing skull thickness values evaluated from CT and B-Mode show that the values between them were comparable. f) 2D raster scan of skull #2 helped reconstruct its 3D skull map. We can see the eye sockets along with other prominent structures. g) Physical skull also shows prominent features present on the 3D skull map. h) A total of 4 measurements were taken to compute dimensions of eye sockets and its neighboring areas using its B-Mode image. i) The B-Mode computed measurements were then compared to physical caliper measurements. These measurements were found to be comparable. Fig 4. a) Example of a 3D surface of the skull that was used to calculate incidence angle for both skulls. A total of 9 3D surfaces were used to calculate incidence and angle and establish its relationship with attenuation. b) Pressure fields recorded in free field and transcranially at the best and worst incidence angle show a reduction in attenuation as incidence angle gets further away from 90°. c) Graph that shows relationship between incidence angles, calculated from 3D B-mode images, and attenuation calculated from pressure fields. d) Focal shifts evaluated from pressure field maps recorded using hydrophone at different incidence angles. e) Graph comparing similar incidence angles between ex-vivo and simulations. Fig 5. a) A coronal slice of In-vivo B-Mode scan shows the skin, muscle, and skull. We took multiple B-Mode coronal slices of our first in-vivo NHP and segmented the skull from each of those B-Mode images b) We reconstructed a 3D skull map using multiple coronal slices which shows sutures visible on the skull surface. c) A sagittal slice of our second in-vivo study also reveals the skin, muscle, and skull. We segmented the skull from the B-Mode image. d) After segmenting the skull out from multiple sagittal slices of our second in-vivo study, we reconstructed 3D skull map which shows eye sockets. Competing interests: The author(s) declare no competing interests. ==== Refs Bibliography 1 Abrahao A. , Meng Y. , Llinas M. , Huang Y. , Hamani C. , Mainprize T. , Aubert I. , , “First-in-human trial of blood–brain barrier opening in amyotrophic lateral sclerosis using MR-guided focused ultrasound,” Nat Commun 10 , 4373 (2019). doi:10.1038/s41467-019-12426-9 31558719 2 Airan R. , “Neuromodulation with nanoparticles,” Science (1979) 357 , 465 (2017). doi:10.1126/science.aao1200 3 AU - Bertolo A. , AU - Nouhoum M. , AU - Cazzanelli S. , AU - Ferrier J. , AU - Mariani J.-C. , AU - Kliewer A. , AU - Belliard B. , , “Whole-Brain 3D Activation and Functional Connectivity Mapping in Mice using Transcranial Functional Ultrasound Imaging,” , doi:doi:10.3791/62267. doi:doi:10.3791/62267 4 Bae S. , Pouliopoulos A. , Ji R. , Liu K. , Jiménez-Gambín S. , Yousefian O. , Kokossis D. , , “Transcranial cavitation mapping of blood–brain barrier opening regions in Alzheimer’s disease patients using a neuronavigation-guided focused ultrasound system,” J Acoust Soc Am 153 , A315–A315 (2023). doi:10.1121/10.0018978 5 di Biase L. , Falato E. , Caminiti M. L. , Pecoraro P. M. , Narducci F. , and di Lazzaro V. , “Focused Ultrasound (FUS) for Chronic Pain Management: Approved and Potential Applications.,” doi:10.1155/2021/8438498 6 Chang J. W. , Park C. K. , Lipsman N. , Schwartz M. L. , Ghanouni P. , Henderson J. M. , Gwinn R. , , “A prospective trial of magnetic resonance-guided focused ultrasound thalamotomy for essential tremor: Results at the 2-year follow-up.,” Ann Neurol 83 , 107–114 (2018). doi:10.1002/ana.25126 29265546 7 Choi J. J. , Pernot M. , Small S. A. , and Konofagou E. E. , “Noninvasive, transcranial and localized opening of the blood-brain barrier using focused ultrasound in mice,”, doi:10.1016/j.ultrasmedbio.2006.07.018. doi:10.1016/j.ultrasmedbio.2006.07.018 8 Choi J. J. , Wang S. , Tung Y.-S. , Morrison B. 3rd , and Konofagou E. E. , “Molecules of various pharmacologically-relevant sizes can cross the ultrasound-induced blood-brain barrier opening in vivo,” Ultrasound Med Biol 36 , 58–67 (2010). doi:10.1016/j.ultrasmedbio.2009.08.006 19900750 9 Demene C. , Baranger J. , Bernal M. , Delanoe C. , Auvin S. , Biran V. , Alison M. , , “Functional ultrasound imaging of brain activity in human newborns .,”, doi:10.1126/scitranslmed.aah6756. doi:10.1126/scitranslmed.aah6756 10 Deuschl G. , Schade-Brittinger C. , Krack P. , Volkmann J. , Schäfer H. , Bötzel K. , Daniels C. , , A Randomized Trial of Deep-Brain Stimulation for Parkinson’s Disease A BS TR AC T, (2006). Retrieved from www.nejm.org 11 Errico C. , Pierre J. , Pezet S. , Desailly Y. , Lenkei Z. , Couture O. , and Tanter M. , “Ultrafast ultrasound localization microscopy for deep super-resolution vascular imaging,” Nature 527 , 499–502 (2015). doi:10.1038/nature16066 26607546 12 Fenoy A. J. , and Simpson R. K. J. , “Risks of common complications in deep brain stimulation surgery: management and avoidance.,” J Neurosurg 120 , 132–139 (2014). doi:10.3171/2013.10.JNS131225 24236657 13 Fishman P. S. , Elias W. J. , Ghanouni P. , Gwinn R. , Lipsman N. , Schwartz M. , Chang J. W. , , “Neurological adverse event profile of magnetic resonance imaging-guided focused ultrasound thalamotomy for essential tremor.,” Mov Disord 33 , 843–847 (2018). doi:10.1002/mds.27401 29701263 14 Heiles B. , Chavignon A. , Hingot V. , Lopez P. , Teston E. , and Couture O. , “Performance benchmarking of microbubble-localization algorithms for ultrasound localization microscopy,”, doi:10.1038/s41551-021-00824-8. doi:10.1038/s41551-021-00824-8 15 Jiménez N. , Camarena F. , Redondo J. , Sánchez-Morcillo V. , Hou Y. , and Konofagou E. E. , “Time-domain simulation of ultrasound propagation in a tissue-like medium based on the resolution of the nonlinear acoustic constitutive relations,” Acta Acustica united with Acustica 102 , 876–892 (2016). doi:10.3813/AAA.919002 16 Jiménez-Gambín S. , Jiménez N. , Benlloch J. M. , and Camarena F. , “Holograms to Focus Arbitrary Ultrasonic Fields through the Skull,” Phys Rev Appl 12 , 14016 (2019). doi:10.1103/PhysRevApplied.12.014016 17 Jiménez-Gambín S. , Jiménez N. , Pouliopoulos A. N. , Benlloch J. M. , Konofagou E. E. , and Camarena F. , “Acoustic Holograms for Bilateral Blood-Brain Barrier Opening in a Mouse Model,” IEEE Trans Biomed Eng 69 , 1359–1368 (2022). doi:10.1109/TBME.2021.3115553 34570701 18 Jones R. M. , O’Reilly M. A. , and Hynynen K. , “Experimental demonstration of passive acoustic imaging in the human skull cavity using CT-based aberration corrections,” Med Phys 42 , 4385–4400 (2015). doi:10.1118/1.4922677 26133635 19 Joo E. Y. , Han S. J. , Chung S.-H. , Cho J.-W. , Seo D. W. , and Hong S. B. , “Antiepileptic effects of low-frequency repetitive transcranial magnetic stimulation by different stimulation durations and locations,” Clinical Neurophysiology 118 , 702–708 (2007). doi:10.1016/j.clinph.2006.11.008 17223384 20 Karakatsani M. E. M. , Samiotaki G. M. , Downs M. E. , Ferrera V. P. , and Konofagou E. E. , “Targeting Effects on the Volume of the Focused Ultrasound-Induced Blood–Brain Barrier Opening in Nonhuman Primates <italic>In Vivo</italic>,” IEEE Trans Ultrason Ferroelectr Freq Control 64 , 798–810 (2017). doi:10.1109/TUFFC.2017.2681695 28320656 21 Karakatsani M. E. , Samiotaki G. , Downs M. , Ferrera V. , and Konofagou E. , “Targeting effects on the volume of the focused-ultrasound-induced blood-brain barrier opening in Non-Human Primates in vivo,” in 2015 IEEE International Ultrasonics Symposium, IUS 2015 (2015). doi:10.1109/ULTSYM.2015.0071 22 Konofagou E. , “Neuronavigated focused ultrasound for clinical bbb opening in Alzheimer’s and brain cancer patients,” J Acoust Soc Am 153 , A98–A98 (2023). doi:10.1121/10.0018295 23 Konofagou E. , “Real-time transcranial mapping in non-human primates and human subjects during opening of the blood-brain barrier,” J Acoust Soc Am 153 , A314–A314 (2023). doi:10.1121/10.0018977 24 Lea-Banks H. , O’Reilly M. A. , Hamani C. , and Hynynen K. , “Localized anesthesia of a specific brain region using ultrasound-responsive barbiturate nanodroplets.,” Theranostics 10 , 2849–2858 (2020). doi:10.7150/thno.41566 32194839 25 Lee S. A. , Kamimura H. A. S. , Burgess M. T. , Pouliopoulos A. , and Konofagou E. E. , “Real-Time Displacement and Cavitation Imaging of Non-Invasive Neuromodulation of the Peripheral Nervous System via Focused Ultrasound,” in 2018 IEEE International Ultrasonics Symposium (IUS) (2018) pp. 1–4. doi:10.1109/ULTSYM.2018.8580011 26 Lipsman N. , Schwartz M. L. , Huang Y. , Lee L. , Sankar T. , Chapman M. , Hynynen K. , , “MR-guided focused ultrasound thalamotomy for essential tremor: a proof-of-concept study,” Lancet Neurol 12 , 462–468 (2013). doi:10.1016/S1474-4422(13)70048-6 23523144 27 Liu H. , Sigona M. K. , Manuel T. J. , Chen L. M. , Caskey C. F. , and Dawant B. M. , “Synthetic CT skull generation for transcranial MR imaging–guided focused ultrasound interventions with conditional adversarial networks,” in Proc.SPIE (2022). doi:10.1117/12.2612946 28 MacÉ E. , Montaldo G. , Cohen I. , Baulac M. , Fink M. , and Tanter M. , “Functional ultrasound imaging of the brain,” Nat Methods 8 , 662–664 (2011). doi:10.1038/nmeth.1641 21725300 29 Mainprize T. , Lipsman N. , Huang Y. , Meng Y. , Bethune A. , Ironside S. , Heyn C. , , “Blood-Brain Barrier Opening in Primary Brain Tumors with Non-invasive MR-Guided Focused Ultrasound: A Clinical Safety and Feasibility Study,” Sci Rep 9 , 1–7 (2019). doi:10.1038/s41598-018-36340-0 30626917 30 Mast T. D. , “Empirical relationships between acoustic parameters in human soft tissues,” Acoustics Research Letters Online 1 , 37–42 (2000). doi:10.1121/1.1336896 31 Mozaffarzadeh M. , Verweij M. D. , Daeichin V. , De Jong N. , and Renaud G. , “Transcranial Ultrasound Imaging with Estimating the Geometry, Position and Wave-Speed of Temporal Bone,” in 2021 IEEE International Ultrasonics Symposium (IUS) (2021) pp. 1–4. doi:10.1109/IUS52206.2021.9593826 32 Paff M. , Boutet A. , Boutet A. , Neudorfer C. , Elias G. J. B. , Germann J. , Loh A. , , “Magnetic Resonance-Guided Focused Ultrasound Thalamotomy to Treat Essential Tremor in Nonagenarians,” Stereotact Funct Neurosurg 98 , 182–186 (2020). doi:10.1159/000506817 32224617 33 Pinton G. , Aubry J.-F. , Bossy E. , Muller M. , Pernot M. , and Tanter M. , “Attenuation, scattering, and absorption of ultrasound in the skull bone,” Med Phys 39 , 299–307 (2012). doi:10.1118/1.3668316 22225300 34 Pouliopoulos A. N. , Kwon N. , Jensen G. , Meaney A. , Niimi Y. , Burgess M. T. , Ji R. , , “Safety evaluation of a clinical focused ultrasound system for neuronavigation guided blood-brain barrier opening in non-human primates,”, doi:10.1038/s41598-021-94188-3. doi:10.1038/s41598-021-94188-3 35 Pouliopoulos A. N. , Wu S.-Y. , Burgess M. T. , Karakatsani M. E. , Kamimura H. A. S. , and Konofagou E. E. , “A Clinical System for Non-invasive Blood–Brain Barrier Opening Using a Neuronavigation-Guided Single-Element Focused Ultrasound Transducer,” Ultrasound Med Biol 46 , 73–89 (2020). doi:10.1016/j.ultrasmedbio.2019.09.010 31668690 36 Singh A. , Kusunose J. , Phipps M. A. , Wang F. , Chen L. M. , and Caskey C. F. , “Guiding and monitoring focused ultrasound mediated blood–brain barrier opening in rats using power Doppler imaging and passive acoustic mapping,” Sci Rep 12 , 14758 (2022). doi:10.1038/s41598-022-18328-z 36042266 37 Treeby B. E. , and Cox B. T. , “Modeling power law absorption and dispersion for acoustic propagation using the fractional Laplacian,” J Acoust Soc Am 127 , 2741–2748 (2010). doi:10.1121/1.3377056 21117722 38 Treeby B. E. , Jaros J. , Rendell A. P. , and Cox B. T. , “Modeling nonlinear ultrasound propagation in heterogeneous media with power law absorption using a k-space pseudospectral method,” J Acoust Soc Am 131 , 4324–4336 (2012). doi:10.1121/1.4712021 22712907 39 Tung Y.-S. , Choi J. J. , Baseri B. , and Konofagou E. E. , “Identifying the inertial cavitation threshold and skull effects in a vessel phantom using focused ultrasound and microbubbles,” Ultrasound Med Biol 36 , 840–852 (2010). doi:10.1016/j.ultrasmedbio.2010.02.009 20420973 40 Tung Y.-S. , Vlachos F. , Choi J. J. , Deffieux T. , Selert K. , and Konofagou E. E. , “In vivo transcranial cavitation threshold detection during ultrasound-induced blood-brain barrier opening in mice.,” Phys Med Biol 55 , 6141–6155 (2010). doi:10.1088/0031-9155/55/20/007 20876972 41 Schneider Uwe , Pedroni Eros , and Lomax Antony , “The calibration of CT Hounsfield units for radiotherapy treatment planning,” Phys Med Biol 41 , 111 (1996). doi:10.1088/0031-9155/41/1/009 8685250 42 Voogt M. J. , Trillaud H. , Kim Y. S. , Th W. P. . Mali M. , Barkhausen J. , Bartels L. W. , Deckers R. , , “Volumetric feedback ablation of uterine fibroids using magnetic resonance-guided high intensity focused ultrasound therapy,” Eur Radiol 22 , 411–417 (2012). doi:10.1007/s00330-011-2262-8 21901565 43 Woods A. J. , Antal A. , Bikson M. , Boggio P. S. , Brunoni A. R. , Celnik P. , Cohen L. G. , , “A technical guide to tDCS, and related non-invasive brain stimulation tools,” Clinical Neurophysiology 127 , 1031–1048 (2016). doi:10.1016/j.clinph.2015.11.012 26652115 44 Wu S. Y. , Aurup C. , Sanchez C. S. , Grondin J. , Zheng W. , Kamimura H. , Ferrera V. P. , , “Efficient blood-brain barrier opening in primates with neuronavigation-guided ultrasound and real-time acoustic mapping,”, doi:10.1038/s41598–018-25904–9. doi:10.1038/s41598-018-25904-9 45 Wu S. Y. , Sanchez C. S. , Samiotaki G. , Buch A. , Ferrera V. P. , and Konofagou E. E. , “Characterizing Focused-Ultrasound Mediated Drug Delivery to the Heterogeneous Primate Brain in Vivo with Acoustic Monitoring,”, doi:10.1038/srep37094. doi:10.1038/srep37094 46 Zhou H. , Liu Y. , Long X. , Qiao Y. , Lee J. , Liu X. , Zheng H. , , “MR-guided blood-brain barrier opening induced by rapid short-pulse ultrasound in non-human primates,” Retrieved from https://qims.amegroups.com/article/view/64169. Retrieved from https://qims.amegroups.com/article/view/64169

PMC010xxxxxx/PMC10312926.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37397981 10.21203/rs.3.rs-2588168/v1 10.21203/rs.3.rs-2588168 preprint 1 Article Impact of Body Mass Index on Pathological Response after Neoadjuvant Chemotherapy: Results from the I-SPY 2 trial Wang Haiyun Cancer Care Associates Yee Douglas University of Minnesota Department of Medicine: University of Minnesota Twin Cities Department of Medicine Potter David University of Minnesota Department of Medicine: University of Minnesota Twin Cities Department of Medicine Jewett Patricia University of Minnesota Department of Medicine: University of Minnesota Twin Cities Department of Medicine Yau Christina University of California San Francisco Beckwith Heather University of Minnesota Department of Medicine: University of Minnesota Twin Cities Department of Medicine Watson Allison Sanford Health O’Grady Nicholas University of California San Francisco Wilson Amy Quantumleap Brain Susie University of California San Francisco Pohlmann Paula MD Anderson Nellie B Connally Breast Center: The University of Texas MD Anderson Cancer Center Nellie B Connally Breast Center http://orcid.org/0000-0002-5433-4810 Blaes Anne University of Minnesota Medical Center Contributions: This study was designed and conducted by all authors on the manuscript. The concept was developed by H.W. with additional input from A.B. and D.Y. Statistical analyses were performed by P.J. The primary manuscript was drafted by H.W. and A.B., and was reviewed in its entirety by all of the authors. ✉ blaes004@umn.edu 31 5 2023 rs.3.rs-2588168https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs2588168v1.pdf Purpose Increased body mass index (BMI) has been associated with poor outcomes in women with breast cancer. We evaluated the association between BMI and pathological complete response (pCR) in the I-SPY 2 trial. Methods 978 patientsenrolled in the I-SPY 2 trial 3/2010–11/2016 and had a recorded baseline BMI prior to treatment were included in the analysis. Tumor subtypes were defined by hormone receptor and HER2 status. Pretreatment BMI was categorized as obese (BMI≥30 kg/m2), overweight (25≤BMI < 30 kg/m2), and normal/underweight (< 25 kg/m2). pCR was defined as elimination of detectable invasive cancer in the breast and lymph nodes (ypT0/Tis and ypN0) at the time of surgery. Logistic regression analysis was used to determine associations between BMI and pCR. Event-free survival (EFS) and overall survival (OS) between different BMI categories were examined using Cox proportional hazards regression. Results The median age in the study population was 49 years. pCR rates were 32.8% in normal/underweight, 31.4% in overweight, and 32.5% in obese patients. In univariable analysis, there was no significant difference in pCR with BMI. In multivariable analysis adjusted for race/ethnicity, age, menopausal status, breast cancer subtype, and clinical stage, there was no significant difference in pCR after neoadjuvant chemotherapy for obese compared with normal/underweight patients (OR = 1.1, 95% CI: 0.68–1.63, p = 0.83), and for overweight compared with normal/underweight (OR = 1, 95% CI: 0.64–1.47, p = 0.88). We tested for potential interaction between BMI and breast cancer subtype; however, the interaction was not significant in the multivariable model (p = 0.09). Multivariate Cox regression showed there was no difference in EFS (p = 0.81) or OS (p = 0.52) between obese, overweight, and normal/underweight breast cancer patients with a median follow-up time of 3.8 years. Conclusions We found no difference in pCR rates by BMI with actual body weight based neoadjuvant chemotherapy in this biologically high-risk breast cancer population in the I-SPY2 trial. Obesity Body Mass Index Breast Cancer Neoadjuvant Chemotherapy Pathological Complete Response NIH/NCIPO1 CA210961-01A1 ==== Body pmcIntroduction Observational studies have shown increased body mass index (BMI) is a risk factor for developing breast cancer, especially hormone receptor positive breast cancers [1, 2]. Obesity and being overweight are also associated with advanced stage of breast cancer at diagnosis and have been independently associated with poor breast cancer outcomes [3–5]. Pathological complete response (pCR) is a surrogate of long-term outcomes of locally advanced breast cancer such as event-free survival (EFS) and overall survival (OS) [6, 7]. Studies investigating the relationship between BMI and pCR after neoadjuvant chemotherapy in breast cancer have demonstrated mixed results, with some revealing an association of increased BMI with poorer pCR rates after neoadjuvant chemotherapy [8–11], while others did not reveal any significant association [12–14]. Most were retrospective studies, some using data from more than a decade ago [8]. Chemotherapy regimens varied substantially from study to study, as did chemotherapy dosage. Since oncology clinical practices may cap chemotherapy dosage to a maximum body-surface area (BSA) of 2.0m2 to avoid increased toxicity [11, 15], it is not clear if the observed worse pCR rate in obese breast cancer patients is related to chemotherapy underdosing rather than BMI itself [8, 11]; and it is also unclear whether the observed worse pCR in obese patients has any correlation with breast cancer biological subtypes. The I-SPY 2 (Investigation of Serial studies to Predict Your Therapeutic Response with Imaging and Molecular AnaLysis 2, NCT01042379) trial is an ongoing, multicenter, adaptive, phase II clinical trial platform that includes multiple experimental arms to evaluate new agents combined with standard neoadjuvant chemotherapy for the treatment of breast cancers with a high risk of recurrence, in comparison to standard chemotherapy regimen in a common control arm [16]. The trial uses pCR as primary end point. Importantly, chemotherapy dosage is not capped, but is given based on actual body weight [17]. The I-SPY 2 trial platform provides the advantage of eliminating some of the above-mentioned confounding factors while studying the association of BMI and neoadjuvant chemotherapy outcomes of breast cancer. The purpose of this study was to examine the association of BMI with pCR, EFS, and OS in women with high-risk early stage breast cancer enrolled in the I-SPY 2 trial. Methods Study population and data collection Women aged 18 years or older with a diagnosis of clinical stage II or III breast cancer, with a tumor diameter of at least 2.5cm by clinical examination and at least 2cm as assessed by imaging were eligible to participate in the I-SPY 2 trial. Exclusion criteria were an Eastern Cooperative Oncology Group performance status score greater than 1, and prior chemotherapy for this cancer. Patients with hormone receptor positive tumors and low risk MammaPrint® scores were also excluded given the lack of benefit from systemic chemotherapy [18]. In this trial, participants were randomized to different neoadjuvant treatment regimens based on biomarker status, determined by Bayesian probabilities of pCR within a specific biomarker subtype with the experimental regimen. The biomarker status was based on hormone receptors (HR), human epidermal growth factor receptor 2 (HER2) and a 70-gene assay of MammaPrint® at baseline. All participants received weekly intravenous paclitaxel (12 doses of 80mg per square meter of BSA) alone (control arm), or in combination with the assigned experimental regimen (experiment arm), followed by four doses of intravenous doxorubicin (60mg per square meter of BSA) and cyclophosphamide (600mg per square meter of BSA) every two to three weeks, with myeloid growth factor support if needed. Patients with HER2 + cancer also received trastuzumab for the first 12 weeks, given with a loading dose of 4mg per kilogram of body weight (week 1), followed by a maintenance dose of 2mg per kilogram every 3 weeks (weeks 4, 7, and 10). After receiving accelerated approval from the FDA [19], Pertuzumab was added to standard therapy for HER2 + patients, given with a loading dose of 840mg (week 1), followed by a maintenance dose of 420mg every 3 weeks (weeks 4, 7, and 10). All chemotherapy drugs were dosed based on actual body weight. Patients then underwent surgery which included axillary lymph node sampling. Radiation and adjuvant endocrine therapy after surgery were recommended in accordance with standard guidelines. All participants provided written informed consent before undergoing screening for the study, and a second consent was obtained before treatment was initiated if the individual was eligible after random assignment to open-label treatment arms. All participating sites of this trial received approval from an institutional review board. Measures The primary outcomes in this analysis were pCR [20], defined as elimination of detectable invasive cancer in the breast and lymph nodes (ypT0/Tis and ypN0) at the time of surgery; RCB (residual tumor burden) if pCR was not achieved [21]; and EFS and OS. The primary exposure of interest for this analysis was pretreatment BMI, categorized as obese (BMI≥30 kg/m2), overweight (25≤BMI < 30 kg/m2), and normal/underweight (< 25 kg/m2) based on World Health Organization criteria. Demographic and clinical covariates included in multivariate analysis and defined a priori were age at screening (years); race/ethnicity (Non-Hispanic White vs. Non-Hispanic Black/African American vs. Latinx vs. other); breast cancer subtype including HR+/HER2+, HR+/HER2−, HR−/HER2 + and triple negative (HR−/HER2−); menopausal status (pre- vs. peri- vs. post-menopausal); and advanced vs. early tumor stage (stage III vs. I or II). Statistical analysis Chi-squared and Anova were used to evaluate the association between BMI category and patient characteristics as appropriate. Logistic regression analysis was used to estimate associations between BMI and pCR, and linear regression to estimate the association between BMI and RCB; and Cox proportional hazards regression to estimate the associations between BMI and EFS, and between BMI and OS. These models were adjusted for the covariates listed above; because of limited degrees of freedom due to the total number of events, we excluded race/ethnicity from the covariates in the survival analyses. We report odds ratios (OR), linear coefficients, hazard ratios (HR), and respective 95% confidence intervals (CI). OR > 1 indicate greater odds of having pCR; hazard ratios > 1 indicate greater hazard of dying or having a major event. Analyses were run in SAS 9.4. All statistical tests were two-sided, and P values less than 0.05 were considered statistically significant. Results Patient characteristics In total, 977 patients with a recorded baseline BMI were included in this study. Of these, 35.6% (N = 348) were normal/underweight, 31.6% (N = 309) overweight, and 32.8% (N = 320) obese (Table 1). The mean age was 48.7 ± 10.6 years. Overweight (mean age 49.9 years) and obese (mean age 49.7 years) patients were significantly older than normal/underweight patients (mean age 46.8; p < 0.0001). There were more Non-Hispanic Black / African American and Hispanic participants among those who were obese compared to normal/underweight and overweight. BMI category was not significantly associated with menopausal status, cancer stage, or cancer hormonal subtype. Relationship between BMI and Pathological Response The overall pCR rate after neoadjuvant chemotherapy was 32.2%. pCR rates were 32.8% in normal/underweight, 31.4% in overweight, and 32.5% in obese patients, with no significant difference in the unadjusted or adjusted analysis (obese vs. normal/underweight, unadjusted OR = 0.99, 95% CI 0.71–1.37, adjusted OR = 1.05, 95% CI 0.68–1.63; overweight vs. normal/underweight, unadjusted OR = 0.94, 95% CI 0.68–1.30, adjusted OR = 0.97, 95% CI 0.64–1.47, Table 2). We ran an additional sensitivity analysis with continuous BMI as predictor, and this association was not significant, either. Although an interaction between BMI and hormonal breast cancer subtype was not significant, we ran the unadjusted logistic regression models (predictor: continuous BMI) stratified by cancer hormonal subtype because cancer outcomes typically differ by hormonal subtypes. The association of BMI with PCR status was not significant in any of these models (Table 2). We did notice a trend towards decreased pCR rates with increasing BMI in the HR−/HER2 + subgroup (N = 88, Table 3) which did not reach statistical significance. The pCR rate in this hormonal subtype group was 75% in normal/underweight, 64.5% in overweight, and 48.3% in obese patients (overall p = 0.11). Using linear regression to compare Residual Cancer Burden (RCB) by BMI, RCB index was not associated with BMI category in either the unadjusted or the adjusted model, or for any cancer hormonal subtype after stratification (Table 4). Relationship of BMI with EFS and OS With a median follow-up time of 3.8 years, estimated OS at 5 years was 85.3% (95% CI 82.3–87.8%; 111 deaths out of 895 participants with known survival status) and estimated EFS at 5 years was 76.5% (95% CI 73.1–79.5%; 182 events out of 895 participants with known event status). BMI was not associated with EFS or OS in this study population (Tables 5 and 6). Due to limited events, we were unable to stratify these models by cancer hormonal subtype. Kaplan Meier curves for EFS and OS in different BMI categories are shown in Figs. 1 and 2. Discussion In this clinical trial using actual body weight-based chemotherapy, higher baseline BMI was not associated with decreasing pCR rate after neoadjuvant chemotherapy in biologically high-risk early stage breast cancer patients, nor was it associated with worse EFS or OS. The overall pCR rate was 32.2% in our study, which was modest in comparison of other studies [22, 11]. The I-SPY 2 trial used standard chemotherapy regimen +/− HER2 targeted therapy depending on the HER2 status. It should also be noted, however, that this clinical trial also included patients with HR + breast cancer which have historically demonstrated lower response rates to chemotherapy [16]. This may explain why the overall pCR rate was modest after including the HR+/HER2− population, as HER2 + patients had a considerably higher pCR rate of 68% in our study [20]. Although several prospective studies and meta-analyses have reported that increased body weight was associated with poorer breast cancer outcomes such as OS and EFS, especially in postmenopausal women [23, 15, 24], it has been a challenge to clarify the underlying cause. In part, this has been attributed to possible interactions between BMI and comorbidities such as diabetes, coronary artery disease, cerebral artery disease, and socioeconomic status [25–28]. Neoadjuvant chemotherapy has recently become the standard of care for biologically high-risk breast cancers. Achieving pCR at the time of surgery is a surrogate marker for better long-term breast cancer outcomes [6, 29]. The Collaborative Trials in Neoadjuvant Breast Cancer (CTNeoBC) results indicated a long-term benefit for patients achieving pCR, as pCR was positively associated with overall EFS (hazard ratio 0.48, 95% CI 0.43–0.54) and overall OS (hazard ratio 0.36, 95% CI 0.31–0.42) [7]. Monitoring pCR rates among overweight and obese breast cancer patients who received neoadjuvant chemotherapy may help us understand why higher BMI is associated with poorer breast cancer outcomes. Litton et al did the first large retrospective study in this regard, finding that patients with higher BMI were more likely to present with high-risk tumor characteristics and were less likely to achieve pCR after neoadjuvant chemotherapy; and that higher BMI was associated with worse OS [8]. Elsamany and colleagues performed a similar retrospective analysis in Saudi Arabian and Egyptian populations, and Fontanella et al did a pooled analysis of four clinical trials in Germany, both studies showed high BMI was associated with worse pCR rate [10, 11]. However, similar studies by Erbes et al and Kogawa et al did not reveal any statistically significant association between increased BMI and worse pCR [12, 14]. We previously performed a meta-analysis with total of 18,702 patients, with pooled univariable analysis demonstrating increased BMI was associated with worse pCR rate in overweight and obese patients [30]. Yet this meta-analysis has limitations given most included studies were retrospective in nature, multivariable analysis and subgroup analysis based on different subtypes of breast cancer were not able to be performed due to lack of standardization of patient characteristics; there were significant variations of chemotherapy regimens, and inclusion of non-weight based chemotherapy dosing [30]. Using the I-SPY 2 trial data to investigate the association of increased BMI with pCR outcome has several advantages. First, this is a currently active clinical trial platform using standard concurrent treatment regimens for each subtype of breast cancer, with a focus on treating high risk, biologically active breast cancer. Second, the I-SPY 2 trial uses standard treatment protocols and chemotherapy is given based on actual body weight. Lastly, it is one of the largest multicenter randomized clinical trials focusing on neoadjuvant therapy for breast cancer. These advantages may eliminate the potential biases originating from the variation of chemotherapy regimens and the underdosing of chemotherapy agents in patients with elevated BMI. In this strictly designed clinical trial, we did not identify any statistically significant evidence that higher BMI was associated with decreasing pCR rate in the high-risk early stage breast cancer group with various hormonal subtypes; nor within each hormonal subtype group after stratification. This result was different from most of the retrospective studies discussed above. Our study reinforced the potential importance of dosing chemotherapy based on actual body weight. Some clinicians may reduce chemotherapy dosage in overweight and obese patients because of the fear of overdosing and excessive toxicity with higher chemotherapy dosage, although randomized clinical trials have demonstrated that this practice contributes to worse outcomes and guidelines recommend against this practice [31–33]. In the I-SPY 2 trial, chemotherapy dosing is strictly based on actual body weight, even if patients’ BSA is above 2.0m2. In Litton’s study, the chemotherapy dose of each patient was not documented and not able to be verified [8]. In Fontanella’s study, more than half of the study population had chemotherapy dosage capped at 2.0m2 [11]. It is possible that the poorer breast cancer outcomes in overweight and obese patients from these studies was attributable to chemotherapy underdosing rather than the influence of BMI on the chemotherapy effectiveness in these patients. Our study has several limitations. Although the I-SPY 2 trial is a prospective study, the correlation of BMI to pCR is not the predetermined primary end point of this trial. While our analysis included almost 1000 women, dividing the study population by tumor subtype, ethnicity and BMI limited our statistical power; especially in the subgroup analysis of BMI in different breast cancer subtypes and its impact on breast cancer outcomes. As there were too few deaths/recurrences in patients who achieved pCR (RCB = 0), we were not able to run a meaningful survival analysis to determine whether BMI has an impact on OS and EFS regardless of patients’ pCR status. Longer follow up is needed to understand the overall impact on OS and EFS. Conclusion We observed no difference in pCR rates by baseline BMI in this biologically high-risk breast cancer population receiving actual body weight-based neoadjuvant chemotherapy. These findings suggest the importance of treating overweight and obese patients with chemotherapy dosage based on actual weight. Longer follow up and further work, however, is needed to understand the role of body mass and breast cancer outcomes across all breast cancer subtypes. Acknowledgements: We would like to thank the ISPY2 clinical trial team for their enrollment of subjects to this study, and the ongoing commitment to care of breast cancer patients. Funding: This study was conducted by volunteer members of ISPY2 clinical trial research team. The clinical trial itself is funded by NIH/NCI PO1 CA210961-01A1 for DY receives research support. Data: Data is available upon request. Abbreviations BMI Increased body mass index pCR Pathological complete response EFS Event-free survival OS Overall survival BSA Body-surface area I-SPY 2 Investigation of Serial studies to Predict Your Therapeutic Response with Imaging and Molecular AnaLysis 2 HR Hormone receptors HER2 Human epidermal growth factor receptor 2 RCB Residual cancer burden OR Odds ratios CI Confidence intervals Figure 1 Kaplan Meier curve for event-free survival based on BMI category Figure 2 Kaplan Meier curve for overall survival based on BMI category Table 1 Patient characteristics according to BMI category. Patient Characteristics BMI < 25 (N = 348) 25 ≤BMI < 30 (N=309) BMI ≥30 (N=320) Total (N=977) P # of pts (%) # of pts (%) # of pts (%) # of pts (%) Age at Initial Treatment Mean (STD) 46.8 (10.6) 49.9 (10.7) 49.7 (10.1) 48.7 (10.6) <.0001 Race/Ethnicity <.0001 African American / Black 21 (6) 26 (8.4) 71 (22.2) 118 (12.1) American Indian / Native or Hawaiian / Pacific Islander 4 (1.2) 2 (0.7) 5 (1.6) 11 (1.1) Asian 35 (10.1) 27 (8.7) 9 (2.8) 71 (7.3) Hispanic 23 (6.6) 43 (13.9) 49 (15.3) 115 (11.8) Non-Hispanic White 265 (76.2) 211 (68.3) 186 (58.1) 662 (67.8) Menopausal Status 0.14 Pre- 193 (64.3) 147 (57.0) 135 (54.4) 475 (58.9) Peri- 12 (4.0) 9 (3.5) 13 (5.2) 34 (4.2) Post- 95 (31.7) 102 (39.5) 100 (40.3) 297 (36.9) Missing 48 51 72 171 Hormonal and HER2 Status 0.53 HR+/HER2+ 63 (18.1) 44 (14.2) 48 (15.0) 155 (15.9) HR+/HER2- 130 (37.4) 130 (42.1) 118 (36.9) 378 (38.7) HR-/HER2+ 28 (8.1) 31 (10.0) 29 (9.1) 88 (9.0) Age at Initial Treatment HR-/HER2- 127 (36.5) 104 (33.7) 125 (39.1) 356 (36.4) Cancer Stage 0.14 I 8 (2.7) 9 (3.5) 4 (1.6) 21 (2.6) II 222 (75.5) 173 (66.8) 183 (70.9) 578 (71.3) III 64 (21.8) 77 (29.7) 71 (27.5) 212 (26.1) Missing 54 50 62 166 Table 2: Odds Ratios (OR) of pathological complete response (pCR) by BMI categories and adjusted variables # of pts pCR OR (95% CI) P OR (95% CI) P (N) N (%) Unadjusted Adjusted* BMI Normal / Underweight 348 114 (32.8) 1 (Ref.) 1 (Ref.) Overweight 309 97 (31.4) 0.94 (0.68–1.30) 0.71 0.97 (0.64– 1.47) 0.88 Obese 320 104 (32.5) 0.99 (0.71–1.37) 0.94 1.05 (0.68– 1.63) 0.83 BMI (continuous, per additonal unit) 0.99 (0.97–1.01) 0.50 0.99 (0.96– 1.02) 0.64 Model stratified by cancer status (unadjusted odds ratios for continuous BMI, per additional unit) Subtype stratum HR+/HER2− 378 64 (16.9) 1.01 (0.96–1.05) 0.81 HR+/HER2+ 155 57 (36.8) 0.97 (0.91–1.03) 0.32 HR−/HER2+ 88 55 (62.5) 0.93 (0.86–1.00) 0.06 HR−/HER2− 356 139 (39.0) 1.00 (0.97–1.04) 0.99 * Adjusted for age at screening, hormonal cancer subtype, race, stage, and menopausal status Table 3 pCR rate of different BMI categories by breast cancer subtypes. Breast Cancer Subtype pCR Normal/Underweight N (%) Overweight N (%) Obese N (%) P HR+/HER2+ No 39 (61.9) 27 (61.4) 32 (66.7) 0.83 Yes 24 (38.1) 17 (38.6) 16 (33.3) HR+/HER2- No 105 (80.8) 116 (89.2) 93 (78.8) 0.06 Yes 25 (19.2) 14 (10.8) 25 (21.2) HR-/HER2+ No 7 (25.0) 11 (35.5) 15 (51.7) 0.11 Yes 21 (75.0) 20 (64.5) 14 (48.3) HR-/HER2- No 83 (65.4) 58 (55.8) 76 (60.8) 0.33 Yes 44 (34.7) 46 (44.2) 49 (39.2) (HR: hormone receptor, HER2: human epidermal growth factor receptor 2, pCR: pathological complete response) Table 4 Association of BMI categories and continuous RCB index in patients who did not achieve pCR. Coefficient (95% CI) P Coefficient (95% CI) P Unadjusted Adjusted* BMI 0.95 0.56 Normal / Underweight 0 (Ref.) 0 (Ref.) Overweight 0.02 (−0.20–0.24) 0.86 −0.14 (−0.38–0.11) 0.28 Obese 0.03 (−0.18–0.25) 0.75 −0.06 (−0.32–0.20) 0.67 BMI (continuous, per additonal unit) 0.00 (−0.01–0.02) 0.54 0.00 (−0.02–0.02) 0.84 Model stratified by cancer status (coefficients for continuous BMI, per additional unit) Subtype stratum HR+/HER2− 0.00 (−0.02–0.02) 0.87 HR+/HER2+ 0.01 (−0.03–0.05) 0.61 HR−/HER2+ 0.03 (−0.01–0.07) 0.13 HR−/HER2− 0.00 (−0.02–0.03) 0.80 * Adjusted for age at screening, hormonal cancer subtype, race, stage, and menopausal status Table 5 Association of BMI with overall survival (hazard ratios, HR) BMI Adjusted HR* 95% CI P Normal / Underweight 1 (Ref.) Overweight 1.13 0.69–1.87 0.63 Obese 0.82 0.46–1.44 0.48 * Adjusted for age, hormonal cancer subtype, stage, and menopausal status Table 6: Association of BMI with Event-free Survival BMI Adjusted HR* 95% CI P Normal / Underweight 1 (Ref.) Overweight 0.99 0.67–1.49 0.98 Obese 0.88 0.57–1.35 0.56 * Adjusted for age, hormonal cancer subtype, stage, and menopausal status Declarations Disclosures: DP, CY, and DY receive research funding from Quantum Leap Collaborative. The remaining authors have no disclosures. Ethics consideration: All participants provided written informed consent before undergoing screening and a second consent was obtained before treatment was initiated if the individual was eligible after random assignment to open-label treatment arms. All participating sites of this trial received approval from an institutional review board. ==== Refs References 1. Pischon T , Nimptsch K (2016) Obesity and Risk of Cancer: An Introductory Overview. Recent Results Cancer Res 208 :1–15. 10.1007/978-3-319-42542-9_1 27909899 2. Munsell MF , Sprague BL , Berry DA , Chisholm G , Trentham-Dietz A (2014) Body mass index and breast cancer risk according to postmenopausal estrogen-progestin use and hormone receptor status. Epidemiol Rev 36 :114–136. 10.1093/epirev/mxt010 24375928 3. Dawood S , Broglio K , Gonzalez-Angulo AM , Kau SW , Islam R , Hortobagyi GN , Cristofanilli M (2008) Prognostic value of body mass index in locally advanced breast cancer. Clin Cancer Res 14 (6 ):1718–1725. 10.1158/1078-0432.ccr-07-1479 18347172 4. Kroenke CH , Chen WY , Rosner B , Holmes MD (2005) Weight, weight gain, and survival after breast cancer diagnosis. J Clin Oncol 23 (7 ):1370–1378. 10.1200/jco.2005.01.079 15684320 5. Protani M , Coory M , Martin JH (2010) Effect of obesity on survival of women with breast cancer: systematic review and meta-analysis. Breast Cancer Res Treat 123 (3 ):627–635. 10.1007/s10549-010-0990-0 20571870 6. Rastogi P , Anderson SJ , Bear HD , Geyer CE , Kahlenberg MS , Robidoux A , Margolese RG , Hoehn JL , Vogel VG , Dakhil SR , Tamkus D , King KM , Pajon ER , Wright MJ , Robert J , Paik S , Mamounas EP , Wolmark N (2008) Preoperative chemotherapy: updates of National Surgical Adjuvant Breast and Bowel Project Protocols B-18 and B-27. J Clin Oncol 26 (5 ):778–785. 10.1200/jco.2007.15.0235 18258986 7. Cortazar P , Zhang L , Untch M , Mehta K , Costantino JP , Wolmark N , Bonnefoi H , Cameron D , Gianni L , Valagussa P , Swain SM , Prowell T , Loibl S , Wickerham DL , Bogaerts J , Baselga J , Perou C , Blumenthal G , Blohmer J , Mamounas EP , Bergh J , Semiglazov V , Justice R , Eidtmann H , Paik S , Piccart M , Sridhara R , Fasching PA , Slaets L , Tang S , Gerber B , Geyer CE , Pazdur R , Ditsch N , Rastogi P , Eiermann W , von Minckwitz G (2014) Pathological complete response and long-term clinical benefit in breast cancer: the CTNeoBC pooled analysis. Lancet 384 (9938 ):164–172. 10.1016/S0140-6736(13)62422-8 24529560 8. Litton JK , Gonzalez-Angulo AM , Warneke CL , Buzdar AU , Kau SW , Bondy M , Mahabir S , Hortobagyi GN , Brewster AM (2008) Relationship between obesity and pathologic response to neoadjuvant chemotherapy among women with operable breast cancer. J Clin Oncol 26 (25 ):4072–4077. 10.1200/JCO.2007.14.4527 18757321 9. Chen S , Chen CM , Zhou Y , Zhou RJ , Yu KD , Shao ZM (2012) Obesity or overweight is associated with worse pathological response to neoadjuvant chemotherapy among Chinese women with breast cancer. PLoS ONE 7 (7 ):e41380. 10.1371/journal.pone.0041380 22848477 10. Elsamany S , Alzahrani A , Abozeed WN , Rasmy A , Farooq MU , Elbiomy MA , Rawah E , Alsaleh K , Abdel-Aziz NM (2015) Mammographic breast density: Predictive value for pathological response to neoadjuvant chemotherapy in breast cancer patients. Breast 24 (5 ):576–581. 10.1016/j.breast.2015.05.007 26071795 11. Fontanella C , Lederer B , Gade S , Vanoppen M , Blohmer JU , Costa SD , Denkert C , Eidtmann H , Gerber B , Hanusch C , Hilfrich J , Huober J , Schneeweiss A , Paepke S , Jackisch C , Mehta K , Nekljudova V , Untch M , Neven P , von Minckwitz G , Loibl S (2015) Impact of body mass index on neoadjuvant treatment outcome: a pooled analysis of eight prospective neoadjuvant breast cancer trials. Breast Cancer Res Treat 150 (1 ):127–139. 10.1007/s10549-015-3287-5 25677740 12. Erbes T , Stickeler E , Rücker G , Buroh S , Asberger J , Dany N , Thornton S , Iborra S , Hirschfeld M , Gitsch G , Mayer S (2016) BMI and Pathologic Complete Response to Neoadjuvant Chemotherapy in Breast Cancer: A Study and Meta-Analysis. Clin Breast Cancer 16 (4 ):e119–132. 10.1016/j.clbc.2016.02.018 27067040 13. Eralp Y , Smith TL , Altundağ K , Kau SW , Litton J , Valero V , Buzdar A , Hortobagyi GN , Arun B (2009) Clinical features associated with a favorable outcome following neoadjuvant chemotherapy in women with localized breast cancer aged 35 years or younger. J Cancer Res Clin Oncol 135 (1 ):141–148. 10.1007/s00432-008-0428-9 18581139 14. Kogawa T , Fujii T , Fouad TM , Liu DD , Harano K , Masuda H , Iwase T , Barnett C , Park YS , Lim B , Tripathy D , Litton JK , Ueno NT (2018) Impact of change in body mass index during neoadjuvant chemotherapy and survival among breast cancer subtypes. Breast Cancer Res Treat 171 (2 ):501–511. 10.1007/s10549-018-4853-4 29915946 15. Arce-Salinas C , Aguilar-Ponce JL , Villarreal-Garza C , Lara-Medina FU , Olvera-Caraza D , Alvarado Miranda A , Flores-Diaz D , Mohar A (2014) Overweight and obesity as poor prognostic factors in locally advanced breast cancer patients. Breast Cancer Res Treat 146 (1 ):183–188. 10.1007/s10549-014-2977-8 24842266 16. Barker AD , Sigman CC , Kelloff GJ , Hylton NM , Berry DA , Esserman LJ (2009) I-SPY 2: an adaptive breast cancer trial design in the setting of neoadjuvant chemotherapy. Clin Pharmacol Ther 86 (1 ):97–100. 10.1038/clpt.2009.68 19440188 17. Rugo HS , Olopade OI , DeMichele A , Yau C , van ‘t Veer LJ , Buxton MB , Hogarth M , Hylton NM , Paoloni M , Perlmutter J , Symmans WF , Yee D , Chien AJ , Wallace AM , Kaplan HG , Boughey JC , Haddad TC , Albain KS , Liu MC , Isaacs C , Khan QJ , Lang JE , Viscusi RK , Pusztai L , Moulder SL , Chui SY , Kemmer KA , Elias AD , Edmiston KK , Euhus DM , Haley BB , Nanda R , Northfelt DW , Tripathy D , Wood WC , Ewing C , Schwab R , Lyandres J , Davis SE , Hirst GL , Sanil A , Berry DA , Esserman LJ , Investigators I-S (2016) Adaptive Randomization of Veliparib-Carboplatin Treatment in Breast Cancer. N Engl J Med 375 (1 ):23–34. 10.1056/NEJMoa1513749 27406347 18. Cardoso F , van’t Veer LJ , Bogaerts J , Slaets L , Viale G , Delaloge S , Pierga JY , Brain E , Causeret S , DeLorenzi M , Glas AM , Golfinopoulos V , Goulioti T , Knox S , Matos E , Meulemans B , Neijenhuis PA , Nitz U , Passalacqua R , Ravdin P , Rubio IT , Saghatchian M , Smilde TJ , Sotiriou C , Stork L , Straehle C , Thomas G , Thompson AM , van der Hoeven JM , Vuylsteke P , Bernards R , Tryfonidis K , Rutgers E , Piccart M , Investigators M (2016) 70-Gene Signature as an Aid to Treatment Decisions in Early-Stage Breast Cancer. N Engl J Med 375 (8 ):717–729. 10.1056/NEJMoa1602253 27557300 19. von Minckwitz G , Procter M , de Azambuja E , Zardavas D , Benyunes M , Viale G , Suter T , Arahmani A , Rouchet N , Clark E , Knott A , Lang I , Levy C , Yardley DA , Bines J , Gelber RD , Piccart M , Baselga J (2017) Adjuvant Pertuzumab and Trastuzumab in Early HER2-Positive Breast Cancer. N Engl J Med 377 (2 ):122–131. 10.1056/NEJMoa1703643 28581356 20. Yee D , DeMichele AM , Yau C , Isaacs C , Symmans WF , Albain KS , Chen YY , Krings G , Wei S , Harada S , Datnow B , Fadare O , Klein M , Pambuccian S , Chen B , Adamson K , Sams S , Mhawech-Fauceglia P , Magliocco A , Feldman M , Rendi M , Sattar H , Zeck J , Ocal IT , Tawfik O , LeBeau LG , Sahoo S , Vinh T , Chien AJ , Forero-Torres A , Stringer-Reasor E , Wallace AM , Pusztai L , Boughey JC , Ellis ED , Elias AD , Lu J , Lang JE , Han HS , Clark AS , Nanda R , Northfelt DW , Khan QJ , Viscusi RK , Euhus DM , Edmiston KK , Chui SY , Kemmer K , Park JW , Liu MC , Olopade O , Leyland-Jones B , Tripathy D , Moulder SL , Rugo HS , Schwab R , Lo S , Helsten T , Beckwith H , Haugen P , Hylton NM , Van’t Veer LJ , Perlmutter J , Melisko ME , Wilson A , Peterson G , Asare AL , Buxton MB , Paoloni M , Clennell JL , Hirst GL , Singhrao R , Steeg K , Matthews JB , Asare SM , Sanil A , Berry SM , Esserman LJ , Berry DA (2020) Association of Event-Free and Distant Recurrence-Free Survival With Individual-Level Pathologic Complete Response in Neoadjuvant Treatment of Stages 2 and 3 Breast Cancer: Three-Year Follow-up Analysis for the I-SPY2 Adaptively Randomized Clinical Trial. JAMA Oncol 6 (9 ):1355–1362. 10.1001/jamaoncol.2020.2535 32701140 21. Symmans WF , Yau C , Chen YY , Balassanian R , Klein ME , Pusztai L , Nanda R , Parker BA , Datnow B , Krings G , Wei S , Feldman MD , Duan X , Chen B , Sattar H , Khazai L , Zeck JC , Sams S , Mhawech-Fauceglia P , Rendi M , Sahoo S , Ocal IT , Fan F , LeBeau LG , Vinh T , Troxell ML , Chien AJ , Wallace AM , Forero-Torres A , Ellis E , Albain KS , Murthy RK , Boughey JC , Liu MC , Haley BB , Elias AD , Clark AS , Kemmer K , Isaacs C , Lang JE , Han HS , Edmiston K , Viscusi RK , Northfelt DW , Khan QJ , Leyland-Jones B , Venters SJ , Shad S , Matthews JB , Asare SM , Buxton M , Asare AL , Rugo HS , Schwab RB , Helsten T , Hylton NM , van ‘t Veer L , Perlmutter J , DeMichele AM , Yee D , Berry DA , Esserman LJ (2021) Assessment of Residual Cancer Burden and Event-Free Survival in Neoadjuvant Treatment for High-risk Breast Cancer: An Analysis of Data From the I-SPY2 Randomized Clinical Trial. JAMA Oncol 7 (11 ):1654–1663. 10.1001/jamaoncol.2021.3690 34529000 22. Warner ET , Ballman KV , Strand C , Boughey JC , Buzdar AU , Carey LA , Sikov WM , Partridge AH (2016) Impact of race, ethnicity, and BMI on achievement of pathologic complete response following neoadjuvant chemotherapy for breast cancer: a pooled analysis of four prospective Alliance clinical trials (A151426). Breast Cancer Res Treat 159 (1 ):109–118. 10.1007/s10549-016-3918-5 27449492 23. Chan DS , Vieira AR , Aune D , Bandera EV , Greenwood DC , McTiernan A , Navarro Rosenblatt D , Thune I , Vieira R , Norat T (2014) Body mass index and survival in women with breast cancer-systematic literature review and meta-analysis of 82 follow-up studies. Ann Oncol 25 (10 ):1901–1914. 10.1093/annonc/mdu042 24769692 24. Borugian MJ , Sheps SB , Kim-Sing C , Olivotto IA , Van Patten C , Dunn BP , Coldman AJ , Potter JD , Gallagher RP , Hislop TG (2003) Waist-to-hip ratio and breast cancer mortality. Am J Epidemiol 158 (10 ):963–968. 10.1093/aje/kwg236 14607804 25. Kang C , LeRoith D , Gallagher EJ (2018) Diabetes, Obesity, and Breast Cancer. Endocrinology 159 (11 ):3801–3812. 10.1210/en.2018-00574 30215698 26. Moghani Lankarani M , Assari S (2017) Diabetes, hypertension, obesity, and long-term risk of renal disease mortality: Racial and socioeconomic differences. J Diabetes Investig 8 (4 ):590–599. 10.1111/jdi.12618 27. Moustsen IR , Friberg AS , Larsen SB , Duun-Henriksen AK , Tjønneland A , Kjaer SK , Brasso K , Johansen C , Dalton SO (2019) The association between education and risk of major cardiovascular events among prostate cancer patients: a study from the Diet, Cancer and Health study. Acta Oncol 58 (5 ):715–721. 10.1080/0284186x.2019.1572924 30757932 28. den Hartigh LJ (2019) Conjugated Linoleic Acid Effects on Cancer, Obesity, and Atherosclerosis: A Review of Pre-Clinical and Human Trials with Current Perspectives. Nutrients 11 (2 ). 10.3390/nu11020370 29. Wolmark N , Wang J , Mamounas E , Bryant J , Fisher B (2001) Preoperative chemotherapy in patients with operable breast cancer: nine-year results from National Surgical Adjuvant Breast and Bowel Project B-18. J Natl Cancer Inst Monogr 3096–102. 10.1093/oxfordjournals.jncimonographs.a003469 30. Wang H , Zhang S , Yee D , Basu S , Beckwith H , Potter D , Blaes A (2021) Impact of body mass index on pathological complete response following neoadjuvant chemotherapy in operable breast cancer: a meta-analysis. Breast Cancer. 10.1007/s12282-020-01194-w 31. Rosner GL , Hargis JB , Hollis DR , Budman DR , Weiss RB , Henderson IC , Schilsky RL (1996) Relationship between toxicity and obesity in women receiving adjuvant chemotherapy for breast cancer: results from cancer and leukemia group B study 8541. J Clin Oncol 14 (11 ):3000–3008. 10.1200/jco.1996.14.11.3000 8918498 32. Griggs JJ , Sorbero ME , Lyman GH (2005) Undertreatment of obese women receiving breast cancer chemotherapy. Arch Intern Med 165 (11 ):1267–1273. 10.1001/archinte.165.11.1267 15956006 33. Griggs JJ , Mangu PB , Anderson H , Balaban EP , Dignam JJ , Hryniuk WM , Morrison VA , Pini TM , Runowicz CD , Rosner GL , Shayne M , Sparreboom A , Sucheston LE , Lyman GH (2012) Appropriate chemotherapy dosing for obese adult patients with cancer: American Society of Clinical Oncology clinical practice guideline. J Clin Oncol 30 (13 ):1553–1561. 10.1200/jco.2011.39.9436 22473167 34. Oldenhove G , Boucquey E , Taquin A , Acolty V , Bonetti L , Ryffel B , Le Bert M , Englebert K , Boon L , Moser M (2018) PD-1 Is Involved in the Dysregulation of Type 2 Innate Lymphoid Cells in a Murine Model of Obesity. Cell Rep 25 (8 ):2053–2060e2054. 10.1016/j.celrep.2018.10.091 30463004 35. Wang Z , Aguilar EG , Luna JI , Dunai C , Khuat LT , Le CT , Mirsoian A , Minnar CM , Stoffel KM , Sturgill IR , Grossenbacher SK , Withers SS , Rebhun RB , Hartigan-O’Connor DJ , Mendez-Lagares G , Tarantal AF , Isseroff RR , Griffith TS , Schalper KA , Merleev A , Saha A , Maverakis E , Kelly K , Aljumaily R , Ibrahimi S , Mukherjee S , Machiorlatti M , Vesely SK , Longo DL , Blazar BR , Canter RJ , Murphy WJ , Monjazeb AM (2019) Paradoxical effects of obesity on T cell function during tumor progression and PD-1 checkpoint blockade. Nat Med 25 (1 ):141–151. 10.1038/s41591-018-0221-5 30420753 36. Naik A , Monjazeb AM , Decock J (2019) The Obesity Paradox in Cancer, Tumor Immunology, and Immunotherapy: Potential Therapeutic Implications in Triple Negative Breast Cancer. Front Immunol 10 :1940. 10.3389/fimmu.2019.01940 31475003 37. McQuade JL , Daniel CR , Hess KR , Mak C , Wang DY , Rai RR , Park JJ , Haydu LE , Spencer C , Wongchenko M , Lane S , Lee DY , Kaper M , McKean M , Beckermann KE , Rubinstein SM , Rooney I , Musib L , Budha N , Hsu J , Nowicki TS , Avila A , Haas T , Puligandla M , Lee S , Fang S , Wargo JA , Gershenwald JE , Lee JE , Hwu P , Chapman PB , Sosman JA , Schadendorf D , Grob JJ , Flaherty KT , Walker D , Yan Y , McKenna E , Legos JJ , Carlino MS , Ribas A , Kirkwood JM , Long GV , Johnson DB , Menzies AM , Davies MA (2018) Association of body-mass index and outcomes in patients with metastatic melanoma treated with targeted therapy, immunotherapy, or chemotherapy: a retrospective, multicohort analysis. Lancet Oncol 19 (3 ):310–322. 10.1016/s1470-2045(18)30078–029449192

PMC010xxxxxx/PMC10312927.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398311 10.21203/rs.3.rs-2922367/v1 10.21203/rs.3.rs-2922367 preprint 1 Article Ubiquitin-specific peptidase 22 controls integrin-dependent cancer cell stemness and metastasis Liu Kun Northwestern University Gao Qiong Northwestern university Jia Yuzhi Northwestern university Wei Juncheng Northwestern university Chaudhuri Shuvam Northwestern university Wang Shengnan Northwestern university Tang Amy Northwestern University Mani Nikita Northwestern University Iyer Radhika Northwestern University Cheng Yang Northwestern University Gao Beixue Northwestern University Lu Weiyuan Northwestern University Sun Zhaolin Dalian Medical University http://orcid.org/0000-0003-4822-7995 Liu Huiping Northwestern University http://orcid.org/0000-0002-4211-2751 Fang Deyu Northwestern University Author contribution D.F. designed and supervised this study, and wrote the manuscript. K.L. performed almost the experiments and wrote the manuscript. Q.G. developed the concepts and assisted with K.L. in this study. H.L. and Y. J. provided breast cancer patient-derived xenograft models. J.W. and Y. J. helped the animal studies. A.T. and N.L.M. helped the flow cytometry experiments, R.L, B.G., W.L. and Z.S. helped the data analysis. H.L. and S.M.C. edited the manuscript. ✉ fangd@northwestern.edu 16 6 2023 rs.3.rs-2922367https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs2922367v1.pdf Integrins plays critical roles in connecting the extracellular matrix and actin skeleton for cell adhesion, migration, signal transduction, and gene transcription, which upregulation is involved in cancer stemness and metastasis. However, the molecular mechanisms underlying how integrins are upregulated in cancer stem cells (CSCs) remain as a biomedical mystery. Herein, we show that the death from cancer signature gene USP22 is essential to maintain the stemness of breast cancer cells through promoting the transcription of a group of integrin family members in particular integrin β1 (ITGB1). Both genetic and pharmacological USP22 inhibition largely impaired breast cancer stem cell self-renewal and prevented their metastasis. Integrin β1 reconstitution partially rescued USP22-null breast cancer stemness and their metastasis. At the molecular level, USP22 functions as a bona fide deubiquitinase to protect the proteasomal degradation of the forkhead box M1 (FoxM1), a transcription factor for tumoral ITGB1 gene transcription. Importantly unbiased analysis of the TCGA database revealed a strong positive correlation between the death from cancer signature gene ubiquitin-specific peptidase 22 (USP22) and ITGB1, both of which are critical for cancer stemness, in more than 90% of human cancer types, implying that USP22 functions as a key factor to maintain stemness for a broad spectrum of human cancer types possibly through regulating ITGB1. To support this notion, immunohistochemistry staining detected a positive correlation among USP22, FoxM1 and integrin β1 in human breast cancers. Collectively, our study identifies the USP22-FoxM1-integrin β1 signaling axis critical for cancer stemness and offers a potential target for antitumor therapy. Cancer stem cells USP22 Integrins metastasis ==== Body pmcIntroduction Despite recent therapeutic advances in tumor treatment, metastasis to the nearby or distal organs remains the main cause of cancer-related death1. It has been proposed that only a small portion of primary tumor cells, termed cancer stem cells (CSCs), are responsible for metastasis 2. CSCs are a small population in tumor that are self-renewable, preferentially aggressive and responsible for cancer initiation, metastasis, and recurrence 3. Breast cancer stem cells (BCSCs), for example, have antioxidative, tumor sphere formation, and chemoresistance properties. Based on cell surface marker expression, BCSCs are CD44(+)/CD24(−/low) tumorigenic cells that initiate tumors in xenografts4. CD44 is a cell surface glycoprotein and stemness marker in BCSCs. CD44 binds to hyaluronic acid and mediates the interactions between cell/cell and cell/matrix proteins such as matrix metalloprotease and osteopontin 5. We have recently discovered that CD44 homophilic interactions and subsequent CD44-PAK2 interactions mediate tumor cluster aggregation and metastasis 6. While some progress has been made in characterization of CSCs over the last decade, the cellular and molecular mechanisms underlying how CSCs are generated and how their self-renewal is maintained remain largely unknown. The ubiquitin-specific peptidase 22 (USP22) was initially identified as one of the 11 genes in cancer-related death signatures and referred to as the Polycomb/cancer stem cell signature group 7. Further survey of gene expression has shown that the elevated expression of USP22 correlates with poor prognosis in a variety of human tumors including the invasive breast cancer 8, 9. At the molecular level, we and others have recently demonstrated that USP22 functions as an oncogene by inhibiting cell apoptosis and promoting cell cycle progression through targeting cyclins, c-MYC, BMI-1, TRF1, and SIRT1, which controls p53 expression 10, 11, 12, 13, 14, 15. USP22 promotes chemotherapeutic resistance by inhibiting Bax-mediated apoptosis, improving HSP90 function, inhibiting Estrogen receptor a degradation, and driving EGFR recirculation 16. Genetic USP22 suppression inhibits cancer cell growth and induces apoptosis 10, 13. USP22 has been speculated to act as a critical cancer stem cell gene 17, however, the molecular pathways underlying if and how USP22 maintain cancer cell stemness and control CSC self-renewal remain to be fully defined. In this study, we present evidence that USP22 is highly expressed in breast cancer stem cells and required for both breast cancer initiation and metastasis. Both genetic and pharmacological USP22 inhibition largely reduced the breast cancer stem cell pool through down-regulating integrin b1, also known as CD29, a cell surface glycoprotein that is critical in almost every step of cancer progression, including cancer initiation, proliferation, local invasion, and metastatic colonization of the new tissue 18, 19. Interestingly, integrin b1 has been used as a biomarker for isolating breast cancer stem cells 20. Indeed, reconstitution of integrin b1 expression fully rescued the BCSCs pool impaired by USP22 deficiency. At the molecular level, we identified the ITGB1 transcription factor FoxM1 as a de novo substrate of the USP22 deubiquitinase. Therefore, USP22 controls breast cancer stem cell self-renewal through protecting FoxM1 from ubiquitination-mediated proteasomal degradation to enhance ITGB1 transcription. Our study defines the USP22-FoxM-integrin b1 axis as a previously unappreciated pathway in breast cancer initiation and metastasis, that can be therapeutically targeted to antagonize invasive breast cancers. Results USP22 is required for the tumorigenicity of breast cancer stem cells. USP22 has been suggested as a cancer stem cell gene or death-from cancer signature gene and its high expression often predicts the poor clinical outcomes of cancer patients 7, but its role in maintaining CSC stemness remains to be defined. We sorted CD24−CD44+ breast CSCs from the patient-derived Luc2-eGFP (L2G)-labeled breast triple-negative (TN1) cancer cells 6 as well as in the murine breast cancer 4T1 cells (Figure s1 A), and found a significantly higher USP22 expression in breast CSCs compared to that in the CD24+CD44− non-CSCs by western blotting (Fig. 1A & 1B and Figure s1 B). To decipher the USP22 functions in generating and/or maintaining breast cancer cell stemness, we generated USP22 targeted deletion in mouse 4T1 and human breast cancer-derived L2G+ TN1 cells by a CRISPR-Cas9 approach. A complete USP22 deletion was validated by immunoblot analysis (Fig. 1C & s1 C). Importantly, silencing USP22 dramatically reduced the CD24−CD44+ breast CSCs population in L2G+ TN1 and 4T1 cells (Fig. 1D and s1D & s1 E), indicating that USP22 is important for breast CSCs self-renewal. We then utilized a well-established tumor sphere formation assay 21,22 to evaluate the role of USP22 in breast CSCs self-renewal. Indeed, the tumor sphere formation from both patient derived L2G+ TN1 and mouse 4T1 breast cancer cells was largely impaired by USP22 CRISPR deletion, which was further confirmed by an in vitro extremely limiting dilution assay (Fig. 1E–G, s1F-H). Consistently, USP22 inhibition in 4T1 cells resulted in a substantial reduction in colony formation (Figure s1I & J). Therefore, these results indicate that USP22 is required to maintain an optimal CSC population, possibly by controlling CSC self-renewal in vitro. CSCs are a critical small population of cancer cells with potent capability for tumor initiation. To test USP22’s function in promoting tumor initiation in vivo. We orthotopically injected 102, 103, and 104 USP22 knockout or control 4T1 breast cancer cells into BALB/c mice. Surprisingly, in contrast to the fact that five out of eight mice implanted with 102 WT 4T1 cells developed cancer three months after implantation, none of the eight mice received with USP22-deficient 4T1 cells developed breast cancer. Even when a higher number of 4T1 cells, 103 and 104, were orthotopically injected, USP22 deletion dramatically inhibited the development of syngeneic tumors (Fig. 1H & 1I), indicating that USP22 is critical for in vivo tumor initiation. Cancer metastases, per prevailing theory, are predominantly initiated by rare cancer cells that bear stem cell properties 23, 24. We then determined whether USP22 exerted a driving role in breast cancer metastasis by intravenously injection of 4T1 USP22-null or its control WT cells into BALB/c mice. As expected, USP22 deletion dramatically inhibited 4T1 cancer colonization to the lung by reducing more than 60% of tumor nodules with further reduced metastatic foci size (Fig. 1J–L). Immunohistochemistry staining confirmed the deletion of USP22 and detected a significant decrease in the levels of stem cell marker CD44 expression in lung metastasis (Fig. 1M & 1N). As a consequence, USP22 ablation significantly improved the overall survival of the mice with 4T1 lung metastasis (Fig. 10). Collectively, our results revealed that USP22 play an important role in breast CSC maintenance, which is critical for breast cancer initiation and metastasis. USP22 promotes breast CSC self-renewal through upregulating ITGB1 expression Integrin family members are known as key regulators in cancer cell stemness, epithelial-mesenchymal transformation and extracellular matrix to initiate the metastatic process for multiple cancer types including breast cancer 19, 25. Importantly, unbiased analysis of the public database TMIER2 26 revealed a strong and statistically significant positive correlation nearly in all types, 37 out of total 40 of human cancers (Fig. 2A and Supplementary table 1). Further analysis of integrin a and b family revealed a positive correlation of several integrin family members in particular integrin a1,2, 8 & 9 (Figure s2A and Supplementary table 1). These results suggest a possibility that USP22 regulates cancer stem cell self-renewal through regulating the transcription of some of integrin family members. Indeed, comparison analysis by western blotting and flow cytometry of integrin family members between WT and USP22-null breast cancer cells detected a dramatic reduction in integrin b1 expression by USP22 inhibition in both mouse 4T1 and patient-derived L2G+ TN1 cells (Fig. 2B–D). USP22 appears to positively regulate integrin b1 expression at the transcriptional level because its targeted deletion resulted in a more than 70% reduction in ITGB1 mRNA levels (Fig. 2E). In addition to integrin b1, USP22 deletion led to a modest but statistically significant reduction in several additional integrin family members including integrin a1–6 and integrin b2–3, b5–7, but not b4, b8 and a7–8 expression (Figure s2B), In contrast, integrin b6 expression is slightly increased in USP22-null breast cancer cells (Figure s2B). Therefore, USP22 appears to regulate the expression of multiple integrin family members with b1 as the dominant one. We then focused on studying the functional consequences of USP22-mediated integrin b1 upregulation and assessed whether USP22 maintains breast CSC self-renewal and promotes breast cancer metastasis through integrin b1 upregulation by ectopic reconstitution of ITGB1 in USP22 knockout cells (Figure s2C-E). Indeed, ectopic ITGB1 expression partially rescued tumor sphere formation from in both mouse 4T1 and patient-derived L2G+ TN1 USP22-deficient breast cancer cells (Fig. 2F–H). Consequently, expression of integrin b1 largely, but not totally restored 4T1 breast cancer lung metastasis of USP22-null cells as documented by analyzing both lung tumor nodule numbers and the metastatic foci size (Fig. 2I–K). Collectively, these results demonstrate that USP22 enhances BCSCs tumorigenic potential, in part, through integrin b1 upregulation. USP22 functions as a de novo FoxM1-specific deubiquitinase in breast cancer cells. The fact that USP22 deletion reduced ITGB1 mRNA expression suggest that USP22 regulates integrin b1 expression at transcriptional level. Indeed, western blotting analysis revealed a significant reduction in the protein expression of FoxM1, a critical transcription factor for ITGB1 expression 27, in USP22-null breast cancer cells (Fig. 3A). In contrast, USP22 ablation did not alter FoxM1 mRNA levels (Fig. 3B). Together with the fact that USP22 is a deubiquitinase, these results imply that the USP22 exerts its regulatory function on FoxM1 protein expression at the post-transcriptional level. Indeed, treatment of USP22-null cells with the proteasome inhibitor MG132 largely restored FoxM1 expression to a level comparable to that in WT breast cancer cells (Fig. 3A). By contrast, treatment with NH4Cl, an inhibitor of endosome-lysosome degradation pathway, fails to protect FoxM1 from degradation (Figure s3A), suggesting that USP22 promotes FoxM1 level through inhibiting its proteasomal degradation. As a deubiquitinase, USP22 exerts its biological function largely through protecting its downstream substrates from ubiquitination-mediated degradation 28. Accordingly, we speculated that USP22 could be a deubiquitinase of FoxM1. Indeed, USP22 interaction with FoxM1 was detected in HEK-293T cells transiently transfected Myc-USP22 and Flag-FoxM1, but not in control cells transfected with Flag-FoxM1 or Myc-USP22 alone (Fig. 3C). The endogenous interaction between USP22 and FoxM1 in patient-derived breast cancer L2G+ TN1 cells was further validated (Fig. 3D and s3B). USP22 protein carries an N-terminal zinc finger and C-terminal U19 peptidase catalytic domain (Fig. 3E). Truncated mutation analysis revealed that the zinc finger-containing N-terminus is sufficient for USP22 interaction with FoxM1, while the C-terminus ubiquitin-specific peptidase domain is not involved in mediating its FoxM1 interaction (Fig. 3F). These results indicate that FoxM1 physically interacts with USP22 in breast cancer cells. A ubiquitin-specific peptidase often inhibits ubiquitination of its interacting proteins 29. Thus, we determined the effect of USP22 on FoxM1 ubiquitination. Higher molecular weight bands were detected in FoxM1 immunoprecipitants, indicating FoxM1 is ubiquitinated possibly by its endogenous E3 ubiquitin ligases such as FBWX7 30. Importantly, transient USP22 expression largely diminished FoxM1 ubiquitination (Fig. 3G). Conversely, loss of USP22 expression resulted in a significant increase in FoxM1 ubiquitination in both mouse 4T1 and patient-derived breast cancer cells (Fig. 3H). Our data indicate physical interaction between USP22 and FoxM1 is required for USP22-mediated suppression of FoxM1 ubiquitination, because mutation of cystines 61 and 63, which disrupts the zinc finger structure and its interaction with FoxM1 (Fig. 3I), totally abolished USP22 activity in suppressing FoxM1 ubiquitination (Fig. 3G). As expected, expression of the catalytically inactive deubiquitinase, through C185A mutation of USP22, failed to inhibit FoxM1 ubiquitination despite not altering its interaction with FoxM1 (Fig. 3F & 3G). These results indicate that USP22 is a bona fide FoxM1-specific deubiquitinase in breast cancer cells. In concordance with this conclusion, USP22 overexpression dramatically prolonged FoxM1 half-life as measured by pulse-chase analysis (Fig. 3J & 3K). Consistent with the ubiquitination data, neither USP22 C185A nor C61/63A mutant sustained FoxM1 stability (Fig. 3J & 3K). In line with this, USP22 ablation dramatically decreased FoxM1 half-life (Fig. 3L & 3M). Consistently, re-expression of WT USP22, but not its mutants restored integrin b1 expression in USP22-null breast cancer cells (Fig. 3N & 3O). These results define USP22 as a de novo FoxM1 deubiquitinase in breast cancer cells to protect FoxM1 from ubiquitination-mediated proteasomal degradation for upregulating integrin b1 expression. USP22 promotes integrin b1 expression through FoxM1 stabilization. FoxM1 has been identified as an integrin β1 transcription factor thereby promoting breast cancer progression 27, implying a possibility that USP22 controls breast cancer cell ITGB1 expression through FoxMl stabilization. Indeed, reconstitution of FoxM1 expression fully restored the endogenous integrin b1 expression in both USP22-null 4T1 and L2G+ TN1 breast cancer cells as determined by western blotting and qRT-PCR (Fig. 4A & 4B), which was further confirmed by flow cytometry (Fig. 4C & 4D). In contrast, we observed that FoxM1 expression fails to rescue integrin b2–7 expression (Figure s4A). These results support our hypothesis that USP22 specifically promote integrin b1 expression through FoxM1 stabilization. Consistent with this, we observed that ectopic expression of FoxM1 largely restored the tumor sphere formation ability of USP22-deficent breast cancer cells (Fig. 4E–G). Likewise, the impaired ability in colony formation of 4T1 breast cancer cells by USP22 depletion was largely rescued by exogenous FoxM1 expression (Figure s4B & s4C). We also noticed that, while FoxM1 expression fully rescued integrin b1 expression both in USP22-null 4T1 and TN1 breast cancer cells, but their sphere and colony formation were only partially restored by FoxM1 re-expression (Fig. 4E–G). We then utilized the lung metastasis model to further illustrate the role of USP22-FoxM1-integrin b1 pathway in breast cancer tumorigenesis in BALB/c mice. Indeed, in contrast to the fact that USP22 deletion resulted in a more than 50% reduction in lung metastases 4T1 cancer nodules, FoxM1 re-introduction restored USP22-null 4T1 cancer lung metastasis to a level of about 85–90% of the WT (Fig. 4H–J). As a consequence, FoxM1 expression dramatically attenuated but not totally abolished the protection of mice from lung metastasis-induced lethality by USP22 targeted inhibition (Fig. 4K). Collectively, these results indicate that USP22 promotes breast cancer metastasis at least partially, through promoting FoxM1-mediated integrin b1 expression. Pharmacological inhibition of USP22 abrogates BCSCs tumorigenicity. Our discovery that genetic USP22 deletion hindered breast cancer stem cell self-renewal and inhibited their lung metastasis provides a rationale for USP22 targeting in anticancer therapy. We first analyzed the effects of pharmacological USP22 inhibition on BCSCs self-renewal using a small molecule inhibitor USP22i-S02 that we recently identified (Fig. 5A) 31. Similar to our observation from USP22 CRISPR KO studies, treatment of breast cancer cells 4T1 and TN1 significantly inhibited both integrin b1 and FoxM1 expression. Consistent with our previous observations, S02 treatment also reduced USP22 expression levels presumably because USP22 is a deubiquitinase of itself (Fig. 5B). Further addition of the proteasomal inhibitor MG132, but not with lysosome inhibitor NH4Cl, largely rescued FoxM1 protein levels from USP22i-S02 treatment (Figure s5A & s5B), confirming our observation that USP22 inhibition facilitates proteasomal FoxM1 protein degradation. In line with this, treatment of 4T1 cells with S02 dramatically shortened FoxM1 protein half-life (Figure s5C & 5D). As expected, S02 treatment suppressed ITGB1 and other stemness related genes expression, including CD44, ALDH, and NANOG (Fig. 5C). In contrast, S02 treatment did not alter FoxM1 mRNA transcription (Figure s5E). These results confirm that USP22 is a positive regulator for FoxM1-mediated ITGB1 expression in breast cancer cells by an orthogonal pharmacological approach. We next set out to determine the effects of USP22 pharmacological inhibition on breast cancer stem cell self-renewal. As expected, S02 treatment reduced breast CSCs population for more than 80%, to a level that is comparable to USP22 knockout (Fig. 5D and s5F). Importantly, treatment of USP22-null breast cancer cells did not further reduce the frequency of breast cancer stem cells (Fig. 5D and s5F), supporting the high specificity of this USP22-specific small molecule inhibitor. Consequently, treatment with S02 significantly impaired breast cancer cell sphere and colony formation capability (Fig. 5E & F and s5G & H). Further in vitro extremely limiting dilution assay confirmed that S02 inhibited breast CSCs self-renewal (Fig. 5G), implying for its great therapeutic potential in treatment breast cancer. We then used the preclinical 4T1 breast pulmonary metastasis model to illustrate the potential anti-metastatic effect of S02 (Figure s5L). Of note, a six-day treatment with S02 after tail vein injection of 4T1 breast cancer cells resulted in a significant reduction in 4T1 breast cancer lung metastasis and prolonged mice survival (Fig. 5H–K). Further immunohistochemistry analysis of the lung metastatic cancers detected a reduction in both integrin b1 and FoxM1 levels in the S02 treatment groups (Fig. 5L). Similar to our recent study, administration of S02 did not show any detectable toxicity because the mice body weight was unaltered (Figure s5J), and further hematoxylin-eosin (H&E) staining did not detect obvious liver damage in S02 treatment mice (Figure s5K). Therefore, these results indicate that pharmacological USP22 targeting is a safe and effective therapy in treatment of triple negative breast cancers. We then further evaluated the therapeutic potential of USP22i-S02 in a patient-derive xenograft model by orthotopically implanting TN1 cells to immune compromised RAG1 mutant mice (Fig. 5M). Intriguingly, a 3-day treatment of pre-established PDX tumor significantly hindered patient derived xenograft tumor growth (Fig. 5N & 5O). Further characterization by IHC staining show that the levels of USP22, FoxM1 and integrin b1 protein expression by USP22i-S02 treatment, which consequently inhibited the breast cancer cell growth because the percentage of Ki-67+ proliferative cells was dramatically decreased. Importantly, we detected a significant reduction in CD44+ breast cancer cells in the S02 treat group, implying that USP22 pharmacological inhibition attenuates either the breast cancer stem cell self-renewal or their survival (Fig. 5P). Therefore, pharmacological inhibition of USP22 achieves represents a potentially efficacious treatment for breast cancer and metastasis. Positive correlation of USP22, FoxM1 and integrin b1 in human breast cancer. Our data collectively documented that USP22 maintain breast cancer stemness in part through stabilizing ITGB1 transcription factor FoxM1 to promote breast cancer growth and metastasis, which define a previously unknown USP22-FoxM1-ITGB1 pathway in breast cancer pathogenesis. Further analysis of the sorted integrin b1low, integrin b1middle and integrin b1high 4T1 cells revealed a gradual elevation in USP22 and FoxM1 expressions (Fig. 6A & 6B). We then generated a green fluorescent protein (GFP)-USP22 fusion knock-in in 4T1 cells with the endogenous USP22 ablation (Fig. 6C). Consistently, the expression of both FoxM1 and ITGB1 are profoundly increased in USP22high comparing to that in USP22low 4T1 knock-in cells (Fig. 6D & E). Further, a significant increase in integrin b1 and FoxM1 in breast CSCs versus none breast CSCs population was observed (Fig. 6F). To further determine the critical roles of the USP22-FoxM1-integrin b1 in breast cancer pathogenesis, we utilize the immunohistochemistry staining determined the expression of USP22, FoxM1, and integrin b1 protein in human breast cancer tissue microarray (Supplementary table 2). As expected, the protein levels of USP22, FoxM1, and integrin b1 was markedly higher in the breast tumor tissues than those in begin tumors (Fig. 6G, 6H & s6A-D), and levels were even further elevated in metastatic tissues (Fig. 6G, 6H), further supporting our discovery that upregulated USP22 in breast cancer stem cells though FoxM1-mediated ITGB1 gene transcription for promoting breast cancer lung metastasis. To support this notion, the protein expression levels of USP22, FoxM1 and integrin b1 are strongly correlated in human breast cancers (Fig. 6I & s6E). Collectively, our study identified USP22 as a FoxM1-specific deubiquitinase which promotes FoxM1 transcriptional activation for ITGB1 expression, which consequently promotes breast cancer stem cell self-renewal and drives breast cancer metastasis to distal organs including lung (Fig. 6J). Discussion Our study defines a previously unknown USP22-FoxM1-integrin b1 pathway critically important for both mouse and human breast cancer stem cell self-renewal. This conclusion is documented by the following discoveries: First, USP22 is further upregulated in BCSCs and breast cancer and targeted USP22 deletion dramatically impaired BCSC self-renewal and tumorigenicity; Second, USP22 controls breast cancer cell stemness through integrin b1 upregulation; Third, USP22 functions as a bona fide deubiquitinase of the ITGB1 transcription FoxM1 and promotes BCSC self-renewal through FoxM1-mediated integrin b1 expression. Fourth, pharmacological USP22 inhibition impairs BCSC self-renewal and protects mice from breast cancer lung metastasis-induced mortality; Last but not least, USP22 and ITGB1 are positively correlated in more than 90% of human cancer types, and USP22, integrin b1 and FoxM1 are increased and positively correlated in breast cancers. Integrin signals play critical roles in supporting the function of both normal adult stem cells and their neoplastic derivatives 32. While integrin mutations are rarely identified, most of, if not all integrin family members are often upregulated in cancer cell stem cells and this upregulation often promotes CSC self-renewal, cancer initiation and metastasis 19, 33, 34. Several tumor initiating and/or promoting pathways including epidermal growth factor (EGF) and vascular endothelial growth factor-mediated signaling pathways activate the RAS-MAP kinase cascade for ITGB1 transcription through downstream AP-1 family transcription factors. On the other hand, the tumoral immune suppressive cytokine TGF-b promotes b1 integrin expression through canonical SMAD family transcription factor activation. In addition, the fork head family transcription factors, both FoxO3 and FoxM1 have been identified to promote cancer invasion through promotes b1 27, 35. Our studies here define the USP22-FoxM1-integrin b1 axis as critical regulatory node in control of breast cancer stem cell self-renewal, tumor initiation and metastasis. In addition to integrin b1, USP22 appears to promote the transcription of several additional integrin family members. However, FoxM1 reconstitution only rescued integrin b1 expression, implying that USP22 regulates integrin family members via district molecular mechanism. FoxM1 has been also known as a crucial transcription factor for the maintenance of a variety of human CSCs and its expression is associated with a worse clinical prognosis 36, 37, 38. Therefore, this study links three important cancer stem cell genes teaming together to maintain an optimal breast cancer stem cell pool. Importantly, our unbiased analysis of existing public database revealed a statistically significant positive correlation between USP22 and ITGB1 in 37 total human cancer types, suggesting that the USP22-FoxM1-integrin b1 axis is a common mechanism in CSC self-renewal. We also noticed that, while integrin b1 expression is full restored in USP22-null mouse and human breast cancer cells, FoxM1 expression only achieved a partial rescue in their in vitro tumor sphere formation and in vivo lung metastasis, indicating that USP22 exerts it cancer stem cell gene function in part through an integrin b1-indepent manner. Indeed, it has been shown that USP22 promotes hypoxia-induced hepatocellular carcinoma stemness through a HIF-1a/USP22 positive feedback loop upon TP53 inactivation 39. On the other hand, USP22 regulates embryonic stem cell differentiation via transcriptional repression of sex-determining region Y-box 2 (SOX2) 40. Therefore, USP22 appears to play a diverse role in regulating cell stemness in both physiological and pathological contexts. Our study provides a strong rationale for targeting the USP22-FoxM1-integrin β1 pathway in anticancer therapy. In fact, pharmacological USP22 inhibition dramatically reduced the frequency of breast cancer stem cells and attenuated both mouse and human invasive breast cancer lung metastasis. In addition to its cancer cell-intrinsic roles, USP22 has been recently discovered to suppress tumor immunosurveillance through potentiating Foxp3 + regulatory functions 31,41 as well as upregulating the expression of checkpoint receptors PD-L1 and CD73 42, 43. Therefore, USP22 targeting presumably achieves both chemo- and immune-therapeutic efficacy. On the other hand, either specific antibody or peptide inhibitors of integrin family members has been tested for antitumor therapy and several clinical trials are still on going. The anti-a5b1 integrin antibody volociximab was shown to inhibit angiogenesis and suppress tumor growth and metastasis in mice and show some antitumor efficacy in treatment of the advanced non-small-cell lung cancer and in pancreatic cancer 44, 45. However, directly targeting integrin b1 has only achieved very limited success as integrin b1 is highly expressed in a variety of normal cells and required for critical biological functions including normal mammary stem cells maintenance 46. Importantly, our discovery that USP22-FoxM1-integrin b1 pathway is critical for breast cancer self-renewal indicates that simultaneous USP22 and integrin b1 targeting may achieve a synergistic efficacy in combating human cancers, which leads to reduced therapeutic doses and side effects from both sides. Materials And Methods Cell culture Human HEK-293T cells were cultured in DMEM medium plus 10% FBS (Thermo Fisher Scientific,0437028) and 1% penicillin and streptomycin. 4T1 cells were maintained in RPMI medium supplemented with 10% FBS and 1% penicillin and streptomycin. TN1 cells were cultured in HuMEC-ready medium (Life Technologies) supplemented with 5% FBS and 0.5% P/S in collagen type I (BD Biosciences) coated plates. Molecular cloning and plasmid Mice and human FoxM1 overexpressed plasmid were purchased from addgene and subclone into pCMV plasmids. Human or mouse USP22 single guide RNA sequence was ligated into lentiCRISPR v2 plasmid separately. Indicated cells were transiently transfected using TurboFect (Thermo Fisher). 48 hours after transfection, cells were selected using puromycin for 14 days. The efficacy of USP22 deletion was validated by western blotting. The sequences of each guide RNA used in this study was shown Supplementary Table 3. Tumor sphere formation assay A total of 3×104 4T1 or TN1 cells expressing with or without USP22 sgRNA were plated into ultralow-attachment 6-well plates (Corning, Cat#3471), and maintained in EpiCult-B Basal Medium (Human) (Stem Cell Technologies, BC, Canada) and EpiCult-B Proliferation Supplement (Human) (Stem Cell Technologies, BC, Canada), and supplemented with 2 U/mL heparin and 0.5 mg/mL hydrocortisone (Sigma H0135). After 10 days culture, the spheres were pictured, and the number of spheres in each group were counted. Colony formation assay A total of 300 indicated cells were seeded into 35 mm dishes with triplicates, and maintained in culture for two weeks. The culture medium was changed every 3 days. When colonies grew to visible size, the colonies were then washed twice with phosphate buffered saline and fixed with 4% formaldehyde for 30 min at room temperature and stained for 1 h with 0.1% crystal violet. After staining, the plates were gently washed with distilled water and air-dried. The exact colony number of colonies was then quantified by ImageJ software. In vitro extremely limiting dilution assay Indicated cells were dissociated into single cell suspensions and seeded into 96-well plates at density of 5, 10, 15, 20 cells per well using previous mentioned tumor sphere formation medium. Cells were incubated at 37 °C for 10 to 14 days. At the time of quantification, each well was exactly counted for formed tumor spheres. Stem cell frequency was calculated using extreme limiting dilution analysis online tool 47 (http://bioinf.wehi.edu.au/software/elda/). Real-time PCR Total RNA was extracted from indicated cells using Trizol. The cDNA was synthesized using a Quantifect Reverse Transcription Kit. qRT-PCR was performed using SYBR Premix Ex Taq, primers, H2O, and cDNA (final reaction volume, 20 mL). The sequences of the primers used in this study were shown in Supplementary Table 3. Flow cytometry analysis and cell sorting For CD24−/CD44+ BCSCs sorting, 4T1 and TN1 cells were washed with PBS, dissociated using accutase, counted and incubate with primary antibody against CD44 and CD24 on ice for 60 minutes. FacsAria (BD) cell sorter equipment was used to isolate CD24−/CD44+ and CD24+/CD44− cells, respectively. For the 4T1 GFP-USP22 fusion knock-in cells, cells were dissociated using accutase. Cells were then sorted using FacsAria (BD) cell sorter equipment based on fluorescence intensity. For integrin family expression evaluation, indicated cells were dissociated using accutase and followed by staining with indicated antibodies on ice for 60 minutes. Cells were run on the BD-LSR Fortessa X-20 (BD Biosciences) instrument and flow analyses were done using FlowJo software. The detailed information of antibodies used in flow cytometry were shown in Supplementary Table 3. Immunoblot Indicated cells in this study were lysed with RIPA buffer supplemented with protease inhibitors. The same quality of protein was subjected to SDS-PAGE gel electrophoresis, transferred onto polyvinylidene fluoride membranes, and blocked with 5% skimmed milk for 30 min at room temperature. The membranes were then incubated with primary antibodies. The detailed information of antibodies used in this study were shown in Supplementary Table 3. Co-immunoprecipitation TN1 or 4T1 cells were harvested and lysed with RIPA buffer containing protease inhibitors. Cell lysates were precleared using protein A/G beads (10294276, GE healthcare) for 1 h incubation with gentle shake at 4 °C, and precleared protein A/G beads were removed and followed by adding primary antibody for overnight incubation with gentle shake at 4 °C, and new protein A/G beads were subsequently added for another 2 h incubation, then beads were collected following washing with ice-cold PBS for 4 times. Finally, the bound protein was eluted by boiling for 5 min and subjected to SDS-PAGE. Immunohistochemistry Immunohistochemical (IHC) staining was performed following the standard protocol as reported 48, 49. Briefly, tissue specimens were subjected to deparaffinized in xylene, rehydrated through graded ethanol solutions, antigen retrieval and immersed in a 0.3% hydrogen peroxide solution. After carefully washing three times with phosphate-buffered saline (PBS), and nonspecific antigen was then blocked by incubation with 5% bovine serum albumin for 30 min at room temperature. The tissue slides were subsequently incubated with primary antibodies overnight at 4 °C. Horseradish peroxidase (HRP) conjugated secondary antibody was used to incubate the slides before DAB detection. For the IHC results analysis, the percentage score was assigned as follows: 1 indicated that 0–25% of the tumor cells showed positive signaling, 2 indicated 26–50% of cells were stained, 3 indicated 51–75% stained, and 4 indicated 76–100% stained. We scored the staining intensity as 0 for negative, 1 for weak, 2 for moderate, and 3 for strong. The total score was obtained by multiplying the percentage score by the stain intensity score. The detailed information of antibodies used in IHC were shown in Supplementary Table 3. Animal studies All animal experiments were approved by the respective Institutional Animal Care and Use Committee at Northwestern University. All mice were maintained in a specific pathogen-free facility. BALB/c, and NSG mice at the age of 6–8 weeks were all purchased from Jackson laboratory. For the metastatic mice model, BALB/c mice were intravenously administrated with 5×104 4T1 USP22 ablation or control cells. 20 days later, all the mice were sacrificed and analyzed the metastatic nodules. For the survival analysis of mice, BALB/c mice were intravenously administrated with 5×104 4T1 wildtype cells, mice were euthanized until exhibiting signs of significantly declining their quality of life (e.g., ataxia, lethargy, seizures, inability to feed) and the survival of mice were recorded. For the S02 treatment, BALB/c mice were intravenously administrated with 5×104 4T1 cells. 24 hours later, mice were randomized into treatment groups and treated with S02 (10 mg/kg), or vehicle control by intraperitoneal injection six times (once every day). Mice were sacrificed 3 weeks later after 4T1 cells administration and the lung of mice were taken out to analyze tumor nodules. For the orthotopic xenograft model, 5×104 TN1 cells were orthotopically injected into the mammary fat pad of NSG mice, 2 weeks later, mice were randomized into treatment groups and treated with S02 (20 mg/kg), or vehicle control group by intraperitoneal injection six times (twice every day). Statistical analysis Data are represented as the mean ± SD, and error bars indicate SD. P values were calculated by either unpaired or paired two-tailed Student’s t test, *P< 0.05, **P< 0.01, and ***P< 0.001. All analyses were performed using GraphPad Prism software (GraphPad Software, Inc.). Acknowledgement We thank the Northwestern Lurie Cancer Center flow cytometry core for the service support. This work was supported by National Institutes of Health (NIH) grants R01DK126908, R01DK120330, R01CA257520 and CA232347 (to D.F.), R01CA245699 and UG3CA256967 (to H.L.), National Natural Science Foundation of China (No.82073768) and Dalian High-level Talent Innovation Support Program (No.2019RD03) to Z.S., Department of Defense Breast Cancer Research Program W81XWH2010679 (to H.L), and Lynn Sage Breast Cancer Foundation (D.F., and H.L.) Data availability The data are available to academic researchers from corresponding author upon reasonable request. Figure 1 USP22 is required for BCSCs tumorigenicity. A. Immunoblot assessment of USP22 in the matched pairs of tumors sphere-enriched BCSCs and non BCSCs of TN1 cells. Band intensities of USP22 was quantified and the results are expressed as USP22/GAPDH levels relative to control cells. B. Quantification showing that USP22 was highly expressed in BCSCs than in none BCSCs. C. TN1 cells were transduced with single guide RNA (sgRNA) targeting USP22 or a scrambled control sgRNA, and knockout efficiency of USP22 was determined by immunoblot analyses. D. USP22 ablation decreases BCSCs (CD44+CD24−) population in TN1 cells determined by flow cytometry. Representative FACS data are shown. Quantification data showing that BCSCs population in USP22-deficent cells was dramatically decreased than control cells. E. Tumor sphere formation ability was evaluated in TN1 cells expressing either control or USP22 sgRNA, and the representative images of each group are shown. Scale bar, 500 mm. F. Silencing USP22 markedly impairs TN1 cells tumor sphere formation ability. G. The frequencies of tumor sphere formation of TN1 USP22 ablation or control cells determined by in vitro extremely dilution analysis. The significance of the difference between the indicated groups was evaluated by c2 test. n=10. H. Images of tumors developed from mice orthotopically implanted with indicated different gradients 4T1 USP22 ablation or control cells. n=8. I. Quantification showing that USP22 ablation impairs breast cancer initiation. J. Representative images of lung from mice intravenously injected with 4T1 cells expressed either control or USP22 sgRNA. Scale bar, 1 cm. K. The mice were humanely killed after 20 days injection of indicated 4T1 cells. The numbers of metastatic nodules in the lung were significantly decreased in mice injected with 4T1 USP22 knockout cells compared with the numbers in those injected with 4T1 control cells. L. The haematoxylin and eosin (H&E) staining show metastatic tumor. Scale bar, 2 mm. M. Immunohistochemical staining using anti-USP22 or CD44 antibodies were performed on metastatic nodules. Representative images of each group are shown. Scale bar, 20 mm. N. Quantification showing that USP22 knockout induced the decrease of CD44 positive cells. O. Kaplan-Meier survival curve of mice intravenously injected 4T1 cells expressed with or without USP22 sgRNA. Quantifications showing that injecting USP22 ablation cells extends mice survival relative to inject control cells. The significance of the difference between the indicated groups was evaluated by log rank test. n=8. The error bars show the mean ± SD. The significances of differences between different groups were determined by two-tailed Student’s t test. *, *** indicates P < 0.05, P < 0.001, respectively. Figure 2 Depleting USP22 prohibits integrin b1 expression. A. Heatmap of correlation between USP22 and ITGB1 expression in various cancer types. B. Indicated protein expression in TN1 and 4T1 cells expressing either control or USP22 sgRNA were determined by immunoblot analysis. Band intensities of integrin b1 was quantified and the results are expressed as integrin bVGAPDH levels relative to control cells. C. Integrin b1 level was decreased in USP22 ablation cells determined by flow cytometry. Representative FACS data are shown. D. Mean fluorescence intensity (MFI) of integrin b1 level in TN1 and 4T1 USP22-deficent or control cells was quantified. E. The mRNA expression of ITGB1 (gene encoding for integrin b1) in TN1 and 4T1 USP22-deficent or control cells was determined by real-time PCR. b-actin was used as internal control. F-G The representative images (F) and number (G) of tumor sphere formed from TN1 and 4T1 cells transduced with integrin b1 in the setting of USP22 depletion. Scale bar, 500 mm. H. The frequencies of tumor sphere formation of indicated cells. Quantifications showing that introduction of integrin b1 restores tumor sphere formation frequency caused by USP22 depletion evaluated by in vitro extremely limiting dilution analysis. The significance of the difference between the indicated groups was evaluated by c2 test. n=10. I. Representative images of lung from mice intravenously injected with indicated cells. Scale bar, 1 cm. J. The numbers of metastatic nodules in the lung from mice intravenously injected with indicated cells. K. H&E staining of lung metastasis of indicated group. Scale bar, 2 mm. The error bars show the mean ± SD. The significances of differences between different group were determined by two-tailed Student’s t test. **, *** indicates P < 0.01, P < 0.001, respectively. Figure 3 USP22 interacts with and stabilizes FoxM1. A. TN1 and 4T1 stably expressing control or USP22 sgRNA were treated with or without the proteasome inhibitor MG132 (10 mM, 12 h), and then FoxM1 protein expression level was evaluated. Band intensities of FoxM1 were quantified and the results are expressed as FoxM1/GAPDH levels relative to control cells. B. Total RNA was isolated from TN1 and 4T1 cells stably expressing control or USP22 sgRNA. The mRNA levels of FoxM1 were determined by real-time PCR. b-actin was used as internal control. ns means no significant difference. C. Interaction of USP22 with FoxM1. HEK-293T cells were transiently transfected with Flag tagged FoxM1 and Myc tagged USP22. Cell extracts were immunoprecipitated (IP) using primary antibodies against Myc and then subjected to immunoblotting (IB) analysis. WCL means whole cell lysates. D. Endogenous USP22 and control IgG were immunoprecipitated from TN1 cell lysates and then subjected to immunoblotting for analyzing associated proteins. Rabbit IgG was used as the isotype control. E. Schematic representation of the N-terminal Myc-tagged full-length USP22, and various corresponding truncation mutants. F. HEK-293T cells were transfected with the indicated truncated constructs, followed by IP with Myc antibody and followed by immunoblot (IB) with antibodies against Flag. EV means empty vector. G. HEK-293T cells were transiently transfected with Flag tagged FoxM1, HA tagged ubiquitin and Myc tagged USP22 or USP22 C61/63A, C185A mutant. Cell extracts were IP using primary antibodies against Flag and then subjected to IB analysis to analyze FoxM1 ubiquitylation linkage. H. TN1 and 4T1 stably expressing control or USP22 sgRNA cell lysates were subjected to IP with FoxM1 antibody, followed by IB with antibodies against ubiquitin. Cells were treated with 10 mM MG132 for 12 hours before harvesting. I. HEK-293T cells were transiently transfected with Flag tagged FoxM1 and Myc tagged USP22 or USP22 C61/63A, C185A mutant. Cell extracts were IP using primary antibodies against Myc and then subjected to IB analysis. J. HEK-293T cells were co-transfected with Flag tagged FoxM1 and Myc tagged USP22 WT or USP22 C61/63A, C185A mutant for 24 h, followed by treating with 20 mg ml−1 cycloheximide for the indicated times, and cell lysates were subjected to immunoblot with indicated antibodies. CHX means cycloheximide. K. Quantification showing that overexpression USP22 WT, but not USP22 C61/63A, C185A mutant augments FoxM1 half-life. Quantification of FoxM1 relative to GAPDH was quantified by Image J. n=3. L. 4T1 USP22-deficent or control cells were treated with 20 mg ml−1 CHX for the indicated times and cell lysates were examined by immunoblotting. M. Quantification showing that USP22 ablation attenuates FoxM1 half-life. FoxM1 band intensity was quantified and the results are expressed as FoxM1/GAPDH levels relative to untreated cells. n=3. N. Immunoblot analyses of indicated proteins of 4T1 USP22-deficent cells transduced with USP22 WT or indicated mutants. Enforced expression of USP22 WT, but not indicated mutants, rescued the level of FoxM1 and integrin b1 in USP22-deficent cells. O. 4T1 USP22-deficent cells were transiently transduced with USP22 WT or indicated mutants for 48 h. The mRNA levels of FoxM1 or ITGB1 were determined by real-time PCR. Enforced expression of USP22 WT, but not indicated mutants, rescued the level of ITGB1, but not FoxM1, mRNA levels in USP22-deficent cells. b-actin was used as internal control. The error bars show the mean ± SD. The significances of differences between different group were determined by two-tailed Student’s t test. *** indicates P < 0.001. Figure 4 FoxM1 introduction partially rescues the suppressive effects caused by USP22 depletion. A.Immunoblot analysis of integrin b1, Flag and USP22 in TN1 and 4T1 cells transduced with Flag-FoxM1 in the setting of depleted USP22. Enforced expression of FoxM1 completely rescued the protein level of integrin b1 in USP22-deficent cells. B. The mRNA levels of ITGB1 in TN1 and 4T1 cells were determined by real-time PCR. Enforced expression of FoxM1 rescued the mRNA level of ITGB1 in USP22-deficent cells. b-actin was used as internal control. C. Ectopic expression of FoxM1 completely rescued the protein level of integrin b1 in USP22-deficent cells determined by flow cytometry. Representative FACS data are shown. D. Quantification showing that ectopic expression of FoxM1 completely rescued the protein level of integrin b1 in USP22-deficent cells. E. Tumor sphere formed from TN1 and 4T1 cells transduced with Flag-FoxM1 in the setting of depleted USP22. The representative images of tumor sphere are shown. Scale bar, 500 mm. F. Quantifications showing that FoxM1 expression partially rescues the decreased tumor sphere formation ability caused by USP22 depletion. G. The frequencies of tumor sphere formation in TN1 and 4T1 cells expressing Flag-FoxM1 or empty control in the setting of USP22 deficient. H. Representative images of lung derived from mice injected 4T1 cells expressing either control or USP22 sgRNA in combination with vector control or Flag-FoxM1. Scale bar 1 cm. I. Quantification result showing that introduction of Flag-FoxM1 partially rescues the inhibitory effects caused by USP22 depletion. J. H&E staining of lung metastasis of indicated group. Scale bar, 2 mm. K. Kaplan-Meier survival curves of mice implanted with indicated cells. Quantification showing that ectopic expression of FoxM1 in 4T1 USP22-deficient cells shorten mice survival than USP22-deficient cells. Significance testing was done by log rank test. The error bars show the mean ± SD. The significances of differences between different group were determined by two-tailed Student’s t test. *, **, *** indicates P < 0.05, P < 0.01, P < 0.001, respectively. Figure 5 The USP22 inhibitor attenuates breast CSCs self-renewal. A. The chemical structure of USP22 inhibitor S02. B. TN1 and 4T1 WT or USP22-null cells were treated with 20 mM S02 for 24 h. Cell lysates were analysed by immunoblotting using the indicated antibodies. Dimethyl sulfoxide (DMSO) vehicle was used as a control. C. TN1 and 4T1 WT or USP22-null cells were treated with or without 20 mM S02 for 48 h. The mRNA levels of indicated genes in TN1 and 4T1 cells were determined by real-time PCR. b-actin was used as internal control. D. 4T1 WT or USP22-null cells were treated with 20 mM S02 for 48 h, the cells were subsequently stained with CD44 and CD24 antibodies, and then analyzed by flow cytometry. S02 treatment decreases BCSCs (CD44+CD24−) population determined by flow cytometry. Quantification data are shown. E. Tumor sphere formation ability was evaluated in TN1 and 4T1 cells treated with or without 20 mM S02 for 10 days. Representative images of each group are shown. Scale bars, 500 mm. F. Quantification showing that tumor sphere formation ability was restricted by S02 treatment in TN1 and 4T1 WT, but not USP22-deficient cells. G. In vitro extremely limiting dilution assay by plating gradient numbers of TN1 and 4T1 control or USP22 ablation cells showed the frequencies of tumor sphere formation in indicated cells treated with 20 mM S02 for 10 days. H. Representative mages of lungs from mice given intravenous injection of 5×104 4T1 cells. 24 hours later, mice were randomized into treatment groups and treated with S02 (10 mg/kg), or vehicle control by intraperitoneal injection six times (once every day). I. Tumor nodules on the lung of mice injected with S02 or vehicle control. Scale bar, 1 cm. J. The mice were humanely killed after 20 days injection of 4T1 cells. The H&E staining sections show representative metastatic tumour. Scale bar, 2 mm. K. 4T1 cells (5×104 cells per mouse) were intravenously injected into BLAB/c mice. Mice were treated as described in H. The survival of mice was evaluated (n = 8. Kaplan-Meier plotter with two-sided log-rank test). L. Immunohistochemical staining of sections from nodules in the lung as in J stained with antibody against USP22 and integrin b1. M. The scheme for mouse breast cancer treatment model. TN1 cells (5×104 cells per mouse) were orthotopically injected into NOD/SCG mice. Two weeks later when the tumors were reached to around 100 mm3, mice were randomized into treatment groups and treated with S02 (20 mg/kg) or vehicle control by intraperitoneal injection six times (twice a day). N. Images of xenograft tumors after orthotopically injecting TN1 cells and treated with vehicle or S02 by the indicated conditions. Scale bar, 1 cm. O. Weights of xenograft tumor treated with vehicle or S02. P. Immunohistochemical analysis of sections from xenograft tumors treated with vehicle or S02 stained with indicated antibodies. Three individual samples were analyzed and quantification data are shown. The error bars show the mean ± SD. The significances of differences between different group were determined by two-tailed Student’s t test. *, **, *** indicates P < 0.05, P < 0.01, P < 0.001, respectively. Figure 6 Clinical significance of USP22/FoxM1/Integrin b1 signaling axis in breast cancer. A. Flow cytometry was used to isolate low, middle and high integrin b1 cell from 4T1 cells. B. Immunoblot analysis for USP22, FoxM1, and integrin b1 in 4T1 cells isolated according to integrin b1 intensity. L, M, H indicate integrin b1 intensity low, middle, and high, respectively. C. The scheme for GFP-USP22 knock-in 4T1 USP22 knockout cells. D. FACS sorting of GFPlow or GFPhigh cells isolated from GFP-USP22 knock-in in 4T1 USP22 knockout cells. Representative FACS data are shown. E. Immunoblot analysis for FoxM1 and integrin b1 of USP22 knock-in in 4T1 USP22 knockout cells isolated according to GFP intensity. F. BCSCs and none BCSCs were isolated from TN1 and 4T1 cells, respectively. The cell extracts were analyzed by immunoblotting using FoxM1 and integrin b1 antibody. G. Immunohistochemical staining of tissue microarray #1 including 55 specimens (n=8 breast para-tumor specimens, n=20 breast cancer specimens, and n=22 metastatic specimens) for USP22, FoxM1 or integrin b1. Representative consecutive sections from 3 specimens are shown. Scale bars: 200 mm. The clinicopathological characteristics of tissue microarrays are shown in Supplementary Table 2. H. The integrated optical density of USP22 (left panel), FoxM1 (middle panel) or integrin b1 (right panel) was compared with those in indicated groups. I. Linear regression analysis of the integrated optical density of USP22 and FoxM1 (left panel), USP22 and Integrin b1 (middle panel), and FoxM1 and Integrin b1 (right panel) showed a significant positive correlation. Pearson’s R correlation test and the Pearson correlation coefficients are shown in the matrix. n=55. J. Proposed working model. USP22 is a de novo FoxM1 deubiquitinase, which plays essential roles in triggering transcriptional activation of ITGB1, thereby promoting breast cancer stem cells self-renewal and drives breast cancer metastasis. The error bars show the mean ± SD. The significances of differences between different group were determined by two-tailed Student’s t test. *, **, *** indicates P < 0.05, P < 0.01, P < 0.001, respectively. Declarations Conflicts of interest Drs. Deyu Fang and Huiping Liu are co-founders and equity owners of ExoMira Medicine Inc. This is a list of supplementary files associated with this preprint. Click to download. Supplementaryinformation.pdf SupplementaryinformationOriginalimagesofWesternblot.pdf ==== Refs References 1. Ganesh K , Massague J . Targeting metastatic cancer. Nat Med 2021,27 (1 ): 34–44.33442008 2. Valastyan S , Weinberg RA . Tumor metastasis: molecular insights and evolving paradigms. Cell 2011, 147 (2 ): 275–292.22000009 3. Lytle NK , Barber AG , Reya T . Stem cell fate in cancer growth, progression and therapy resistance. Nat Rev Cancer 2018, 18 (11 ): 669–680.30228301 4. Prager BC , Xie Q , Bao S , Rich JN . Cancer Stem Cells: The Architects of the Tumor Ecosystem. Cell Stem Cell 2019, 24 (1 ): 41–53.30609398 5. Weber GF , Ashkar S , Cantor H . Interaction between CD44 and osteopontin as a potential basis for metastasis formation. Proceedings of the Association of American Physicians 1997, 109 (1 ): 1–9.9010911 6. Liu X , Taftaf R , Kawaguchi M , Chang YF , Chen W , Entenberg D , Homophilic CD44 Interactions Mediate Tumor Cell Aggregation and Polyclonal Metastasis in Patient-Derived Breast Cancer Models. Cancer Discov 2019, 9 (1 ): 96–113.30361447 7. Glinsky GV , Berezovska O , Glinskii AB . Microarray analysis identifies a death-from-cancer signature predicting therapy failure in patients with multiple types of cancer. J Clin Invest 2005, 115 (6 ): 1503–1521.15931389 8. Zhang Y , Yao L , Zhang X , Ji H , Wang L , Sun S , Elevated expression of USP22 in correlation with poor prognosis in patients with invasive breast cancer. J Cancer Res Clin Oncol 2011, 137 (8 ): 1245–1253.21691749 9. Prokakis E , Dyas A , Grun R , Fritzsche S , Bedi U , Kazerouni ZB , USP22 promotes HER2-driven mammary carcinoma aggressiveness by suppressing the unfolded protein response. Oncogene 2021,40 (23 ): 4004–4018.34007022 10. Lin ZH , Yang H , Kong QF , Li JP Lee SM , Gao BX , USP22 Antagonizes p53 Transcriptional Activation by Deubiquitinating Sirt1 to Suppress Cell Apoptosis and Is Required for Mouse Embryonic Development. Molecular Cell 2012, 46 (4 ): 484–494.22542455 11. Lin ZH , Tan C , Qiu Q , Kong SY , Yang H , Zhao F , Ubiquitin-specific protease 22 is a deubiquitinase of CCNB1. Cell Discovery 2015, 1 . 12. Zhang XY , Varthi M , Sykes SM , Phillips C , Warzecha C , Zhu W , The putative cancer stem cell marker USP22 is a subunit of the human SAGA complex required for activated transcription and cell-cycle progression. Molecular Cell 2008, 29 (1 ): 102–111.18206973 13. Liu H , Liu N , Zhao Y , Zhu X , Wang C , Liu Q , Oncogenic USP22 supports gastric cancer growth and metastasis by activating c-Myc/NAMPT/SIRT1-dependent FOXO1 and YAP signaling. Aging (Albany NY) 2019, 11 (21 ): 9643–9660.31689236 14. Atanassov BS , Evrard YA , Multani AS , Zhang Z , Tora L , Devys D , Gcn5 and SAGA regulate shelterin protein turnover and telomere maintenance. Mol Cell 2009, 35 (3 ): 352–364.19683498 15. Gennaro VJ , Stanek TJ , Peck AR , Sun YG , Wang F , Qie S , Control of CCND1 ubiquitylation by the catalytic SAGA subunit USP22 is essential for cell cycle progression through G1 in cancer cells. Proceedings of the National Academy of Sciences of the United States of America 2018, 115 (40 ): E9298–E9307.30224477 16. Kosinsky RL , Helms M , Zerche M , Wohn L , Dyas A , Prokakis E , USP22-dependent HSP90AB1 expression promotes resistance to HSP90 inhibition in mammary and colorectal cancer. Cell Death Dis 2019, 10 (12 ): 911.31801945 17. Melo-Cardenas J , Zhang Y , Zhang DD , Fang D . Ubiquitin-specific peptidase 22 functions and its involvement in disease. Oncotarget 2016, 7 (28 ): 44848–44856.27057639 18. Hamidi H , Ivaska J . Every step of the way: integrins in cancer progression and metastasis. Nat Rev Cancer 2018, 18 (9 ): 533–548.30002479 19. Gao Q , Sun Z , Fang D . Integrins in human hepatocellular carcinoma tumorigenesis and therapy. Chin Med J (Engl) 2023. 20. Perez-Nunez I , Rozalen C , Palomeque JA , Sangrador I , Dalmau M , Comerma L , LCOR mediates interferon-independent tumor immunogenicity and responsiveness to immune-checkpoint blockade in triple-negative breast cancer. Nat Cancer 2022, 3 (3 ): 355–370.35301507 21. Liu K , Jiang LP Shi YL , Liu BY , He YM , Shen QS , Hypoxia-induced GLT8D1 promotes glioma stem cell maintenance by inhibiting CD133 degradation through N-linked glycosylation. Cell Death and Differentiation 2022, 29 (9 ): 1834–1849.35301431 22. Ramos EK , Hoffmann AD , Gerson SL , Liu H . New Opportunities and Challenges to Defeat Cancer Stem Cells. Trends Cancer 2017, 3 (11 ): 780–796.29120754 23. Oskarsson T , Batlle E , Massague J . Metastatic stem cells: sources, niches, and vital pathways. Cell Stem Cell 2014, 14 (3 ): 306–321.24607405 24. Pang R , Law WL , Chu AC , Poon JT , Lam CS , Chow AK , A subpopulation of CD26 + cancer stem cells with metastatic capacity in human colorectal cancer. Cell Stem Cell 2010, 6 (6 ): 603–615.20569697 25. Klonisch T , Wiechec E , Hombach-Klonisch S , Ande SR , Wesselborg S , Schulze-Osthoff K , Cancer stem cell markers in common cancers - therapeutic implications. Trends Mol Med 2008, 14 (10 ): 450–460.18775674 26. Li T , Fu J , Zeng Z , Cohen D , Li J , Chen Q , TIMER2.0 for analysis of tumor-infiltrating immune cells. Nucleic Acids Res 2020, 48 (W1 ): W509–W514. 27. Hamurcu Z , Kahraman N , Ashour A , Ozpolat B . FOXM1 transcriptionally regulates expression of integrin beta 1 in triple-negative breast cancer. Breast Cancer Research and Treatment 2017, 163 (3 ): 485–493.28361350 28. Zhou AD , Lin KY , Zhang SC , Chen YH , Zhang N , Xue JF , Nuclear GSK3 beta promotes tumorigenesis by phosphorylating KDM1A and inducing its deubiquitylation by USP22. Nature Cell Biology 2016, 18 (9 ): 954–966.27501329 29. Lin ZH , Yang H , Tan C , Li J F , Liu ZJ , Quan Q , USP10 Antagonizes c-Myc Transcriptional Activation through SIRT6 Stabilization to Suppress Tumor Formation. Cell Reports 2013, 5 (6 ): 1639–1649.24332849 30. Chen YH , Li Y , Xue JF , Gong AH , Yu GZ , Zhou AD , Wnt-induced deubiquitination FoxM1 ensures nucleus beta-catenin transactivation. Embo Journal 2016, 35 (6 ): 668–684.26912724 31. Montauti E , Weinberg SE , Chu P Chaudhuri S , Mani NL , Iyer R , A deubiquitination module essential for T(reg) fitness in the tumor microenvironment. Sci Adv 2022, 8 (47 ): eabo4116.36427305 32. Plaks V , Kong NW , Werb Z . The Cancer Stem Cell Niche: How Essential Is the Niche in Regulating Stemness of Tumor Cells? Cell Stem Cell 2015, 16 (3 ): 225–238.25748930 33. Cooper J , Giancotti FG . Integrin Signaling in Cancer: Mechanotransduction, Stemness, Epithelial Plasticity, and Therapeutic Resistance. Cancer Cell 2019, 35 (3 ): 347–367.30889378 34. Jahangiri A , Aghi MK , Carbonell WS . beta1 integrin: Critical path to antiangiogenic therapy resistance and beyond. Cancer Res 2014, 74 (1 ): 3–7.24327727 35. Hu C , Ni Z , Li BS , Yong X , Yang X , Zhang JW , hTERT promotes the invasion of gastric cancer cells by enhancing FOXO3a ubiquitination and subsequent ITGB1 upregulation. Gut 2017, 66 (1 ): 31–42.26370108 36. Zhang S , Zhao BS , Zhou A , Lin K , Zheng S , Lu Z , m(6)A Demethylase ALKBH5 Maintains Tumorigenicity of Glioblastoma Stem-like Cells by Sustaining FOXM1 Expression and Cell Proliferation Program. Cancer Cell 2017, 31 (4 ): 591–606 e596.28344040 37. Zhang N , Wei P Gong A , Chiu WT , Lee HT , Colman H , FoxM1 promotes beta-catenin nuclear localization and controls Wnt target-gene expression and glioma tumorigenesis. Cancer Cell 2011, 20 (4 ): 427–442.22014570 38. Lam EW , Brosens JJ , Gomes AR , Koo CY . Forkhead box proteins: tuning forks for transcriptional harmony. Nat Rev Cancer 2013, 13 (7 ): 482–495.23792361 39. Ling SB , Shan QN , Zhan QF , Ye QW , Liu P Xu SJ , USP22 promotes hypoxia-induced hepatocellular carcinoma stemness by a HIF1 alpha/USP22 positive feedback loop upon TP53 inactivation. Gut 2020, 69 (7 ): 1322–1334.31776228 40. Sussman RT , Stanek TJ , Esteso P Gearhart JD , Knudsen KE , McMahon SB . The epigenetic modifier ubiquitin-specific protease 22 (USP22) regulates embryonic stem cell differentiation via transcriptional repression of sex-determining region Y-box 2 (SOX2). J Biol Chem 2013, 288 (33 ): 24234–24246.23760504 41. Cortez JT , Montauti E , Shifrut E , Gatchalian J , Zhang Y , Shaked O , CRISPR screen in regulatory T cells reveals modulators of Foxp3. Nature 2020, 582 (7812 ): 416–420.32499641 42. Huang X , Zhang Q , Lou Y , Wang J , Zhao X , Wang L , USP22 Deubiquitinates CD274 to Suppress Anticancer Immunity. Cancer Immunol Res 2019, 7 (10 ): 1580–1590.31399419 43. Gregory S , Xu Y , Xie P Fan J , Gao B , Mani N , The ubiquitin-specific peptidase 22 is a deubiquitinase of CD73 in breast cancer cells. Am J Cancer Res 2022, 12 (12 ): 5564–5575.36628293 44. Li J , Peng L , Chen Q , Ye Z , Zhao T , Hou S , Integrin beta1 in Pancreatic Cancer: Expressions, Functions, and Clinical Implications. Cancers (Basel) 2022, 14 (14 ). 45. Alday-Parejo B , Stupp R , Ruegg C . Are Integrins Still Practicable Targets for Anti-Cancer Therapy? Cancers (Basel) 2019, 11 (7 ). 46. Shackleton M , Vaillant F , Simpson KJ , Stingl J , Smyth GK , Asselin-Labat ML , Generation of a functional mammary gland from a single stem cell. Nature 2006, 439 (7072 ): 84–88.16397499 47. Hu Y , Smyth GK . ELDA: extreme limiting dilution analysis for comparing depleted and enriched populations in stem cell and other assays. J Immunol Methods 2009, 347 (1–2 ): 70–78.19567251 48. Zhang J , Lee S-M , Shannon S , Gao B , Chen W , Chen A , The type III histone deacetylase Sirt1 is essential for maintenance of T cell tolerance in mice. Journal of Clinical Investigation 2009, 119 (10 ): 3048–3058.19729833 49. Lin Z , Yang H , Kong Q , Li J , Lee SM , Gao B , USP22 antagonizes p53 transcriptional activation by deubiquitinating Sirt1 to suppress cell apoptosis and is required for mouse embryonic development. Mol Cell 2012, 46 (4 ): 484–494.22542455

PMC010xxxxxx/PMC10312928.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398265 10.21203/rs.3.rs-3044777/v1 10.21203/rs.3.rs-3044777 preprint 1 Article Analysis of a macrophage carbamylated proteome reveals a function in post-translational modification crosstalk You Youngki Pacific Northwest National Laboratory Tsai Chia-Feng Pacific Northwest National Laboratory Patel Rishi Purdue University Sarkar Soumyadeep Pacific Northwest National Laboratory Clair Geremy Pacific Northwest National Laboratory Zhou Mowei Pacific Northwest National Laboratory Liu Tao Pacific Northwest National Laboratory Metz Thomas O. Pacific Northwest National Laboratory Das Chittaranjan Purdue University Nakayasu Ernesto S. Pacific Northwest National Laboratory Authors’ contributions Contributions: Concept and idea: YY, C-FT, RP, TL, and ESN; Perform the experiments: YY, C-FT, RP, WZ, and GC; Analysis: All authors; Writing: YY, C-FT, RP, TL, and ESN; All authors read and approved the final version of the manuscript. ✉ ernesto.nakayasu@pnnl.gov 16 6 2023 rs.3.rs-3044777https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs3044777v1.pdf Background. Lysine carbamylation is a biomarker of rheumatoid arthritis and kidney diseases. However, its cellular function is understudied due to the lack of tools for systematic analysis of this post-translational modification (PTM). Methods. We adapted a method to analyze carbamylated peptides by co-affinity purification with acetylated peptides based on the cross-reactivity of anti-acetyllysine antibodies. We integrated this method into a mass spectrometry-based multi-PTM pipeline to simultaneously analyze carbamylated and acetylated peptides in addition to phosphopeptides were enriched by sequential immobilized-metal affinity chromatography. Results. By testing the pipeline with RAW 264.7 macrophages treated with bacterial lipopolysaccharide, 7,299, 8,923 and 47,637 acetylated, carbamylated, and phosphorylated peptides were identified, respectively. Our analysis showed that carbamylation occurs on proteins from a variety of functions on sites with similar as well as distinct motifs compared to acetylation. To investigate possible PTM crosstalk, we integrated the carbamylation data with acetylation and phosphorylation data, leading to the identification 1,183 proteins that were modified by all 3 PTMs. Among these proteins, 54 had all 3 PTMs regulated by lipopolysaccharide and were enriched in immune signaling pathways, and in particular, the ubiquitin-proteasome pathway. We found that carbamylation of linear diubiquitin blocks the activity of the anti-inflammatory deubiquitinase OTULIN. Conclusions Overall, our data show that anti-acetyllysine antibodies can be used for effective enrichment of carbamylated peptides. Moreover, carbamylation may play a role in PTM crosstalk with acetylation and phosphorylation, and that it is involved in regulating ubiquitination in vitro. National Institutes of Health (NIH)National Institute of Diabetes and Digestive and Kidney DiseasesU01 DK127786 U01 DK127505 National Cancer Institute’s Clinical Proteomic Tumor Analysis ConsortiumU24CA210955 U24CA271012 National Institute of General Medical SciencesR01GM126296 NIH fellowship1F31CA275390 ==== Body pmcBackground Carbamylation (also known as carbamoylation) is a modification of lysine residue side chains, generating Nε’-carbamyl-lysine or homocitrulline [1]. To date, all known lysine carbamylation of protein and peptides are products of non-enzymatic reactions induced by cyanate and isocyanic acid, formed from urea and thiocyanate, respectively, or by carbamoyl phosphate, an intermediate metabolite of arginine metabolism and nucleotide biosynthesis [1, 2]. Cyanate can be formed in the body by uremia in kidney disease, while isocyanic acid is a product of the pro-inflammatory enzyme myeloperoxidase whose expression is elevated in infectious or autoimmune diseases, such as rheumatoid arthritis. As a result, carbamylation is considered a biomarker for these diseases [3–6]. Despite carbamylation links to various diseases, its roles in pathogenesis and physiology are understudied, mainly due to the lack of tools available for systematic analysis of its function. Proteomics has been an important tool for studying a variety of protein post-translational modifications (PTMs). However, carbamylation represents a major hurdle for the proteomics community. Carbamylation can be generated as an artifact when denaturing proteins with urea [7], a crucial step for efficient proteolysis during proteomics sample preparation. In addition, carbamylation is a major contaminant and confounding factor of lysine acetylome analysis. Lysine acetylation (+ 42.0103 Da) and carbamylation (+ 43.00543 Da) have similar mass; there is particular overlap in mass when comparing carbamylation to the 13C isotope of acetylation (+ 43.0137 Da), a difference of only 8 mDa. Thus, mis-identifications can occur depending on the database searching tool [8]. Another issue is that peptides containing lysine carbamylation can be co-purified with those containing acetylation when using anti-acetyllysine antibodies due to their structural similarities [9]. In this study, we sought to investigate the roles of carbamylation in the RAW 264.7 macrophage cell line by performing a global analysis of this PTM in response to an inflammatory stimulus with bacterial lipopolysaccharide. We took advantage of the co-purification of acetylation and carbamylation to simultaneously analyze both PTMs and performed isotope correction and recalibration to accurately distinguish between the two. We also integrated the data with phosphorylation through a sequential phosphopeptide enrichment of the same sample and further analyzed the data to investigate possible PTM crosstalk and pathways that they might affect. Our data show that carbamylation can be effectively enriched with anti-acetyllysine antibodies. In addition, we showed some characteristics of protein carbamylation and identified a potential role in crosstalk with other PTMs. Methods Cell culture and treatments RAW 264.7 cells were cultivated in DMEM medium containing 10% fetal bovine serum and penicillin/streptomycin at 5% CO2 atmosphere at 37 °C. Cells were treated with 100 ng/mL Salmonella lipopolysaccharide (Invitrogen, cat. No. 00-4976-93) for 24 h at 5% CO2 atmosphere at 37 °C. Cells were washed twice with cold PBS (4 °C), scraped, and harvested into centrifuge tubes. Cells were centrifuged for 5 min at 500 g, the supernatant was discarded, and the pellet was stored at −80 °C for multi-omics analysis. Protein digestion, labeling and peptide enrichment Cell pellets were lysed in 50 mM triethylammonium bicarbonate buffer containing 8 M urea or 12 mM sodium deoxycholate (SDC) at 4 °C for 15 min followed by 95 °C for 5 min. Cell lysates were reduced with 10 mM dithiothreitol at room temperature for 30 min, and cysteine residues were alkylated with 50 mM iodoacetamide at room temperature for 30 min, protected from the light. The reaction was diluted 5-fold with 50 mM triethylammonium bicarbonate buffer and digested with 1:25 trypsin:protein ratio and 1:50 endoproteinase Lys-C/protein ratio overnight at room temperature. Enzyme reactions were stopped by adding trifluoroacetic acid (0.5% final concentration). Peptides were desalted by solid phase extraction using C18 cartridges (Phenomenex) and dried in a vacuum centrifuge. An optimized ratio of TMT to peptide amount of 1:1 (w/w), recently reported by Zecha et al. [10] was used, and samples were fractionated by high pH reverse phase chromatography and concatenated into 12 fractions. Carbamylated and acetylated peptides were enriched with anti-acetyllysine antibodies using PTMScan® Acetyl-Lysine Motif Immunoaffinity Beads (Cell Signaling), following manufacturer recommendations. For analysis of all three PTMs, the samples were first subjected to phosphopeptide enrichment using a recently developed tip-based immobilized metal affinity chromatography (IMAC) method [11], followed co-enrichment of carbamylated and acetylated peptides from the IMAC flow-through using the procedure described above. Mass spectrometry and data analysis Peptides dissolved in 2% acetonitrile and 0.1% trifluoroacetic acid were separated using a reversed-phase column (packed in-house into a 25-cm length of 360 μm o.d. × 75 μm i.d. fused silica picofrit New Objective capillary tubing using ReproSil-Pur 120 C18-AQ 1.9 μm stationary phase) connected to a nanoACQUITY UPLC system (Waters). The analytical column was heated to 50°C using an AgileSLEEVE column heater (Analytical Sales and Services). Peptides were separated through a linear gradient from 8–35% buffer B over 100 min at a flow rate of 200 nL/min. MS analysis was performed using an Orbitrap Fusion Lumos mass spectrometer (ThermoFisher Scientific). Orbitrap precursor spectra (AGC 4×105) were collected from 350–1800 m/z for 110 min at a resolution of 60K along with data-dependent Orbitrap HCD MS/MS spectra (centroid) at a resolution of 50K (AGC 1×105). For acetylation and carbamylation peptide analyses, max ion injection time was set at 125 ms. For phosphopeptide analysis, max ion injection time was set at 105 ms. The total duty cycle was 2 seconds. Precursor ions for MS/MS were isolated (quadrupole) at a width of 0.7 m/z and fragmented using a normalized collision energy of 30%. Peptide mode was selected for monoisotopic precursor scan and charge state screening was enabled to reject unassigned 1 + and > 7+-charged ions with a dynamic exclusion time of 45 seconds. Data were processed with MaxQuant software (v2.1.0.0) [8] by matching against the mouse reference proteome database from Uniprot Knowledge Base (downloaded on August 31, 2020). Searching parameters included protein N-terminal acetylation and oxidation of methionine as variable modifications, and carbamidomethylation of cysteine residues as fixed modification. Searching parameters included protein N-terminal acetylation and oxidation of methionine as variable modifications, and carbamidomethylation of cysteine residues as fixed modification. For phosphoproteome analysis, phosphorylation of serine, threonine and tyrosine residues was set as a variable modification. Acetylation and carbamylation of lysine residues were set as variable modifications. For the motif analysis (see below) data processed with MSFragger (v3.5) [12] using FragPipe (v18.0) with the same parameters than the MaxQuant analysis. Statistical and pathway analysis Reporter ion intensity values were normalized by median centering before submitting to Student’s t-test. Functional-enrichment analysis was done with Database for Annotation, Visualization and Integrated Discovery (DAVID) [13], using the KEGG annotation. Connectivity between proteins was queried in the String database.[14] Additional pathway analysis for the proteins which showed carbamylation, acetylation, phosphorylation was conducted using Reactome [15]. Motif analysis The unique sequences of carbamylated or acetylated lysine residues at the center ± seven adjacent residues. The carbamylation and acetylation motifs were generated with 6,455 carbamylated sites and 5,278 acetylated sites using pLogo (v1.2.0) [16] and 636,113 mouse sequences in pLogo were used for the background sequences. We regenerated the additional motifs after fixing a specific residue that was placed at the ± 1 site of carbamylated or acetylated lysine residues and ranked within the top three in the first motif analysis. Structural analysis Protein structures were downloaded from the Protein Data Bank (PDB) and analyzed with Discovery Studio Visualizer 4.5 program. Cloning, expression, and purification of recombinant proteins OTULINFL and M1-linear diubiquitin was cloned into expression plasmids pCOLD-HisSUMO (Takara Bio) and pET-26b respectively. These plasmids were transformed into E. coli BL21 DE3 and plated against LB-agar containing 100 μg/mL ampicillin and 50 μg/mL kanamycin respectively. A single colony was inoculated into LB media containing the respective antibiotics and grown overnight at 37°C. Cells were grown with shaking at 37°C until an OD600 = 0.6–0.8 was reached. Protein expression was induced with 0.35 mM IPTG for 18 h at 18°C. Cells expressing OTULINFL were harvested at 7000 rpm for 10 min and were resuspended in pH 7.4 phosphate-buffered saline (PBS) with 400 mM KCl containing 0.5 mg/mL lysozyme and lysed using a French press. Lysate was clarified by ultracentrifugation for 1 h at 100,000 g at 4°C and applied to 5 mL of Ni-NTA agarose (Qiagen) resin pre-equilibrated with the respective lysis buffer. The resin is washed with 20 column volumes (CV) of 1X PBS 400 mM KCl, 20 CV of 1X PBS 400 mM KCl containing 25 mM imidazole, and finally eluted with 8 CV of 1X PBS 400 mM KCl containing 300 mM imidazole. The elution was concentrated, and buffer exchanged 1X PBS 1 mM DTT using an Amicon 10 kDa molecular weight cut-off concentrator (Millipore-Sigma). Cells expressing M1-linear diubiquitin were resuspended in 50 mM sodium acetate pH 4.5 and disrupted by French press as described earlier. Cell lysates were heated to 70–80°C for 15 min prior to ultracentrifugation as described above. The clarified supernatant was applied to a self-packed SP Sepharose Fast Flow resin (GE Healthcare) column and eluted with a gradient 1 M NaCl in 50mM sodium acetate buffer. Protein fractions had the purity determined by SDS-PAGE analysis, and were pooled, concentrated, and exchanged into 1X PBS. All protein purity and homogeneity described here is monitored by SDS-PAGE. M1-linear Diubiquitin Carbamylation Carbamylation of M1-linear diubiquitin, purified as described earlier, was carried out by incubation of the dimer in 0.1 M potassium cyanate (AK Scientific) in 1X PBS pH 7.4 at 37°C for 24 h. The carbamylated M1-linear diubiquitin was buffer exchanged by size exclusion chromatography the next day into 1X PBS. Carbamylation was assessed by direct infusion on an Orbitrap Exploris 480 mass spectrometer (Thermo Fisher Scientific). The protein solution was desalted using Amicon Unltra 0.5 mL centrifugal filters (MWCO 3 kDa, Millipore). Then the diluted protein solution at ~ 0.1 μM in 50% acetonitrile 0.1% formic acid was infused at 3 μL/min with HESI source at sheath gas 5, auxiliary gas 7, spray voltage 3.2 kV, ion transfer tube at 320°C, and funnel RF level at 50%. Spectra were acquired at 240k resolution setting in positive mode, and deconvoluted by Xtract within FreeStyle v1.5 (Thermo Fisher Scientific). Deubiquitylating Assay OTULIN activity towards the substrates unmodified and carbamylated M1-linear diubiquitin was carried out in 1X PBS 1mM DTT buffer. The reactions were started by adding equal volumes of enzyme (500 pM final concentration) and substrate (20 μM final concentration) solutions. The reaction was quenched at differing time points (t = 0, 1, 6, and 24 h) by the addition of 5X SDS-PAGE loading dye. The experiment was carried out in triplicate. Results Establishing a method for global lysine carbamylation analysis To establish a lysine carbamylation enrichment method, we tested the ability of anti-acetyllysine antibody to co-capture carbamylated peptides (Fig. 1A). We digested RAW 264.7 cell lysate replicates in buffer containing urea or SDC as denaturing agents. Urea causes carbamylation in vitro and therefore was used as a positive control for carbamylation, whereas SDC does not induce carbamylation and therefore, it was used to detect endogenous sites. After digestion, peptides were labeled with TMT and phosphopeptides were captured by IMAC. The unbound fraction from the IMAC had both acetylated and carbamylated peptides co-captured with anti-acetyllysine antibodies. All fractions were then analyzed by LC-MS/MS. To distinguish between carbamylation and acetylation, we used MaxQuant software, which automatically performs isotope correction of parent ions, recalibrates the mass spectrometry measurements, and performs searches with 4.5 ppm mass tolerance. This process reduces the chances of mismatching carbamylation and acetylation. A total of 63,859 modified peptides were identified, including 7,299, 8,923, and 47,637 acetylated, carbamylated, and phosphorylated peptides, respectively (Fig. 1B, Tab. S1–3). The quantitative analysis showed that the samples digested in buffer containing urea had carbamylated peptides with 2 logs (4-fold) higher average intensity of the reporter ions, confirming that they are in fact carbamylated (Fig. 1C–D). These results showed that anti-acetyllysine can efficiently enrich carbamylated peptides in addition to acetylated peptides. Furthermore, our pipeline provides in-depth coverage of multiple PTMs from the same samples. Pathways differentially carbamylated in RAW 264.7 cells treated with LPS Out of the 8,468 quantifiable carbamylated peptides, 2,378 proteins were found in the samples digested with SDC, showing that carbamylation occurs endogenously in a large number of proteins in cells. To study possible functions of carbamylation in inflammation, we treated RAW 264.7 cells with bacterial lipopolysaccharide (LPS) and analyzed these in parallel with untreated controls (Fig. 2). The carbamylated peptides of both control and LPS treatment groups showed similar average intensity of TMT reporter ions (Fig. 2A). A principal component analysis showed that the carbamylated peptides of the LPS treatment group were clustered together and segregated from the carbamylated peptides of the control group (Fig. 2B). The quantitative analysis showed that 195 endogenously carbamylated peptides from 186 proteins were upregulated by the LPS treatment, while 165 endogenously carbamylated peptides from 148 proteins were downregulated (Fig. 2C). A functional-enrichment analysis showed an overrepresentation of differentially abundant carbamylation in proteins in 38 pathways (Fig. 2D). This included a variety of metabolic pathways (e.g., carbon, amino acid, and porphyrin metabolisms), protein synthesis and processing (e.g., ribosomes and protein processing in the endoplasmic reticulum), RNA synthesis, processing, and degradation (e.g., spliceosome, t-RNA biosynthesis, and RNA degradation), and signaling pathways (e.g., HIF-1 signaling pathway) (Fig. 2D). These results showed that carbamylated proteins from a variety of processes are regulated in the cells by LPS, ranging from cellular metabolism to protein synthesis and signaling pathways. Endogenous carbamylation motif analysis We performed a motif analysis to study carbamylation specificity and to compare against acetyllysine motifs (Fig. 3). Lysine carbamylation was significantly enriched with glutamic acid or phenylalanine at the − 1 position. In the case of glutamic acid, carbamylation occurred nearby hydrophobic residues (Fig. 3A). Acetyllysine was also significantly enriched with glutamic acid at the − 1 position, but an adjacent hydrophobic residue was not as evident (Fig. 3B). Negatively charged residues (aspartic and glutamic acids) were overrepresented nearby carbamylated lysine sites with phenylamine at the – 1 position (Fig. 3C). The same was observed for acetyllysine (Fig. 3D). Positively charged residues (arginine and lysine) were underrepresented near the modified lysine in both motifs with both glutamic acid and phenylalanine at the − 1 position. We also found carbamylation and acetylation motifs containing aspartic acid at the − 1 position. However, in the case of carbamyllysine, this motif was accompanied by an enrichment of proline or lysine at the + 1 position (Fig. 3E–F). In carbamylation, another motif was enriched with phenylalanine at the + 1 position, with adjacent negatively charged amino acids (Fig. 3G). These results show that carbamylation occurs preferentially at the lysine adjacent to negatively charged residues nearby hydrophobic ones. Integration of carbamylation with acetylation and phosphorylation We next investigated possible carbamylation crosstalk with acetylation and phosphorylation by searching for proteins that were commonly modified by these 3 PTMs. We found 1,183 proteins that were commonly modified by all 3 PTMs (Fig. 4A) and they were overrepresented in a variety of pathways such as metabolic pathways and protein degradation pathways (Fig. 4B). Among the 1,183 commonly modified proteins, 54 proteins had the levels of all 3 PTMs regulated by the LPS treatment. We queried the String database to investigate if these proteins were somehow involved in similar functions (Fig. 5). The analysis reviewed a high connectivity between the proteins, indicating that they interact or participate in the same pathways. Of these 54 proteins, 14 were signaling proteins of the immune system (p = 0.0035), and 6 were from the cytokine signaling (p = 0.0184), based on Reactome pathway analysis. There was also an enrichment of proteins related to PTMs involved in the ubiquitin-proteasome pathway, including proteasome subunit alpha type-6 (Psma6), ubiquitin carboxyl-terminal hydrolase 14 (Usp14), Ubiquitin-40S ribosomal protein S27a (Rps27a) (ubiquitin), ubiquitin-activating enzyme E1 (Uba1), and E3 SUMO protein ligase (RanBP2) (Fig. 5). These results suggest a regulation of immune signaling pathways by carbamylation, phosphorylation and acetylation via another PTM, i.e. ubiquitination. Effect of carbamylation on protein ubiquitination We next focused our attention on ubiquitin since this protein was heavily modified by carbamylation, acetylation, and phosphorylation, including sites of each PTM that were regulated by the LPS treatment (Fig. 6A). Because of the ubiquitin role in regulating inflammatory signaling, we asked if any of the carbamylation sites were on lysine residues that might interfere with interaction with other proteins. We examined the complex structure of linear M1-linear diubiquitin bound with the deubiquitinase OTULIN (PDB accession number 3ZNZ) since it is a mechanism of shutting off NF-κB inflammatory signaling [17]. K29, K33, K63 from proximal ubiquitin unit and K11 from the distal one interface with OTULIN within to 2.6 to 5.2 Å by forming hydrogen bonds, electrostatic interactions, or hydrophobic interactions (Fig. 6B). Of these sites, K33 had its carbamylation levels reduced by the LPS treatment (Fig. 6A). To determine if ubiquitin carbamylation can interfere in OTULIN activity, we carbamylated M1-linear diubiquitin with potassium cyanate and incubated with OTULIN for various times. Deubiquitinase activity was assessed by the increase in deconjugated ubiquitin units were analyzed via SDS-PAGE, which indicated that carbamylation completely abolished M1-linear diubiquitin cleavage by OTULIN (Fig. 6C). The carbamylation efficiency was assessed by mass spectrometry and showed the addition of 7–12 carbamyl groups to the M1-linear diubiquitin (Fig. 6D). These results show that carbamylation regulates protein ubiquitination, at least in vitro conditions. Discussion Carbamylation has been viewed by the proteomics community as a major artifact and confounding factor for lysine acetylation analysis by leading to mis-identification and affinity co-purification with anti-acetyllysine antibodies (hence decreased specificity and overall effectiveness in acetylation analysis). Here, we showed that the affinity co-purification with anti-acetyllysine antibodies can be effectively used to study the endogenous carbamylome of a cell. With minor modifications in sample preparation protocol and data analysis, we showed that it is possible to identify and quantify over 7,000 acetylated peptides in addition to over 8,000 carbamylated peptides. This opens opportunity to understand the physiological roles of carbamylation in vivo. The motif analysis showed an enrichment in glutamate, aspartate, and phenylalanine residues close to the carbamylation site. Since carbamylation is non-enzymatic, we believe that these residues can help attract the chemical donor to the modification site. For instance, carbamoyl-phosphate synthase, which produces the carbamylation donor carbamoyl-phosphate, has phenylalanine and glutamate in its catalytic pocket [18]. Myeloperoxidase, which produces the carbamylation donor isocyanic acid, has multiple aspartates and glutamates in its catalytic pocket [19]. We found that carbamylation motifs partially overlap with the acetylation ones. This is expected to some extent. Like carbamylation, acetylation can also occur non-enzymatically. Indeed, in bacteria the major mechanism of lysine acetylation occurs non-enzymatically by acylation of amine groups with acetyl-phosphate [20], a carbamoyl-phosphate analog that increases with the excess of carbon availability in cells. The presence of acetyl-phosphate has not been reported in mammalian cells. However, acetyl-coA, the universal donor for acetyltransferases, can also induce acetylation non-enzymatically, mainly in CoA-binding proteins [21]. Carbamoyl-phosphate is produced in cells as intermediates of arginine and nucleotide synthesis metabolism. Levels of carbamoyl-phosphate increase in cells with excess of nitrogen availability to increase the production of arginine, which is used as an intermediate for urea production and subsequent secretion in urine [22]. Therefore, acetylation and carbamylation could represent an alternating mechanism of protein regulation in carbon and nitrogen excess, respectively. Our data also showed an extensive co-modification of proteins with carbamylation, acetylation and phosphorylation. We found that the ubiquitination machinery to be one of those pathways that were co-modified and regulated by the LPS treatment. Different modifications have been shown to occur in ubiquitin. For instance, phosphorylation of Thr-12 on ubiquitin unit modifying histone H2A has been shown to regulate DNA damage response [23]. Moreover, phosphorylation of Ser-65 inhibits polyubiquitin formation and deconjugation of K63-linked polyubiquitin chains by deubiquitinases [24]. Lysine acetylation inhibits polyubiquitin chain elongation. This phenomenon is not only due to blocking the modification site since it also inhibits ubiquitination of other lysine residues [25]. It has been recently reported that ubiquitin carbamylation inhibits polyubiquitin formation [26]. We now show that protein carbamylation blocks M1-linear ubiquitin chains to be deconjugated by the deubiquitinase OTULIN. The downregulation of K33 carbamylation 24 h after LPS treatment may play a role in regulating inflammation by increasing the activity of OTULIN. However, whether all these modifications work together or in specific processes during inflammation still needs to be further studied. In conclusion, we developed a method to analyze the carbamylomes of samples within a pipeline that simultaneously analyzes acetylomes and phosphoproteomes. This opens opportunities to study post-translational modification crosstalk and novel functions of carbamylation in cells. Acknowledgements Parts of this work were performed in the Environmental Molecular Science Laboratory, a U.S. Department of Energy (DOE) national scientific user facility at Pacific Northwest National Laboratory (PNNL) in Richland, WA. Battelle operates PNNL for the DOE under contract DE-AC05-76RLO01830. Funding National Institutes of Health (NIH), National Institute of Diabetes and Digestive and Kidney Diseases grants U01 DK127786 (to E.S.N.) and U01 DK127505 (to T.O.M.). National Cancer Institute’s Clinical Proteomic Tumor Analysis Consortium grant U24CA210955 and U24CA271012 (to T.L.). National Institute of General Medical Sciences R01GM126296 (to C. D.). R. P. is supported by NIH fellowship under the award 1F31CA275390. Availability of data and materials Mass spectrometry raw data was deposited into the MassIVE repository, which is a member of the ProteomeXchange Consortium. MassIVE accession: MSV000092020 Server: massive.ucsd.edu User: MSV000092020 Password: Acetyl4872 Abbreviations IMAC immobilized metal affinity chromatography LC-MS/MS liquid chromatography tandem-mass spectrometry LPS lipopolysaccharide PTM post-translational modification SDC sodium deoxycholate TMT tandem mass tag Figure 1 Quantitative multi-post-translational modification analysis of RAW 264.7 cells treated with bacterial lipopolysaccharide. (a) Workflow of the sample preparation procedure and proteomics analysis of phosphorylated, carbamylated and acetylated peptides. (b) Number of unique carbamylated, acetylated, and phosphorylated peptides. (c) Box plot of intensities of carbamylated peptides obtained from samples prepared with urea or sodium deoxycholate (SDC) denaturing. (d) Volcano plot of carbamylated peptides intensities from samples denatured with urea vs. SDC. Figure 2 Carbamylated proteomics analysis of macrophages treated with bacterial lipopolysaccharide (LPS). (a) Box plot of carbamylated peptide intensities of LPS-treated vs untreated group. (b) Principal component analysis of carbamylated peptides from LPS-treated and control samples. (c) Heatmap of carbamylated peptides of LPS-treated and control samples. (d) A DAVID functional-enrichment analysis of proteins with carbamylation sites significantly regulated by the LPS treatment. Figure 3 Motif analysis of carbamylation site. Sequence motif logo of carbamylated lysine (a) and acetylated lysine (b). Glutamic acid (c and d), phenylalanine (e and f), or aspartic acid (g and h) at −1 position from the carbamylated or acetylated lysine was fixed in the sequence logos. Phenylamine at +1 position from the carbamylated lysine was fixed in the sequence logo (i). Figure 4 Number of proteins modified with carbamylation, acetylation, or phosphorylation and a functional enrichment of proteins commonly modified by all three post-translational modifications. (a) Venn diagram of carbamylated, acetylated, or phosphorylated proteins. (b) A functional-enrichment analysis of the 1183 proteins commonly modified with all three post-translational modifications. Figure 5 Protein-protein interaction network and functional enrichment analysis of proteins with acetylation, carbamylation and phosphorylation sites regulated by bacterial lipopolysaccharide. The network contains 54 common proteins which had altered levels of carbamylation, acetylation, and phosphorylation by the LPS treatment. The bar graph shows the direction of regulation (p < 0.05) of the three modifications after LPS treatment. The network was enriched in proteins of the innate immune system, cytokine signaling pathway in immune system, and protein post-translation modification pathway using Reactome and they are highlighted the color of pink, blue, green, respectively. The line colors represent if the interactions were experimental determined (pink) and curated in databases (purple). Figure 6 Ubiquitination carbamylation, acetylation, and phosphorylation sites, their regulation by bacterial lipopolysaccharide, and the effect of carbamylation in deubiquitination. (a) Carbamylated, acetylated, or phosphorylated residues on ubiquitin and their regulation by the LPS treatment. (b) Interactions between lysine residues of M1-linear diubiquitin and OTULIN. Proximal ubiquitin and distal ubiquitin are shown in dark and light green ribbon structures, respectively. OTULIN ribbon structure is shown in blue. Lysine residues of M1-linear diubiquitin and residues of OTULIN in magenta and cyan, respectively. Met-1 of the distal ubiquitin, which is linked to the proximal ubiquitin C-terminus, is highlighted as green line structure. Hydrogen bonds, electrostatic interactions, and hydrophobic interactions are shown in green, orange, and pink dashed lines, respectively. All the interactions were identified using Discovery Studio Visualizer 4.5 program. (c) SDS-PAGE of unmodified and carbamylated M1-linear diubiquitin chains incubated with OTULIN. Image is representative of 3 replicates. (d) Mass spectra of unmodified and carbamylated M1-linear diubiquitin chains. Competing interests The authors declare that they have no competing interests. Ethics approval and consent to participate Not applicable Consent for publication Not applicable. ==== Refs References 1. Delanghe S , Delanghe JR , Speeckaert R , Van Biesen W , Speeckaert MM . Mechanisms and consequences of carbamoylation. Nat Rev Nephrol. 2017;13 :580–93.28757635 2. Joshi AD , Mustafa MG , Lichti CF , Elferink CJ . Homocitrullination Is a Novel Histone H1 Epigenetic Mark Dependent on Aryl Hydrocarbon Receptor Recruitment of Carbamoyl Phosphate Synthase 1. J Biol Chem. 2015;290 :27767–78.26424795 3. Wang Z , Nicholls SJ , Rodriguez ER , Kummu O , Horkko S , Barnard J , Reynolds WF , Topol EJ , DiDonato JA , Hazen SL . Protein carbamylation links inflammation, smoking, uremia and atherogenesis. Nat Med. 2007;13 :1176–84.17828273 4. Jaisson S , Pietrement C , Gillery P . Protein Carbamylation: Chemistry, Pathophysiological Involvement, and Biomarkers. Adv Clin Chem. 2018;84 :1–38.29478512 5. Kalim S , Karumanchi SA , Thadhani RI , Berg AH . Protein carbamylation in kidney disease: pathogenesis and clinical implications. Am J Kidney Dis. 2014;64 :793–803.25037561 6. Jaisson S , Kazes I , Desmons A , Fadel F , Oudart JB , Santos-Weiss IC , Millart H , Toure F , Rieu P , Gillery P . Homocitrulline as marker of protein carbamylation in hemodialyzed patients. Clin Chim Acta. 2016;460 :5–10.27302313 7. Kollipara L , Zahedi RP . Protein carbamylation: in vivo modification or in vitro artefact? Proteomics. 2013;13 :941–4.23335428 8. Cox J , Mann M . MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat Biotechnol. 2008;26 :1367–72.19029910 9. Martinez-Val A , Garcia F , Ximenez-Embun P , Martinez Teresa-Calleja A , Ibarz N , Ruppen I , Munoz J . Urea Artifacts Interfere with Immuno-Purification of Lysine Acetylation. J Proteome Res. 2017;16 :1061–8.28067524 10. Zecha J , Satpathy S , Kanashova T , Avanessian SC , Kane MH , Clauser KR , Mertins P , Carr SA , Kuster B . TMT Labeling for the Masses: A Robust and Cost-efficient, In-solution Labeling Approach. Mol Cell Proteomics. 2019;18 :1468–78.30967486 11. Tsai CF , Wang YT , Hsu CC , Kitata RB , Chu RK , Velickovic M , Zhao R , Williams SM , Chrisler WB , Jorgensen ML , A streamlined tandem tip-based workflow for sensitive nanoscale phosphoproteomics. Commun Biol. 2023;6 :70.36653408 12. Kong AT , Leprevost FV , Avtonomov DM , Mellacheruvu D , Nesvizhskii AI . MSFragger: ultrafast and comprehensive peptide identification in mass spectrometry-based proteomics. Nat Methods. 2017;14 :513–20.28394336 13. Sherman BT , Hao M , Qiu J , Jiao X , Baseler MW , Lane HC , Imamichi T , Chang W . DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res. 2022;50 :W216–221.35325185 14. Szklarczyk D , Gable AL , Lyon D , Junge A , Wyder S , Huerta-Cepas J , Simonovic M , Doncheva NT , Morris JH , Bork P , STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47 :D607–13.30476243 15. Fabregat A , Sidiropoulos K , Viteri G , Marin-Garcia P , Ping P , Stein L , D’Eustachio P , Hermjakob H . Reactome diagram viewer: data structures and strategies to boost performance. Bioinformatics. 2018;34 :1208–14.29186351 16. O’Shea JP , Chou MF , Quader SA , Ryan JK , Church GM , Schwartz D . pLogo: a probabilistic approach to visualizing sequence motifs. Nat Methods. 2013;10 :1211–2.24097270 17. Verboom L , Hoste E , van Loo G . OTULIN in NF-kappaB signaling, cell death, and disease. Trends Immunol. 2021;42 :590–603.34074601 18. de Cima S , Polo LM , Diez-Fernandez C , Martinez AI , Cervera J , Fita I , Rubio V . Structure of human carbamoyl phosphate synthetase: deciphering the on/off switch of human ureagenesis. Sci Rep. 2015;5 :16950.26592762 19. Grishkovskaya I , Paumann-Page M , Tscheliessnig R , Stampler J , Hofbauer S , Soudi M , Sevcnikar B , Oostenbrink C , Furtmuller PG , Djinovic-Carugo K , Structure of human promyeloperoxidase (proMPO) and the role of the propeptide in processing and maturation. J Biol Chem. 2017;292 :8244–61.28348079 20. Weinert BT , Iesmantavicius V , Wagner SA , Scholz C , Gummesson B , Beli P , Nystrom T , Choudhary C . Acetyl-phosphate is a critical determinant of lysine acetylation in E. coli. Mol Cell. 2013;51 :265–72.23830618 21. Carrico C , Cruz A , Walter M , Meyer J , Wehrfritz C , Shah S , Wei L , Schilling B , Verdin E . Coenzyme A binding sites induce proximal acylation across protein families. Sci Rep. 2023;13 :5029.36977698 22. Morris SM Jr . Regulation of enzymes of urea and arginine synthesis. Annu Rev Nutr. 1992;12 :81–101.1503815 23. Walser F , Mulder MPC , Bragantini B , Burger S , Gubser T , Gatti M , Botuyan MV , Villa A , Altmeyer M , Neri D , Ubiquitin Phosphorylation at Thr12 Modulates the DNA Damage Response. Mol Cell. 2020;80 :423–436e429.33022275 24. Wauer T , Swatek KN , Wagstaff JL , Gladkova C , Pruneda JN , Michel MA , Gersch M , Johnson CM , Freund SM , Komander D . Ubiquitin Ser65 phosphorylation affects ubiquitin structure, chain assembly and hydrolysis. EMBO J. 2015;34 :307–25.25527291 25. Ohtake F , Saeki Y , Sakamoto K , Ohtake K , Nishikawa H , Tsuchiya H , Ohta T , Tanaka K , Kanno J . Ubiquitin acetylation inhibits polyubiquitin chain elongation. EMBO Rep. 2015;16 :192–201.25527407 26. Pawloski W , Komiyama T , Kougentakis C , Majumdar A , Fushman D . Site-Specific Detection and Characterization of Ubiquitin Carbamylation. Biochemistry. 2022;61 :712–21.35380792

PMC010xxxxxx/PMC10312932.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398446 10.21203/rs.3.rs-2987105/v1 10.21203/rs.3.rs-2987105 preprint 1 Article Thermoring basis for the TRPV3 bio-thermometer Wang Guangyu University of California, Davis Author Contributions Statement: G.W. wrote the main manuscript text and prepared figures 1–6 and Table 1 and Supplementary Information, and reviewed the manuscript. ✉ gary.wang10@gmail.com 02 6 2023 rs.3.rs-2987105https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs2987105v1.pdf The thermosensitive transient receptor potential (TRP) channels are well-known as bio-thermometers with specific temperature thresholds and sensitivity. However, their structural origins are still mysterious. Here, graph theory was used to test how the temperature-dependent non-covalent interactions as identified in the 3D structures of thermo-gated TRPV3 could form a systematic fluidic grid-like mesh network with the thermal rings from the biggest grids to the smallest ones as necessary structural motifs for the variable temperature thresholds and sensitivity. The results showed that the heat-evoked melting of the biggest grids may control temperature thresholds to activate the channel while the smaller grids may act as thermo-stable anchors to secure the channel activity. Together, all the grids along the gating pathway may be necessary for the specific temperature sensitivity. Therefore, this grid thermodynamic model may provide an extensive structural basis for the thermo-gated TRP channels. ==== Body pmcIntroduction The thermosensitive transient receptor potential (TRP) channels are well known as biothermometers involving TRPV (vanilloid), TRPM (melastatin), TRPC (canonical), and TRPA (ankyrin). Their temperature thresholds (Tth) for activation range from noxious cold, cold, warm to noxious heat. Specifically, TRPV1 (> 42°C), TRPV2 (> 52°C), TRPV3 (> 32–39°C), TRPV4 (> 25–35°C), TRPM2, TRPM3, TRPM4, and TRPM5 are involved in warm to hot sensation. In contrast, TRPA1 (< 17°C) or TRPM8 (< 20–28°C) and TRPC5 (< 25–37°C) are sensitive to cold and cool temperatures. When compared to non-temperature-sensitive ones, they also have a high temperature sensitivity Q10 (the ratio of rates or open probabilities Po of an ion channel measured 10°C apart) [1–16]. However, the structural origins of the specific temperature thresholds and sensitivity are still known. Of special interest, TRPV3, which is mainly expressed in skin keratinocytes and oral and nasal epithelia mediating thermal reception and pain sensation [5–6], undergoes sensitization together with TRPV2 while TRPV1 and 4 channels desensitize in response to successive heat stimuli [1, 7–8, 17–19]. Upon initial short heat stimulation within 100 ms, TRPV3 exhibits the high temperature threshold and sensitivity in the noxious temperature range above 50°C. After that intensive stimulation, it becomes responsive to warm temperatures with the low sensitivity. Further studies showed that the insertion of valine at position 412 dramatically eliminates the use-dependent heat sensitization of TRPV3 [19]. Following those findings, the primary cryo-electron microscopy (cryo-EM) structural studies indicated that TRPV3 is a homotetramer. Each monomer has S1-S6 as a transmembrane domain (TMD) and a large intracellular amino- (N-) terminal as an ankyrin repeat domain (ARD). S1-S4 form a voltage-sensor-like domain (VSLD) while S5-S6 and the pore helix and two pore loops are folded as a pore domain. Both the VSLD and the pore domain are swapped via a S4-S5 linker. The TRP helices, which are almost parallel to the membrane, interact with both the skirt ARD and the TMD. Several lipid sites were also found in their interfaces [20]. The pre-S1 domain, together with the carboxyl- (C-) terminal loop domain, couples the TMD with the ARD. The residues 638GLGD641 in the P-loop-extended region line the selectivity filter to permeate partially hydrated Na+, K+ or Ca2+ ions but not to function as an upper gate. In contrast, the narrowest pore constriction around M677 on S6 may act as a lower gate [20–21]. Although the state- and redox-dependent cryo-EM structures of mouse TRPV3 (mTRPV3) with or without the Y564A mutation at different temperatures are available [22–23], the specific structural motifs responsible for the use-dependent temperature threshold and sensitivity have not been pinpointed. On the other hand, following the findings that a nucleic acid hairpin can function as a thermal ring with the number of H-bonds in the stem and the loop length to regulate the melting temperature threshold Tm [24–26], a graph theory-based grid thermodynamic model has been developed to describe proteins as a systematic fluidic grid-like noncovalent interaction mesh network along a single polypeptide chain. Further, the Tm of each grid and the grid-based systematic thermal instability Ti have been defined and calculated and compared with relevant experimental values. In this way, the theoretical and experimental match allows the thermal rings from the biggest grid to the smallest one to be identified as the necessary structural motifs for the thermal stability and activity of globular proteins such as two classes of fructose aldolases from psychrophilic to mesophilic and hyperthermophilic [27–29]. In this regard, it is necessary to test if membrane proteins such as TRPV3 also use such a series of thermo-rings as necessary structural motifs to achieve the use-dependent thermal sensitization. In this computational, graph theory was used to examine this hypothesis by carefully decrypting each grid in the grid-like non-covalently interacting mesh networks as identified in the cryo-EM structures of mTRPV3 with or without the Y564A mutation at different temperatures [22–23]. Once the biggest grid was identified, the calculated Tm was compared with the experimental threshold. Further, the grid-based systematic thermal instability Ti was also calculated as important energetic references to identify different gating states for the use-dependent sensitization. Finally, the systematic structural thermo-sensitivity (Ω10) between any two gating states was also calculated and compared with the experimental Q10 once defined as a heat-evoked change of the total chemical potential of all the grids upon a change in the total enthalpy included in non-covalent interactions along the same gating pathway of one subunit between two gating states within 10°C apart. Once all the three lines of calculated parameters were found to be close to the experimental ones of some redox- and lipid-dependent gating states, a closed and reduced state, a sensitized but oxidized state, and an open and oxidized state were identified with a reasonable energetic sequence for the use-dependent heat sensitization of TRPV3. Results Reduced and PC-free mTRPV3-Y564A had the biggest Grid 13 in the Pre-S1/TRP interface for a calculated Tm 38°C The cryo-EM structures of both closed and open states in detergent-solubilized PC-free mTRPV3-Y564A were first sampled at 37°C after heat sensitization. Therefore, it is necessary to examine if the release of the phosphatidylcholine (PC) lipid from the vanilloid site by the Y546A mutation is responsible for the lower experimental temperature threshold and sensitivity [22]. The previous chimera studies between rat TRPV1 (rTRPV1) and mTRPV3 indicated that the pre-S1 segment 358–434 plays a critical role in mediating the temperature threshold and sensitivity Q10 [30]. On the other hand, the chimera investigations between heat-sensing TRPV1 and cold-sensing TRP melastatin 8 (TRPM8) showed that the C-terminal including the TRP domain (693–710) is required for the polarity of thermal sensitivity [31]. In this regard, the segment from D396 in the pre-S1 domain to K705 in the TRP domain should be at least included as the necessary gating pathway for the temperature threshold and sensitivity and the systemic thermal instability. Along such a gating pathway, the diversity of non-covalent interactions between amino acid side chains in the closed and PC-free Y564 mutant was found after heat sensitization (Fig. 1A). They included 9 H-bonds emerged between different hydrophilic residues, twenty-six π interactions between aromatic residues and nearby residues, and 3 salt bridges between several charged pairs (Fig. 1A, Table S1). When these non-covalent interactions formed a systematic grid-like non-covalent interaction mesh network, the total non-covalent interactions and grid sizes along the gating pathway from D396 to K705 were 38 and 77, respectively (Fig. 1A). Thus, the systemic thermal instability Ti was 2.03 (Table 1). Meanwhile, in addition to the smallest grid with a 0-residue size in the VSLD, the biggest Grid13 with a 13-residue size appeared in the pre-S1/TRP interface via the shortest path from D396 to Y409, R698, R696, W433, K432 and back to D396 to control the D396-K432 salt bridge (Fig. 1B–D). When 1.0 equivalent H-bond sealed this grid, the removal of the PC lipid from the vanilloid site by the Y564A mutation allowed a calculated Tm of 38°C near the experimental Tm 37°C (Table 1) [22]. The melting of the biggest Grid13 at 37°C initiated channel opening of reduced and PC-free mTRPV3-Y564A with a low Ω10 comparable to the low Q10 When the mTRPV3-Y564A mutant opened with the melting of Grid13 in the TRP/pre-S1 interface at 37°C, the disruption of the D396-K432 salt bridge triggered several changes in the systematic grid-like non-covalent interaction mesh network. In addition to one salt bridge, three H-bonds and 17π interactions were conserved, two bridges, six H-bonds and eight π interactions were replaced with three new salt bridges, six new H-bonds and eight new π interactions (Tables S1 and S2). Thus, the total non-covalent interactions along the gating pathway from D396 to K705 had only a minor change from 38 to 39 (Figs. 1A and 2A). On the other hand, the smallest Grid0 was conserved with a zero-residue size via the shortest path from F441 to Y565, Y448, F449, F445 and back to F441 (Figs. 1A and 2A). Thus, it may serve as a thermostable anchor against which two smaller grids may favor channel opening. One was Grid4 with a 4-residue size via the shortest path from D519 to W521, V525, F626, Y565, R567, Q695, R698 and back to D519 in the VSLD/TRP interface (Figs. 2A–B, E); the other was the biggest Grid11 in the TRP/VSLD/pre-S1 interfaces to control the H417-E689 π interaction. It had an 11-residue size via the shortest path from T411 to H417, E689, W692, R696, R698, D519, S515, and back to T411 (Figs. 2A, 2C–E). Once 2 equivalent H-bonds sealed the grid, the calculated Tm was about 52°C (Table 1). In any way, the disruption of the D396-K432 salt bridge in the biggest Grid13 induced a global conformational change from the pre-S1 domain to the VSLD, the TRP domain, the S4–55 linker and the pore domain (Figs. 1A and 2A). However, the grid sizes along the gating pathway from D396 to K705 had only a minor change from 77 to 74 (Figs. 1A and 2A). In this case, the systemic thermal instability Ti was 1.90, and the calculated Ω10 was in a range from 0.76 to 4.30 and with a mean value 1.48, which was comparable to the experimental Q10 (~ 1.21) (Table 1) [22]. In other words, the removal of the PC lipid from the vanilloid site by the Y564A mutation allowed reduced mTRPV3 to have the very low structural and functional thermo-sensitivities. On the other hand, oxidized mTRPV3 with a disulfide bond between C612 and C619 in the outer pore has also been reported to open from a PC-bound closed state at a lower threshold 42°C after repeated heat sensitization from 25°C to 40°C [23]. Therefore, it is exciting to test if oxidation also allows a low structural temperature sensitivity Ω10 to be responsible for the measured functional temperature sensitivity Q10 (1.9–3.1) [19]. Closed PC-bound mTRPV3 with the disulfide bond in the outer pore had the biggest Grid17 in the Pre-S1/VSLD interface for a calculated Tm 40°C after heat sensitization Regarding the PC-bound closed state of oxidized mTRPV3 after heat sensitization, much more non covalent interactions than those in the PC-free closed state of reduced mTRPV3-Y564A shaped a distinct systematic fluidic grid-like non-covalent interaction mesh network (Figs. 1A and 3A). In the presence of the C612-C619 disulfide bond in the pore domain, four salt bridges (E610-K614 was merged into the C612-C619 disulfide bond), fifteen H-bonds and forty π interactions were identified between D396 and K705 (Fig. 3A, Table S3). Since the total non-covalent interactions and grid sizes along the gating pathway from D396 in the pre S1 domain to K705 in the TRP domain were 59 and 72, respectively (Fig. 3A), the grid-based systemic thermal instability Ti was about 1.22 (Table 1). Despite several smallest grids with a zero-residue size, the biggest Grid17 with a 17-residue size was outstanding in the VSLD/pre-S1 interface to control the D519-R416 salt bridge (Fig. 3B–D). It started with D519 and went through W521, F522, Y564, Y565, F441, W433 and ended with R416 (Fig. 3E). When two equivalent H- bonds sealed the grid, the predicted Tm was about 40°C (Table 1), which was close to the measured Tm 42°C. [23] The melting of the biggest Grid17 at 42°C drove oxidized mTRPV3 opening with a low Ω10 comparable to the low Q10 In the heat-activated open state, following the melting of the R416-D517 salt bridge in the biggest Grid17 at 42°C as predicted (Fig. 3D) [23], although one salt bridge, five H-bonds, and 35π interactions were conserved, three salt bridges, ten H-bonds, and six π interactions were substituted by two new salt bridges, eight new H-bonds, and one new π interaction (Figs. 3A and 4A, Tables S3 and S4). However, two smallest Grid0 with a zero-residue size were still conserved as anchors near the R416-D519 salt bridge: one via the shortest path from F445 to Y565, Y448, F449, and back to F445, and the other via the shortest path from Y448 to Y565, Y564, F526 and back to Y448 (Figs. 3A and 4A). Therefore, the following gating pathway against these two anchors was proposed. First, in the VSLD/pre-S1/TRP interfaces, the D519-R416 salt bridge was substituted by the T411-R416 and D519-R567 H-bonds. As a result, the T397-E704 H-bond and the D396-K432 salt bridge were broken with the formation of H417/E418-R690 and E423-T427 H-bonds. In addition, the K432-E704 salt bridge became an H-bond (Fig. 4A). Second, when R567 H-bonded with D519 in the VSLD, a smaller Grid2 with a 2-residue size appeared via the shortest path from D519 to W521, F522, Y564, Y565, R567 and back to D519. Consequently, the V528-F524 π interaction was disconnected (Fig. 4A). Third, when the conformational wave extended to the S4-S5/TRP interface, the R567-T699 H-bond was disrupted and the E689-R693 H-bond changed to a salt bridge (Fig. 4A). Fourth, when this conformational wave continued to the pore domain, the F597-F601 π interaction and the N647-E610/K614 and E682-K686 H-bonds and the E610-K614 salt bridge were disconnected. In the meanwhile, the E631-K634 H-bond and the F633-I637 π interaction were present, and the H-bond moved from S621-Q646 to S620-Q646 (Fig. 4A). Taking all these changes into account, after the biggest Grid17 in the VSLD/pre-S1 interface melted above the predicted 40°C, the PC lipid was released from nearby W521 and Q695 and thus the new biggest Grid9 with a 9-residue size was created in the S5-S6 interface, which may be required for channel opening (Fig. 4B–C). When two equivalent H-bonds sealed Grid9 via the shortest path from D586 to F590 and L673 and T680 and back to D586 (Figs. 4C, 4E), the calculated Tm was about 56°C (Table 1). That may be why the temperature limit is 57°C for stable efficacy [19]. Since Grid9, together with Grid7 with a 7-residue size via the shortest path from F590 to Y594, T636, Y661, T665, L673 and back to F590, was conserved in both closed and open states of oxidized mTRPV3 (Figs. 3A and 4A), they may act as thermostable anchors to secure channel activity. In the meanwhile, a smaller Grid3 with a 3-residue size in the pre-S1/VSLD/S4-S5 linker/TRP/pre-S1 interfaces may be required to stimulate the lower state of the channel. It linked multiple active residues together including W433, F441, Y565, Y564, F522, W521, D519, R567, Q570, W692, and R696 (Figs. 4D, 4E). As a result, the total non-covalent interactions and grid sizes along the gating pathway from D396 to K705 decreased from 59 and 72 to 52 and 65, respectively (Figs. 3A & 4A). Such a decrease produced a low systemic thermal instability Ti value as 1.25. More importantly, the systematic structural thermo-sensitivity Ω10 was in a range from1.88 to 14.3 and with a mean value 4.12 (Table 1), which was close to the experimental Q10 (1.9–3.1) [19]. Therefore, even if the PC lipid at the corresponding vanilloid site was not released, the presence of the C612-C619 disulfide bond in the outer pore may be adequate for mTRPV3 to open with both low Tm and Ω10 to match the measured Tth and Q10 in response to the second heat stimulation [19]. In that regard, it is attractive to test if the disruption of the C612-C619 disulfide bond can increase the Tm and the Ω10 upon channel opening from reduced mTRPV3 to meet the requirement of the higher Tth (> 50 C) and Ω10 (16.4–22.6) [19]. The melting of the biggest Grid12 at the vanilloid PC site above Tm 50°C was required to release PC from reduced mTRPV3 for channel opening with a high Ω10 In the absence of the C612-C619 disulfide bond, reduced mTRPV3 had some different noncovalent interactions to form the systematic grid-like non-covalent interaction mesh network at 4°C when compared with the closed and oxidized one at 42°C after heat sensitization (Figs. 4A & 5A, Table S5) [32]. In the pore domain, after the E610-K614 salt bridge and the E610-N647 and S621-Q646 and E682-K686 H-bonds were disrupted, the F625-V629 and F633-I637 and T649-Y650 π interactions and the Y594-Y661 H-bond emerged (Fig. 5A). When this conformational change extended to the S4-S5 linker/TRP interface, the R567-T699 H-bond and the R698-E702 salt bridge were broken but the Q570-E689 H-bond replaced the Q570-W692 CH- π interaction. As a consequence, along with the T566-S576 H-bond in the S4-S5 linker/VSLD interface, D519 H-bonded with T411 in the VSLD/pre S1 interface and formed an additional salt bridge with R698 in the VSLD/TRP interface (Fig. 5A). When this conformational change extended to the VSLD, the T456-W559 H-bond was broken, the H471-Y540/Y547 π interaction network changed to the Y448/Y551-Q529 H-bonds, the π interaction moved from F542-Y544 to Y540-Y547, and the H-bond shifted from K500-E501 to Q514-S518. When the PC bridge moved from W521-PC-Q695 to W521-PC-F524/R567 (Figs. 4B & 5B), in addition to a salt bridge between R567 and the PC lipid and the CH-π interactions of the PC lipid with W521 and F524, Y564 formed a CH-π interaction with R567 (Fig. 5A). By all account, the disruption of the C612-C619 disulfide bond brought about the biggest Grid12 at the vanilloid PC site (Fig. 5B). When two equivalent H-bonds governed the 12-atom path from W521 to PC to F524 (Figs. 5C–D), the predicted melting temperature was about 50°C (Table 1), which was close to the initial experimental Tth 52°C for TRPV3 opening [19]. On the other hand, when compared with oxidized mTRPV3 in both closed and open states, only four H-bonds and 26 π interactions were conserved, and two new salt bridges and four new H-bonds and six new π interactions were added (Tables S3, S4 and S5). As the total non-covalent interactions and grid sizes along the gating pathway from D396 to K705 were 55 and 96, respectively (Fig. 5A), the systemic thermal instability (Ti) was 1.75 (Table 1). When the same open state as shown in the oxidized and PC-free mTRPV3 was employed (Fig. 4A), the melting of the W521-PC-F524 bridge in Grid12 would produce the calculated Ω10 ranging from 8.76 to 58.5 with a mean value 18.3, which was approximate with the experimental Q10 (16.4–22.6) (Table 1) [19]. Thus, the initial high Tth and Q10 of mTRPV3 upon the brief heat stimulation may result from the gating transition from the reduced and closed state to the open and oxidized one. Discussion Thermo-sensitive TRPV3 is characterized as the use-dependent heat sensitization. Although several cryo-EM structures of mTRPV3 are available in different gating and redox states and at various temperatures, the specific structural motifs responsible for this use-dependent heat sensitization are still missing. This computational study first demonstrated that the calculated melting temperature threshold Tm of the biggest grid along the gating pathway in mTRPV3 was comparable to not only the structural Tm but also the functional activation threshold Tth of mTRPV3. It further confirmed that the functional thermo-sensitivity Q10 was also comparable to the grid-based structural thermo-sensitivity Ω10. Finally, the grid-based systematic thermal instability values of mTRPV3 in different redox- and lipid-dependent gating states were also compared with each other to establish the energetic relationship of different gating states. Taken as a whole, three gating states were completely identified to account for the use-dependent heat sensitization of TRPV3. First, it was further confirmed that the biggest grid may employ its size and strength to determine the melting temperature threshold (Tm) of TRPV3. At a given salt concentration (150 mM NaCl), for reduced and sensitized mTRPV3-Y564A, when 1.0 equivalent H-bond sealed the biggest Grid13, it had a calculated Tm of about 38°C in the closed state. In agreement with this prediction, the D396-K432 salt bridge in this Grid13 was broken in the open state at 37°C (Figs. 1A and 2A, Table 1) [22]. For oxidized and sensitized mTRPV3 in the closed state, when two equivalent H-bonds sealed the biggest Grid17, it had a calculated Tm 40°C, and the R416-D519 salt bridge was also disrupted at 42°C in the open state (Figs. 3A & 4A, Table 1) [23]. Second, if the functional temperature threshold Tth for activation of mTRPV3 is controlled by the melting temperature threshold (Tm) of the biggest grid along the gating pathway, the calculated Tm should be comparable to the measured threshold Tth. In accordance with this prediction, reduced mTRPV3 in the closed state had a Tth around 52°C which was close to the calculated Tm 50°C of the biggest Grid12 at the vanilloid PC site (Fig. 5A, Table 1) [19]. Once oxidized, mTRPV3 had a low calculated Tm of 40°C (Table 1). Since the tunable distance between R416 in the pre-S1 domain and D519 in the VSLD, the R416-D519 salt bridge may change the strength which ranged from 0.5 to 2.0 equivalent H-bonds, That may account for the warm activation of mTRPV3 in a range from around 25–40°C [5–6, 19, 33]. Hence, the activation threshold Tth may be governed by the melting of the biggest grid via the adjustable R416-D519 salt bridge along the gating pathway. Since the calculated Tm values of the biggest Grid9 and Grid11 in the open states of oxidized mTRPV3 and reduced mTRPV3-Y564A were about 56°C and 57°C, respectively (Figs. 2A & 4A, Table 1), the biggest grids along the gating pathway may be responsible for the optimal activity temperature range of the mTRPV3 bio-thermometer [19]. Third, increased temperature has been reported to accelerate the dissociation rate kd of enthalpy-driven non-covalent interaction in a biophysical network but to slow down kd of entropy-driven crosslinks to a different extent [34]. In this study, an increase in the opening rate or the open probability Po of TRPV3 has been observed with raised temperatures [23]. If the temperature threshold for TRPV3 opening is governed by a rate-limiting single step to disrupt a non-covalent interaction in the biggest grid along the gating pathway, TRPV3 opening would be initially enthalpy-driven (△H<0). In this case, when thermo-gated TRPV3 opens from a closed state within 10°C, the functional thermo-sensitivity (Q10) should be comparable to the calculated systematic structural thermo-sensitivity Ω10 because they both factually reflect the change of the total chemical potentials of all the grids upon the alteration of the total enthalpy included in the non-covalent interactions along the same gating pathway from D396 to K705. In agreement with this proposal, if wild-type mTRPV3 had the same open and oxidized state, the calculated mean Ω10 of reduced mTRPV3 would be 18.3, which was close to the measured Q10 (16.4–22.6) (Table 1) [19]. For oxidized and sensitized mTRPV3, the calculated mean Ω10 was 4.12, which was similar to the measured Q10 (1.9–3.1) (Table 1) [19]. For the reduced mTRPV3-Y564A mutant, the calculated mean Ω10 was 1.48, which was near to the measured Q10 1.21 (Table 1) [22]. Thereafter, the functional thermo-sensitivity Q10 may be governed by the grid-based systematic structural thermo-sensitivity Ω10 as defined. In this regard, when the intensity of a non-covalent interaction was in the range from 0.5 to 3 kJ/mol, the resultant Ω10 ranges from the minimum to the maximum may be theoretically calculated as 8.76–58.5 for reduced and closed mTRPV3, 1.88–14.3 for sensitized and oxidized mTRPV3, and 0.76–4.3 for reduced and sensitized mTRPV3-Y564A (Table 1). Taken together, it is proposed that reduced mTRPV3 may start the first activation above the calculated Tm 50°C upon the fast heat stimulation. Once the channel is opened, it is oxidized to form the C612-C619 disulfide bond so that the functional thermo-sensitivity Q10 (16.4–22.6) can keep consistent with the calculated Ω10 (18.3) (Fig. 6A, Table 1). It is further proposed that when the temperature declines, oxidized but sensitized mTRPV3 may decrease a Tth to 30–40°C and Q10 to 1.9–3.1 as a result of the formation of the C612-C619 disulfide-bond (Fig. 6A, Table 1) [19]. In this way, the lower threshold 30–40°C may increase the open probability in response to the same temperature jump from 32°C to 59°C so that mTRPV3 activation exhibits the use-dependent sensitization upon successive heat stimuli [19]. In direct line with this use-dependent sensitization, the grid-based systemic thermal instability (Ti) in the closed state was 1.75 for reduced mTRPV3, and slightly decreased to 1.22 when oxidized by heat sensitization in favor of the use-dependent heat-sensitization during channel opening with a similar Ti of 1.25 (Table 1). In contrast, the Y564A mutation increased the grid-based systemic thermal instability (Ti) from 1.75 to 2.03 for the sensitized but closed state and 1.90 for the open state (Table 1). In other words, the Y564A mutation may increase the systemic thermal instability in favor of spontaneous channel opening [22]. On the other hand, when reduced or Cys-less mTRPV3 is exposed to the long and slow heat stimulation, the channel can be activated above a threshold Tth 30°C [33]. Therefore, it is also possible that either the formation of the C612-C619 disulfide bond by air oxidation or the Cys-less mutation may increase the length of the R416-D519 salt bridge to account for the declined Tth of 30°C so that the PC lipid could be released from the vanilloid site for channel opening (Fig. 6A, Table 1). When the insertion of valine at position 412 disrupts the T411-D519 salt bridge and the related smaller Grid4 via the shortest path from T411 to R416 and D519 and then back to T411 (Figs. 5A, 6B), the same biggest Grid17 as shown in oxidized but closed mTRPV3 may be followed in the VSLD/pre-S1 interface in favor of the release of the vanilloid site lipid for channel opening below 40°C in response to the first fast heat stimulus (Figs. 3A, 5A, Table 1). That may be why the insertion of valine at 412 removes the use-dependent sensitization upon repeated heat stimuli [19]. In support of this proposal, when the Y564A mutation release the PC lipid from the vanilloid site, it also had a low calculated Tm (38°C) and Ω10 (1.48) to keep consistent with the low threshold (< 37°C) and the low Q10 (1.21), respectively (Fig. 1–2, Table 1) [22]. In any way, several smaller anchor grids in the pore domain may be important to stablize the common open state in favor of high heat efficacy. In the pore domain, the first was Grid7 with the shortest path from F590 to Y594, T636, Y661, T665, L673 and back to F590, and the second was Grid9 with the shortest path from D586 to F590, L673 and T680 and then back to D586 (Figs. 4A, 4C, 6B). It has been reported that the T636S mutation decreases the temperature threshold [33], and the mutation N643S, I644S, N647Y, L657I, Y661C or T680A is actually less sensitive to heat or slows down the activation rate [30, 32, 35–36]. Therefore, it is possible that these mutations may affect the thermostability of these smaller anchor grids in the pore domain. In contrast, four smallest grids with a zero-residue size in the VSLD may form a basic stable backbone anchor system for mTRPV3 activation or a fuse group to keep a low systemic thermal instability (Figs. 3– 5, 6B): the first Grid0 via the shortest path from F447 to W493, Y451, Y448 and back to F447; the second Grid0 via the shortest path from Y448 to Y451, N452, W559, F449 and back to Y448; the third Grid0 via the shortest path from Y448 to F526, Y564, Y565 and back to Y448; the forth Grid0 via the shortest path from F445, Y448, F449 and back to F445. Further experiments may be required to test if non-covalent interactions in other grids than anchors are essential for heat-evoked TRPV3 opening. For example, the D519-R416 salt bridge, the H417/E418-R690 and D519-R567 and S620-Q646 and Q529-Y448/Y451 H-bonds, the F441-Y565 and H471-Y540/Y547 and Y622/Q646-F654 π−π interactions, and the Q570-W692-R696-W433-K438 cation/CH-π interactions (Figs. 3A, 4A, 5A). CONCLUSION In this computational study, a graphical grid thermodynamic model has bridged crystallographic static conformations with electrophysiological dynamic findings together by using graph theory in atomic details. Once the thermal rings in the systematic fluidic grid-like mesh network of non-covalent interactions along the gating pathway were tested and identified as key deterministic structural factors or motifs for thermo-gated mTRPV3, three gating states could be in turn established to account for the use-dependent heat sensitization of TRPV3. Accordingly, this model can be used to predict the thermal stability and activity of cellular biological macromolecules including membrane proteins once high-resolution 3D structural data are available. METHODS Data mining resources In this in silico study, two groups of the temperature-dependent cryo-EM structures of mTRPV3 in different gating and redox states were analyzed by graph theory to abstract the structural bioinformation for the use-dependent temperature thresholds and sensitivity. One included reduced and sensitized mTRPV3-Y564A at 37°C in the detergent (PDB ID, 6PVO, model resolution = 5.18 Å) and reduced and open mTRPV3-Y564A (PDB ID, 6PVP, model resolution = 4.4 Å) at 37°C in the detergent [22]; The other covered oxidized and sensitized WT mTRPV3 in cNW11 at 42°C (PDB ID, 7MIN, model resolution = 3.09 Å), oxidized and open WT mTRPV3 in cNW11 at 42°C (PDB ID, 7MIO, model resolution = 3.48 Å) [23], and reduced and closed WT mTRPV3 in MSP2N2 at 4°C (PDB ID, 6LGP, model resolution = 3.31 Å) [32]. Standards for non-covalent interactions In order to secure that results could be reproduced with a high sensitivity, the same strict standard definition as described previously as well as structure visualization software, UCSF Chimera, was exploited to identify stereo- or regio-selective inter-domain diagonal and intra-domain lateral non-covalent interactions in the 3D strcutres of mTRPV3 (Tables S1-S5) [27–29]. They included salt-bridges, CH/cation/lone pair/π−π interactions and H-bonds along the gating pathway from D396 to K705 in mTRPV3 with or without the Y564A mutation. Notably, although the hydrophobic effect and residue hydrophobicity are necessary to drive protein folding, their effects on protein stabilization may be rather marginal [37–38]. Preparation of topological grid maps by using graph theory The identified non-covalent interactions were geometrically mapped as edges along with marked node arrows to represent the positions of the linked residues in the systematic fluidic mesh network according to the same protocol as previously described [27–29]. All the grids were then covered in this network after their ring sizes were constrained as the minimal number of the total free side chains of residues or atoms in the bound lipid that did not participate in any non-covalent interction in a grid. The size constraint was completed by using graph theory and the Floyd-Warshall algorithm to calculate the shortest return path from one end of a non-covalent interaction to the start because the direct path from the start to the end was zero. [39]. For example, in the intra-subunit grid-like biochemical reaction mesh network of Fig. 3A, a direct path length from E610 and N647 was zero because of an H-bond between them. However, there was another shortest return path from N647 to K614 and back to E610 via the N647-K614 H-bond and the K614-E610 salt bridge in this grid. Therefore, the grid size was zero. Once all the grid sizes were available, only the uncommon sizes were marked in black, and a grid with an x-residue or atom size was denoted as Gridx. When the total number of all noncovalent interactions and grid sizes along the gating pathway were calculated, they were displayed in black and blue circles beside the mesh network map, respectively, for the calculations of the systematic thermal instability and the structural temperature sensitivity. Calculation of the temperature threshold of mTRPV3 A DNA hairpin thermo-sensor with a 20-base loop and two G-C base pairs in the stem has a start control melting temperature threshold Tm of 34°C to initiate thermal unfolding of the hairpin loop. When an additional G-C base or five additional bases are included in the hairpin, the Tm is increased by 10°C [26]. In a similar way, when a single polypeptide chain in protein carries out rate-limiting thermal unfolding of the thermal rings from the biggest grid to the smallest grid, the Tm of thermal unfolding of the given grid along the chain was calculated by using the following equation as described previously [27–29]: (1) Tm∘C=34+(n−2)×10+20−Smax×2 where, n is the total number of the grid size-controlled H-bonds equivalent to non-covalent interactions in the given grid, and Smax is the size of the given grid. In this regard, the more grid’s heat capacity will be expected with the decreased grid size or the increased equivalent H-bonds. Calculation of the systemic thermal instability Ti On the other hand, the Tm of the DNA hairpin will be always increased by the more G-C base pairs in the stem or the shorter the poly-A loop [26]. Thus, the grid-based systemic thermal instability Ti along the single polypeptide chain was reasonably defined using the following equation as described previously [27–29]: (2) Ti=S/N where, S is the total grid sizes and N is the total non-covalent interactions along the gating pathway of one subunit in a gating state. Usually, the lower Ti, the less the conformational entropy in the system. Calculation of the systematic temperature sensitivity of mTRPV3 For initial enthalpy-driven TRPV3 opening upon decyclization of the biggest grid (ΔH<0), if a thermosensitive TRPV3 channel changes from a fully closed state to a fully open state within a temperature range ΔT, and if the chemical potential of a grid is defined as the maximal potential for equivalent residues in the grid to form a tight and ideal β -hairpin with the smallest loop via non-covalent interactions [40], the grid-based systematic structural thermo-sensitivity ΩΔT of a single ion channel can be defined and calculated using the following equations: (3) ΩΔT=Sc−SoE/2(Hc/Ho)=Sc−SoE/2[(ENc)/(ENo)]=Sc−SoE/2(Nc/No) where, in the closed and open states along the same gating pathway of one subunit, Hc and Ho are the total enthalpy included in non-covalent interactions, respectively; Nc and No are the total non-covalent interactions, respectively; Sc and So are the total grid sizes, respectively. E is the energy intensity of a non-covalent interaction in a range of 0.5–3 kJ/mol. Usually, E is 1 kJ/mol. Thus, ΩΔT actually mirrors a heat-evoked change in the total chemical potential of all the grids upon a heat-induced change in the total enthalpy included in non-covalent interactions from a closed state to an open state along the same gating pathway of one subunit. When ΔT=10∘C, Ω10 could be comparable to the functional thermo-sensitivity (Q10) of a single ion channel. Q10 was defined and calculated using the following equation: (4) Q10=X2/X110/(T2−T1) where, X1 and X2 are open probability Po values or reaction rates obtained at temperatures T1 and T2 (measured in kelvin), respectively. Acknowledgements: The author’s own studies cited in this article were supported by NIDDK Grant (DK45880 to D.C.D.) and Cystic Fibrosis Foundation grant (DAWSON0210) and NIDDK grant (2R56DK056796-10) and American Heart Association (AHA) Grant (10SDG4120011 to GW). Data availability statement: All data generated or analysed during this study are included in this published article and Supplementary Information. Conventions and abbreviations ARD ankyrin repeat domain cryo-EM cryo-electron microscopy Tm, melting temperature PC phosphatidylcholine Tth temperature threshold TMD transmembrane domain TRP transient receptor potential TRPA TRP ankyrin TRPC TRP canonical TRPVi TRP vanilloid i TRPV3 TRP vanilloid-3 rTRPV1 rat TRPV1 mTRPV3 mouse TRPV3 TRPMi TRP melastatin i TRPM8 TRP melastatin 8 VSLD voltage-sensor-like domain WT wild-type Figure 1 The grid-like non-covalently interacting mesh network along the gating pathway of PC-free reduced mTRPV3-Y564A in the sensitized state at 37 °C after heat sensitization. A, The topological grids in the systemic fluidic grid-like mesh network. The cryo-EM structure of one subunit in detergent-solubilzed mTRPV3-Y564A, which was PC-free, reduced, sensitized but closed at 37 °C (PDB ID, 6PVO), was used for the model. The pore domain, the S4-S5 linker, the TRP domain, the VSLD and the pre-S1 domain are indicated in black. Salt bridges, p interactions, and H-bonds between pairing amino acid side chains along the gating pathway from D396 to K705 are marked in purple, green, and orange, respectively. The grid sizes required to control the relevant non-covalent interactions were calculated with graph theory and labeled in black. The total grid sizes and grid size-controlled non-covalent interactions along the gating pathway are shown in the blue and black circles, respectively. B, The location of the biggest Grid13 is marked in a red circle. C, The structure of the biggest Grid13 with a 13-residue size in the TRP/pre S1 interface to control the D396-K432 salt bridge. D, The sequence of the biggest Grid13 to control the D396-K432 salt bridge in a blue rectangle. Figure 2 The grid-like non-covalently interacting mesh network along the gating pathway of PC-free reduced mTRPV3-Y564A in the open state at 37 °C after heat sensitization. A, The topological grids in the systemic fluidic grid-like mesh network. The cryo-EM structure of one subunit in detergent-solubilzed and open and reduced mTRPV3-Y564A without PC bound in cNW11 at 37 °C (PDB ID, 6PVP) was used for the model. The pore domain, the S4-S5 linker, the TRP domain, the VSLD and the pre-S1 domain are indicated in black. Salt bridges, p interactions, and H-bonds between pairing amino acid side chains along the gating pathway from D396 to K705 are marked in purple, green, and orange, respectively. The grid sizes required to control the relevant non-covalent interactions were calculated with graph theory and labeled in black. The total grid sizes and grid size-controlled non-covalent interactions along the gating pathway are shown in the blue and black circles, respectively. B, The structure of the smaller Grid4 with a 4-residue size to control the D519-R698 salt bridge and F526-Y565 p interaction in the grid. C, The location of the biggest Grid11 is marked in a red circle. D, The structure of the biggest Grid11 with an 11-residue size in the VSLD/TRP/pre S1 interfaces to control the H417-E689 and T411-S515 H-bonds. E, The sequences of two smaller Grid4 and Grid11 to control the aformentioned non-covalent interactions in the blue boxes, respectively. Figure 3 The grid-like non-covalently interacting mesh network along the gating pathway of PC-bound oxidized mTRPV3 in the sensitized state at 42 °C after heat sensitization. A, The topological grids in the systemic fluidic grid-like mesh network. The cryo-EM structure of one subunit in sensitized and oxidized mTRPV3 with PC bound in cNW11 at 42 °C (PDB ID, 7MIN) was used for the model. The pore domain, the S4-S5 linker, the TRP domain, the VSLD and the pre-S1 domain are indicated in black. Salt bridges, p interactions, and H-bonds between pairing amino acid side chains along the gating pathway from D396 to K705 are marked in purple, green, and orange, respectively. The grid sizes required to control the relevant non-covalent interactions were calculated with graph theory and labeled in black. The total grid sizes and grid size-controlled non-covalent interactions along the gating pathway from D396 to K705 are shown in the blue and black circles, respectively. C, The location of the biggest Grid17 is marked in a red circle. D, The structure of the biggest Grid17 with a 17-residue size in the VSLD/pre S1 interface to control the strong R416-D519 salt bridge. E, The sequence of the biggest Gird17 to control the R416-D519 salt bridge in a blue box. Figure 4 The grid-like non-covalently interacting mesh network along the gating pathway of PC-free oxidized mTRPV3 in the open state at 42 °C after heat sensitization. A, The topological grids in the systemic fluidic grid-like mesh network. The cryo-EM structure of one subunit in open and oxidized mTRPV3 without PC bound in cNW11 at 42 °C (PDB ID, 7MIO) was used for the model. The pore domain, the S4-S5 linker, the TRP domain, the VSLD and the pre-S1 domain are indicated in black. Salt bridges, p interactions, and H-bonds between pairing amino acid side chains along the gating pathway from D396 to K705 are marked in purple, green, and orange, respectively. The grid sizes required to control the relevant non-covalent interactions were calculated with graph theory and labeled in black. The total grid sizes and grid size-controlled non-covalent interactions along the gating pathway are shown in the blue and black circles, respectively. B, The location of the biggest Grid9 is marked in a red circle. C, The structure of the biggest Grid9 with a 9-residue size in the S5-S6 interface to control the D586-T680 H-bond. D, The structure of the putative smaller Grid3 with a 3-residue size for the lower gate. E, The sequences of two smaller gating Grid9 and Grid3 to control the D586-T680 H-bond and a group of crtitical cation-p interactions in the blue boxes, respectively. Figure 5 The grid-like non-covalently interacting mesh network along the gating pathway of PC-bound reduced mTRPV3 in the closed state at 4 °C without heat sensitization. A, The topological grids in the systemic fluidic grid-like mesh network. The cryo-EM structure of one subunit in reduced and closed mTRPV3 with PC bound in MSP2N at 4 °C (PDB ID, 6LGP) was used for the model. The pore domain, the S4-S5 linker, the TRP domain, the VSLD and the pre-S1 domain are indicated in black. Salt bridges, p interactions, and H-bonds between pairing amino acid side chains along the gating pathway from D396 to K705 are marked in purple, green, and orange, respectively. The grid sizes required to control the relevant non-covalent crosslinking interactions were calculated with graph theory and labeled in black. The total grid sizes and grid size-controlled non-covalent crosslinking interactions along the gating pathway are shown in the blue and black circles, respectively. The T411-D519 H-bond, which is marked in a dashed line, may be disrupted by the insertion of valine or serine at position 412. B, The location of the biggest Grid12 is marked in a red circle. C, The structure of the biggest Grid12 with a 12-atom size at the PC site to control the W521-PC-F524 bridge. D, The sequence of the biggest Grid12 to control the W521-PC-F524 bridge in a blue box. Figure 6 The tentative model for the use-dependent heat sensitization of thermo-gated TRPV3. A, The homo-tetrameric cryo-EM structures of mTRPV3 in the reduced and closed state (PDB ID: 6LGP), the sensitized and oxidized (PDB ID: 7MIN), and the open and oxidized state (PDB ID: 7MIO) were used for the model. For a convenient view, only two opposite subunits are shown. The dashed rectangle is the membrane area. In the absence of the C612-C619 disulfide bond, reduced mTRPV3 is open with a high Q10 (16.4–22.6) upon the short heat stimulus above 50 °C to melt the biggest Grid12, a thermo-active ring (red), at the vanilloid PC site and then to release PC from the vanilloid site. Meanwhile, C612 is close to C619 enough to form a disulfide bond. When the temperature decreases below 40 °C, oxidized and PC-free mTRPV3 is closed with the biggest Grid17, another thermo-active ring (red), in the VSLD to decrease the threshold from 50 °C to 30–40 °C. Upon the second short heat stimulus, the sensitized and oxidized mTRPV3 has a Q10 as low as 1.9–3.1. However, the long and slow warm stimulation above 30 °C can oxidize mTRPV3 to decrease the threshold from 50 °C to 30 °C in favor of the release of the PC lipid from the vanillloid site for channel opening but with a low Q10 (1.66). B, The proposed smaller thermo-stable anchor grids against which mTRPV3 opens above the temperature thresholds (PDB ID, 7MIO). The grid sizes are shown in the red circles. Table 1 The grid thermodynamic model-based new parameters of the mTRPV3 bio-thermometer along the gating pathway from D396 to K705 Construct WT mTRPV3 mTRPV3-Y564A PDB ID 6LGP 7MIO 7MIN 6PVP 6PVO Lipid PC at the vanilloid site bound free bound free free Redox state reduced oxidized oxidized reduced reduced Lipid environment MSP2N2 cNW11 cNW11 detergent Sampling temperature, °C 4 42 42 37 37 Gating state Closed ↔ Open ↔ Sensitized Open ↔Sensitized # of the biggest grid Grid12 Grid9 Grid17 Gridn Grid13 Biggest grid size (Smax) 12 9 17 11 13 Equivalent H-bonds in Smax 2.0 2.0 2.0 2.0 1.0 Total non-covalent interactions 55 52 59 39 38 Total grid sizes, a.a 96 65 72 74 77 Calculated Tm, °C 50 56 40 52 38 Measured Tm, °C 42 37 Measured Tth, °C 52 57 32–39 Systemic thermal instability (Ti) 1.75 1.25 1.22 1.90 2.03 Calculated Ω10, min at Emin=0.5 kJ/mol 8.76 1.88 0.76 Calculated Ω10, mean at Emean=1.0 kJ/mol 18.3 4.12 1.48 Calculated Ω 10, max at Emax=3.0 kJ/mol 58.5 14.3 4.30 Measured Ω 10, 16.4–22.6 1.9–3.1 1.21 Ref. for measured Tth or Q10 [19] [19] [19] [19] Declarations Conflict of interest: The author declares no conflict of interest. ==== Refs References 1. Caterina M.J. , Schumacher M.A. , Tominaga M. , Rosen T.A , Levine J.D. and Julius D. (1997) The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature. 389 , 816–824.9349813 2. Caterina M.J. , Rosen T. A. , Tominaga M. , Brake A. J , Julius D. (1999)A capsaicin-receptor homologue with a high threshold for noxious heat. Nature, 398 ,436.10201375 3. McKemy D. D. , Neuhausser W. M. & Julius D. (2002) Identification of a cold receptor reveals a general role for TRP channels in thermosensation. Nature, 416 , 52–58.11882888 4. Peier A.M. , Moqrich A. , Hergarden A.C. , Reeve A.J. , Andersson D.A. , Story G.M. , Earley T.J. , Dragoni I. , McIntyre P. , Bevan S. and Patapoutian A. (2002) A TRP channel that senses cold stimuli and menthol. Cell, 108 , 705–715.11893340 5. Peier A. M. , Reeve A. J. , Andersson D.A. , Moqrich A. , Earley T.J. , Hergarden A.C. , Story G.M. , Colley S. , Hogenesch J. B. , McIntyre P. , Bevan S. , Patapoutian A. (2002) A heat-sensitive TRP channel expressed in keratinocytes. Science, 296 , 2046.12016205 6. Smith G. D. , Gunthorpe M. J. , Kelsell R. E. , Hayes P. D. , Reilly P. , Facer P. , Wright J. E. , Jerman J. C. , Walhin J-P. , Ooi L. , Egerton J. , Charles K. J. , Smart D. , Randall A. D. , Anand P. , Davis J. B. (2002) TRPV3 is a temperature-sensitive vanilloid receptor-like protein. Nature, 418 , 186.12077606 7. Xu H. , Ramsey I. S. , Kotecha S. A. , Moran M. M. , Chong J. A. , Lawson D. , Ge P. , Lilly J. , Silos-Santiago I. , Xie Y. , DiStefano P. S. , Curtis R. , Clapham D. E. (2002) TRPV3 is a calcium-permeable temperature-sensitive cation channel. Nature, 418 , 181.12077604 8. Guler A.D. , Lee H. , Iida T. , Shimizu I. , Tominaga M. and Caterina M. J. (2002) Heat-evoked activation of the ion channel, TRPV4. J Neurosci, 22 6408–14.12151520 9. Chung M. K. , Lee H. , Caterina M. J. (2003) Warm temperatures activate TRPV4 in mouse 308 keratinocytes. J. Biol. Chem., 278 , 32037.12783886 10. Story G.M. , Peier A.M. , Reeve A.J. , Eid S.R. , Mosbacher J. , Hricik T.R. , Earley T.J. , Hergarden A.C. , Andersson D.A. , Hwang S.W. , McIntyre P. , Jegla T. , Bevan S. , Patapoutian (2003) A. ANKTM1, a TRP-like channel expressed in nociceptive neurons, is activated by cold temperatures. Cell, 112 , 819–829.12654248 11. Jordt S.E. , Bautista D.M. , Chuang H.H. , McKemy D.D. , Zygmunt P.M. , Högestätt E.D. , Meng I.D. and Julius D. (2004) Mustard oils and cannabinoids excite sensory nerve fibres through the TRP channel ANKTM1.Nature, 427 , 260–265.14712238 12. Talavera K , Yasumatsu K. , Voets T. , Droogmans G. , Shigemura N. , Ninomiya Y. , Margolskee R. F. , Nilius B. (2005) Heat activation of TRPM5 underlies thermal sensitivity of sweet taste. Nature, 438 , 1022.16355226 13. Vriens J , Owsianik G. , Hofmann T. , Philipp S. , Stab J. , Chen X. , Benoit M. , Xue F. , Janssens A. , Kerselaers S. , Oberwinkler J. , Vennekens R. , Gudermann T. , Nilius B. , Voets T. (2011) TRPM3 is a nociceptor channel involved in the detection of noxious heat. Neuron, 70 , 482.21555074 14. Zimmermann K. , Lennerz J.K. , Hein A. , Link A.S. , Kaczmarek J.S. , Delling M. , Uysal S. , Pfeifer J.D. , Riccio A. and Clapham D.E. (2011)Transient receptor potential cation channel, subfamily C, member 5 (TRPC5) is a cold-transducer in the peripheral nervous system. Proc. Natl. Acad. Sci.USA., 108 , 18114–18119.22025699 15. Song K. , Wang H. , Kamm G. B. , Pohle J. , Reis F. C. , Heppenstall P. , Wende H. , Siemens J. (2016) The TRPM2 channel is a hypothalamic heat sensor that limits fever and can drive hypothermia. Science, 353 ,1393.27562954 16. Tan C. H. , McNaughton P. A. (2016) The TRPM2 ion channel is required for sensitivity to warmth. Nature, 536 , 460.27533035 17. Luo L , Wang Y , Li B , Xu L , Kamau PM , Zheng J , Yang F , Yang S , Lai R . (2019) Molecular basis for heat desensitization of TRPV1 ion channels. Nat Commun. 10 , 2134. Correction in: Nat Commun. (2020) 11, 1798. 31086183 18. Liu B. & Qin F. (2016) Use dependence of heat sensitivity of vanilloid receptor TRPV2. Biophys J,. 110 , 1523–1537.27074678 19. Liu B. & Qin F. (2017) Single-residue molecular switch for high-temperature dependence of vanilloid receptor TRPV3. Proc. Natl.Acad. Sci. USA. 114 , 1589–1594.28154143 20. Singh A.K. , McGoldrick L.L. and Sobolevsky A.I. (2018) Structure and gating mechanism of the transient receptor potential channel TRPV3. Nat. Struct. Mol. Biol. 25 , 805–813.30127359 21. Zubcevic L. , Herzik M. A. Jr. , Wu M. , Borschel W.F. , Hirschi M. , Song A.S. , Lander G.C. , and Lee S. Y. (2018) Conformational ensemble of the human TRPV3 ion channel. Nat. Commun. 9 , 4773.30429472 22. Singh A.K. , McGoldrick L.L. , Demirkhanyan L. , Leslie M. , Zakharian E. , and Sobolevsky A.I. (2019) Structural basis of temperature sensation by the TRP channel TRPV3. Nat. Struct. Mol. Biol. 26 , 994–998.31636415 23. Nadezhdin K. D. , Neuberger A. , Trofimov Y. A. , Krylov N. A. , Sinica V. , Kupko N. , Vlachova V. , Zakharian E. , Efremov R. G. , and Sobo-levsky A. I. (2021) Structural mechanism of heat-induced opening of a temperature-sensitive TRP channel. Nat. Struct. Mol. Biol. 28 , 564–572.34239124 24. Kuznetsov S.V. , Ren C.C. , Woodson SA , Ansari A. (2008). Loop dependence of the stability and dynamics of nucleic acid hairpins. Nucleic Acids Res. 36 , 1098–112. doi: 10.1093/nar/gkm1083.18096625 25. Ke G. ; Wang C. ; Ge Y. ; Zheng N. ; Zhu Z. ; Cao Y. ; Yang J. L-DNA molecular beacon: A safe, stable, and accurate intracellular nano-thermometer for temperature sensing in living cells. J. Am. Chem. Soc. 2012, 134 , 18908–18911.23126671 26. Jonstrup A. T. , Fredsøe J. , Andersen A. H. (2013). DNA Hairpins as Temperature Switches, Thermometers and Ionic Detectors. Sensors. 13 , 5937–5944. doi:10.3390/s130505937.23666126 27. Wang G. (2022). The network basis for the structural thermostability and the functional thermoactivity of aldolase B. BioRxiv. DOI: 10.1101/2022.10.20.513014; The network basis for the structural thermostability and the functional thermoactivity of aldolase B. Molecules. 28 , 1850. 28. Wang G. (2023) Network basis for the heat-adapted structural thermostability of bacterial class II fructose bisphosphate aldolase. ACS Omega. 8 , 17731–17739. 10.1021/acsomega.3c00473 37251155 29. Wang G. (2023) Thermal ring-based heat switches in hyperthermophilic class II bacterial fructose aldolase. ACS Omega. 10.1021/acsomega.3c03001. 30. Yao J. , Liu B. , Qin F. (2011).Modular thermal sensors in temperature-gated transient receptor potential (TRP) channels. PNAS 108 , 11109–11114 21690353 31. Brauchi S ,, Orta G. , Salazar M. , Rosenmann E. , Latorre R. (2006).A hot-sensing cold receptor: C-terminal domain determines thermosensation in transient receptor potential channels. J Neurosci. 26 , 4835–4840.16672657 32. Shimada H. , Kusakizako T. , Nguyen THD , Nishizawa T. , Hino T. , Tominaga M. , Nureki O. (2020) The structure of lipid nanodisc-reconstituted TRPV3 reveals the gating mechanism. Nat Struct Mol Biol. 27 , 645–652.32572254 33. Kim S. E. , Patapoutian A. , Grandl J. (2013) Single residues in the outer pore of TRPV1 and TRPV3 Have Temperature-Dependent Conformations. PLOS One 8 , e59593 23555720 34. Yu A. C. (2021) Physical networks from entropy-driven non-covalent interactions. Nat Commun. 12 , 746; 10.1038/s41467-021-21024-7.33531475 35. Myers B. R. , Bohlen C. J. , Julius D. (2008) A yeast genetic screen reveals a critical role for the pore helix domain in TRP channel gating. Neuron 58 , 362–273.18466747 36. Grandl JH , Hu H , Bandell M , Bursulaya B , Schmidt M , Petrus M and Patapoutian A . (2008) Pore region of TRPV3 ion channel is specifically required for heat activation. Nat. Neurosci. 11 , 1007–1013.19160498 37. Priyakumar U. D. (2012) Role of hydrophobic core on the thermal stability of proteins – molecular dynamics simulations on a single point mutant of Sso7d abstract. J. Biomol. Struct. Dyn., 29 , 961–971.22292954 38. Van den Burg B , Dijkstra BW , Vriend G , Van der Vinne B , Venema G , Eijsink VG . (1994) Protein stabilization by hydrophobic interactions at the surface. Eur. J. Biochem., 220 , 981–985.8143751 39. Floyd RW . (1962), Algorithm-97 - Shortest Path. Commun Acm 5 , 345–345. 40. Kiehna S. E. , Waters M. L. (2003) Sequence dependence of b-hairpin structure: Comparison of a salt bridge and an aromatic interaction. Protein Science 12 , 2657–2667.14627727

PMC010xxxxxx/PMC10312933.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398216 10.21203/rs.3.rs-3006203/v1 10.21203/rs.3.rs-3006203 preprint 1 Article Supra-second interval timing in bipolar disorder: examining the role of disorder sub-type, mood, and medication status Müller Ewald Victόria A. *ah Trapp Nicholas T. *ah Sarrett McCall E. j Pace Benjamin D. a Wendt Linder i Richards Jenny G. g Gala Ilisa K. a Miller Jacob N. k Wessel Jan R. cdh Magnotta Vincent A. agh Wemmie John A. aefh Boes Aaron D. abh Parker Krystal L. ah^ a Department of Psychiatry, The University of Iowa, Iowa City, Iowa, United States of America b Department of Pediatrics, The University of Iowa, Iowa City, Iowa, United States of America c Department of Psychological & Brain sciences, The University of Iowa, Iowa City, Iowa, United States of America d Department of Neurology, The University of Iowa, Iowa City, Iowa, United States of America e Department of Molecular Physiology and Biophysics, The University of Iowa, Iowa City, Iowa, United States of America f Department of Neurosurgery, The University of Iowa, Iowa City, Iowa, United States of America g Department of Radiology, The University of Iowa, Iowa City, Iowa, United States of America h Iowa Neuroscience institute, The University of Iowa, Iowa City, Iowa, United States of America i Institute for Clinical and Translational Science, The University of Iowa, Iowa City, Iowa, United States of America j Department of Psychological and Brain sciences, Villanova University, Villanova, Pennsylvania, United States of America k St. Luke’s Hospital, Cedar Rapids, Iowa, United States of America * Both authors contributed equally to the development and publication of this work. Authors’ contributions: VAME – data acquisition, data analysis, data interpretation, creation of new software, manuscript drafting and revision. NTT – study conception, study design, data acquisition, data interpretation, and manuscript revision. MES – data analysis, data interpretation, creation of new software, manuscript revision. BDP – study design, data acquisition, data analysis, manuscript revision. LW – data analysis, data interpretation, and manuscript revision. JGR – study design, data acquisition, and manuscript revision. IKG – data analysis, data interpretation, and manuscript revision. JNM – study conception, study design, data acquisition, data interpretation, and manuscript revision. JRW – data analysis, data interpretation, creation of new software, and manuscript revision. VAM – study conception, study design, data acquisition, data interpretation, and manuscript revision. JAW – study conception, study design, data interpretation, manuscript revision. ADB – study conception, study design, data interpretation, and manuscript revision. KLP - study conception, study design, data interpretation, and manuscript revision. All authors revised and approved the final manuscript. ^ Corresponding author: Address: 200 Hawkins Drive W276GH, Iowa City, IA 52242-1057, krystal-parker@uiowa.edu, Phone: (319) 353-4554 02 6 2023 rs.3.rs-3006203https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs3006203v1.pdf Background: Widely reported by bipolar disorder (BD) patients, cognitive symptoms, including deficits in executive function, memory, attention, and timing are under-studied. Work suggests that individuals with BD show impairments in interval timing tasks, including supra-second, sub-second, and implicit motor timing compared to the neuronormative population. However, how time perception differs within individuals with BD based on BD sub-type (BDI vs II), mood, or antipsychotic medication-use has not been thoroughly investigated. The present work administered a supra-second interval timing task concurrent with electroencephalography (EEG) to patients with BD and a neuronormative comparison group. As this task is known to elicit frontal theta oscillations, signal from the frontal (Fz) lead was analyzed at rest and during the task. Results: Results suggest that individuals with BD show impairments in supra-second interval timing and reduced frontal theta power compared during the task to neuronormative controls. However, within BD sub-groups, neither time perception nor frontal theta differed in accordance with BD sub-type, mood, or antipsychotic medication use. Conclusions: This work suggests that BD sub-type, mood status or antipsychotic medication use does not alter timing profile or frontal theta activity. Together with previous work, these findings point to timing impairments in BD patients across a wide range of modalities and durations indicating that an altered ability to assess the passage of time may be a fundamental cognitive abnormality in BD. Cognition Bipolar disorder Medication status Depression Antipsychotic Bipolar Disorders Research Program of Excellence via the University of Iowa Roy J. Carver Charitable Trust and the University of Iowa Neuroscience InstituteBaszucki Brain Research FundNIH National Institute of Mental HealthR01MH113325 NIH National Institute of Drug AbuseR01DA052953 Roy J. Carver Charitable TrustRoy J. Carver ChairU.S. Department of Veterans Affairs Merit Review AwardU.S. Department of Veterans Affairs ==== Body pmcBackground Current treatments for bipolar disorder (BD) largely focus on mood symptoms (1–4). However, changes in cognitive functioning, including deficits in memory, executive function, attention, planning, and timing (5–8), are common and may even precede a formal BD diagnosis (9). Cognitive symptoms are reported by patients with bipolar I disorder (BDI) and bipolar II disorder (BDII), and are present even in a euthymic state (10). As these symptoms are widely reported and linked to lowered quality of life (11), studies of cognitive symptoms within BD are imperative to a comprehensive understanding of the disorder. Previous work has identified timing deficits as a cognitive abnormality consistently presented by BD patients (12). Using finger tapping, auditory temporal bisection, and single-cue delay eye blink conditioning, Bolbecker and colleagues showed sub-second interval timing impairments in BD (13–15). Using time production and estimation tasks, Bschor and colleagues showed supra-second interval timing deficits in BD ranging from durations of 7 to 109 seconds. Interval timing depends on diffuse neural networks including the cerebello-thalamo-cortical network and the cortico-striatal network (16, 17). In BD, abnormalities in nodes of these networks have been previously reported, including frontal cortex, thalamus, and cerebellum (18–22), providing a mechanistic explanation for the wide-ranging deficits in timing observed in the disorder. Previous work suggests that impairments in interval timing correlate with abnormal fronto-central theta oscillations in patients with SCZ (23, 24). Substantial genetic and symptomatic overlaps between SCZ and BD have been suggested (4, 25). However, it is unclear if the deficits in timing and frontal oscillations observed in SCZ extend to BD. While extensive work has compared time processing between individuals with BD and other clinical populations or neuronormative controls, there is a paucity of work examining which specific characteristics within BD are linked to timing deficits. The relationship between depressive symptoms and time-perception slowing is well established and referred to as depressive time dilation (26). Additionally, work suggests that manic patients also show alterations in time processing, although works differ in effect directionality (12, 27). However, is unknown if the timing deficits observed in BD are linked to mood status at the time of assessment, or if they are a stable characteristic present even in the absence of mood symptoms. Additionally, while some reports suggest that cognitive impairments in BD can improve in conjunction with mood-symptom treatment (28), other reports suggest that cognitive symptoms may worsen in conjunction with mood treatment (25, 28). It has been suggested that antipsychotic medications may impair measures of general intellectual functioning, working memory, and cognitive set-shifting (6), likely due to reductions in information processing speed. Indeed, although the relationship between different medication types, including antidepressants and stimulants, and cognitive functioning in BD have been studied, antipsychotic-use is the only medication related variable with consistently significant impacts on cognition (6). Finally, although it is established that individuals with BD show impairments in timing, work has not explored how this may differ between bipolar disorder sub-types. Given differences in cycling speed and manic episode strength between disorder sub-types (29), differences in time perception could be expected. However, work also suggests similar cognitive profiles between BDI and BDII (10), adding a layer of complexity to this debate. To address these questions, we administered a supra-second interval timing task (ITT) to participants with BD and neuronormative controls (CT), while simultaneously recording electroencephalographic (EEG) activity. We hypothesized that individuals with BD would show impaired supra-second ITT performance, in agreement with previous work. Because of previous work with SCZ patients, we further hypothesized that BD patients would show reduced frontal theta power compared to the CT group during the ITT. Finally, we assessed differences in ITT performance and frontal theta in BD depending on disorder sub-type, mood, or antipsychotic medication status. Methods Subjects Twenty-four participants (20 females, 4 males) with a DSM-IV diagnosis of BDI (16 subjects) or BDII (8 subjects) were recruited from the Iowa Longitudinal Database (Tables 1 and 2). Subjects had diagnoses confirmed by a board-certified psychiatrist at the University of Iowa Hospitals & Clinics. Medication status was stable for a minimum of 30 days prior to enrollment and was not altered for the present study (Table 3). Individuals who reported illicit drug use within 6 months of study commencement were excluded from participation. Mood was assessed via the Montgomery-Asberg depression rating scale (30). A depressed state was defined as a score greater than 10 (31). Six CT subjects were included as a neuronormative comparison group. CT subjects did not have a history of neuropsychiatric disorders. In accordance with federal and institutional guidelines, all procedures including informed consent were approved by the University of Iowa Institutional Review Board and are in accordance with the Declaration of Helsinki. Tasks Interval timing task Concurrent with EEG acquisition, participants performed a supra-second ITT. Participants completed the task sitting in front of a Dell 20” monitor with a 60 Hz refresh rate and 4096×2304 screen resolution. White times new roman size 40 text appeared on a black background in the middle of the screen. Participants received verbal instructions on how to perform the task from the experimenter and read the same set of instructions on the computer screen. All participants were instructed not to count time in their head. To start each trial, a number indicating the interval to be estimated by the participant (“3” for the short interval/SIT or “12” for the long interval/LIT) appeared on the screen. Participants pressed the space bar to start the trial and to indicate their judgement of the elapsed interval, thus ending the trial. Response accuracy feedback was given for every trial immediately following the button press. The experiment consisted of a total of 80 trials (40 SIT trials & 40 LIT trials) presented in pseudo-random order. Resting state Resting-state recordings were conducted before the ITT task, lasting 5 minutes. Participants sat in a chair, were instructed to keep their eyes open, look forward, and let their mind wander. Electroencephalography (EEG) EEG acquisition A BrainVision 64-channel active electrode system with Ag/AgCl electrodes was used to collect EEG (Morrisville, NC). A custom-made electrode cap was utilized, which included electrode placements that are not typical of the International 10–20 system (32). Electrodes PO3 and PO4 were substituted by electrodes I1 and I2 which flanked the Iz electrode. These were situated at the back of the head over the inion and were not analyzed as a part of the present experiment. At the beginning of the recordings, impedances were reduced using high viscosity electrode gel for active electrodes (EASYCAP, Munich, Germany). Impedance for all electrodes was kept at or below 15 kΩ for the duration of the recording. Data were acquired at 500 Hz and referenced online to Pz. EEG analyses Analyses focused on electrode Fz, as frontal theta oscillations are typically maximal at this site (33). Frequency bands were defined as: delta (1–4 Hz), theta (4–8 Hz), alpha (8–13 Hz), beta (13–30 Hz), and gamma (30–50 Hz). Data were preprocessed using custom MATLAB (MathWorks, Natick, MA) scripts based on EEGLAB (34) functions. Data were sequentially high-pass filtered at 1Hz, and low-pass filtered at 50Hz, the transition bandwidth was set to twice the cutoff frequency (−6 dB) for cutoff <= 1Hz and 25% cutoff frequency for cutoff > 8Hz. Trials containing nonstereotypic artifacts were removed manually, resulting in exclusion of 18% of trials on average. Continuous data were rereferenced offline to average voltage. Eyeblinks and saccades were removed using independent component analysis. Data were epoched as follows: for the interval timing task, data were epoched around the presentation of the timing cue. For the SIT, the epoch ranged from 1 second before cue presentation to 5 seconds after cue presentation. For the LIT, data were epoched from 1 second before cue presentation to 15 seconds after cue presentation. For resting-state analyses, continuous data were epoched into 20s intervals to maintain epochs at approximately the same size between the two tasks. Quantification of band power was conducted using the fast-Fourier transform method. Relative power at each frequency band was defined as the proportion of the overall spectral power distribution occupied by each frequency band, quantified using the MATLAB function trapz. Statistical analyses Participant characteristics Demographic characteristics were compared between individuals in the BDI, BDII, and CT groups using a Pearson’s Chi Square for categorical variables and a one-way ANOVA for continuous variables. Categorical variables were: sex, race, education and handedness. Age was the only continuous variable. Additionally, propensity scores were generated to assess whether age, sex, race, and years of education were associated with the probability of a participant being in the BD group vs. the CT group. Interval timing task performance and band power To assess performance on the ITT, participants’ time estimates for the SIT/LIT intervals were fit with Gaussian distributions using custom-written MATLAB routines. Timing accuracy and precision were estimated by calculating peak time and CV measures, respectively. The peak time index represents the accuracy of participants’ responses and was calculated using the best fit estimate of the Gaussian distribution. The CV index represents the precision of participants’ responses and was calculated by dividing the response standard deviation by peak time. T-tests were conducted in GraphPad Prism (San Diego, California) to statistically assess performance differences between groups. Statistical comparisons of power at each oscillation band were compared between groups in GraphPad Prism using t-tests. Multiple comparisons were corrected for using Tukey’s multiple comparisons test. Statistical outliers were defined as individuals with scores 2 standard deviations above/below their group mean and excluded from the analysis. Mean, standard error of the mean (SEM), and number of outliers excluded for each group are expressed as [GROUP NAME mean ± SEM (number of outliers excluded)]. Results Participant characteristics The demographic characteristics age, sex, race education and handedness did not significantly differ between individuals in the BDI, BDII, and CT groups (Table 1). Additionally, although the study population was heavily skewed towards the BD group, propensity scores did not provide strong evidence that participants in the BD vs. the CT groups substantially differed with regards to their age, race, sex, and years of education (Figure 1). Comparison between BD and CT groups Individuals with BD show impaired supra-second ITT performance compared to the CT group (Figure 2A–B). For SIT, individuals with BD show an over-estimation of the target duration compared to the CT group, as quantified by the peak time index (t(27) = 2.61, p = 0.0146 [BD 3.45 ± 0.0616 (1); CT 3.17 ± 0.0.0466 (0)]; Figure 2C [left]). Response distribution, quantified by the CV index, did not differ between groups (t(27) = 1.33, p = 0.192 [BD 0.2032 ± 0.00655 (1); CT 0.184 ± 0.00883 (0)]; Figure 2C [right]). For the LIT, peak response times did not differ between BD and CT groups (t(28) = 1.576, p = 0.1262 [BD 11.36 ± 0.1084 (0); CT 11.73 ± 0.1620 (0)]; Figure 2D[left]). However, individuals with BD showed significantly lower CV indices, indicating higher variability in response times compared to the CT group (t(27) = 3.345, p = 0.0024 [BD 0.198 ± 0.00778 (1); CT 0.1406 ± 0.161 (0)]; Figure 2D[right]). During the ITT, individuals with BD showed lower frontal theta power compared to the CT group (t(27) = 2.992, p = 0.0059 [BD 0.159 ± 0.0116 (1); CT 0.3000 ± 0.0842 (0)]; Figure 2E). No differences in power were detected between the BD and the CT groups during the ITT for any other frequency bands (Supplemental Figure 1). Although theta power values of a single CT subject are markedly higher than the remainder of the CT subjects, this data point was not excluded, as it does not fit the statistical outlier criteria as described in the methods section. To assess if differences in theta power between BD and CT groups were task-specific, resting-state data were analyzed (Figure 3). There were no significant differences in theta power between BD and CT groups (t(30) = 0.8343, p = 0.4107 [BD 0.147 ± 0.009901 (1); CT 0.248 ± 0.0488 (1)]) during rest. Comparisons within BD sub-groups Within our group of individuals with BD, we first assessed if BD disorder sub-type differentially affected supra-second interval timing ability and associated frontal theta power. Response curves suggest that individuals with BDI and BDII did not differ in their supra-second ITT performance (Figure 4A–B). Peak time and CV indices did not differ between groups for either the SIT (Peak time: t(22) = 0.02449, p = 0.9807 [BDI 3.45 ± 0.0760 (0); BDII 3.45 ± 0.153 (0)]; Figure 4C [left]; CV: t(22) = 0.4230, p = 0.6764 [BDI 0.198 ± 0.00879 (0); BDII 0.204 ± 0.0120 (0)]; Figure 4C [right]) or the LIT (Peak time: t(22) = 0.9181, p = 0.9277 [BDI 11.35 ± 0.130 (0); BDII 11.37 ± 0.206 (0)]; Figure 4D [left]; CV: t(22) = 0.3181, p = 0.7534 [BDI 0.192 ± 0.0109 (0); BDII 0.195 ± 0.0130 (0)]; Supplemental Figure 4D [right]) intervals. Additionally, theta power during the ITT did not differ between individuals with BDI vs. BDII (t(21) = 1.268, p = 0.2188, one BD outlier excluded, Figure 4E). Next, we assessed if ITT performance and associated frontal theta differed by mood status (i.e. depressed vs. euthymic) within the BD group. Response curves suggest that supra-second ITT performance does not differ between depressed vs. euthymic individuals (Figure 5A–B). Peak time and CV indices did not differ between groups for the SIT (Peak time: t(20) = 0.05827, p = 0.9541 [BD 3.502 ± 0.0725 (0); CT 3.510 ± 0.121 (0)]; CV: t(21) = 0.3629, p = 0.7203 [BD 0.2011 ± 0.00900 (1); CT 0.195 ± 0.0124 (0)]; Figure 5C) or the LIT (Peak time: t(21) = 1.333, p = 0.1969 [BD 11.30 ± 0.116 (0); CT 11.57 ± 0.186 (0)]; CV: t(21) = 0.3012, p = 0.7662 [BD 0.189 ± 0.0120 (0); CT 0.194 ± 0.00967 (0)]; Figure 5D) intervals. Frontal theta power also did not differ between groups as shown in Figure 5E (t(20) = 0.1963, p = 0.8463 [BD 0.165 ± 0.0216 (1); CT 0.160 ± 0.00999 (0)]). Finally, we assessed if ITT performance and associated frontal theta differed by antipsychotic medication-use within the BD group. Response curves suggest that supra-second ITT performance was not significantly associated with differences in antipsychotic medication status (Figure 6A–B). Peak time and CV indices did not differ between groups for the SIT (Peak time: t(20) = 0.1367, p = 0.8927 [BD 3.46 ± 0.0780 (0); CT 3.47 ± 0.0900 (0)]; CV: t(22) = 0.1525, p = 0.8802 [BD 0.199 ± 0.0103 (0); CT 0.201 ± 0.00898 (0)]; Figure 6C) or the LIT (Peak time: t(21) = 0.2647, p = 0.7938 [BD 11.29 ± 0.118 (1); CT 11.35 ± 0.198 (0)]; CV: t(21) = 0.02555, p = 0.9799 [BD 0.198 ± 0.00886 (1); CT 0.1987 ± 0.01428 (0)], Figure 6D) intervals. Frontal theta power during the ITT also did not differ between groups Figure 6E (t(21) = 1.284, p = 0.2133 [BD 0.171 ± 0.0130 (1); CT 0.140 ± 0.0220 (0)]). Discussion The objective of the present work was to assess supra-second ITT performance in individuals with BD. Specifically, we were interested in whether BD disorder sub-type, mood, or antipsychotic medication-use altered supra-second interval timing in our cohort of patients. Our results suggest that, although ITT performance and frontal theta were decreased in the BD group compared to the CT group, within BD sub-groups there were no differences in ITT performance or frontal theta power. Together with previous work indicating that individuals with BD show impairments in supra-second (12, 27), sub-second (13, 15), and implicit motor timing (14), our work suggest that an altered ability to assess the passage of time may be a fundamental cognitive abnormality in BD. Trait vs. state abnormalities in BD Although cognitive impairments are widely reported by BD patients (10), few reports have attempted to triangulate which cognitive impairments may be state vs. trait characteristics of BD. Because clinical characteristics vary between BDI and BDII, and because previous work suggests a link between depressed mood/antipsychotic medication-use and altered temporal processing, we assessed if these characteristics varied within our group of BD patients. First, our results indicate that supra-second timing performance is not altered as a function of BD disorder sub-type. The extent to which cognitive profiles differ between BDI and BDII is debated in the literature. While some work suggests that BDI presents with more significant cognitive impairments (8), other studies suggest similar levels of cognitive impairments between the two sub-groups (10). Indeed, recent work suggests that BDII patients show impaired performance in neuropsychological tests including attention/working memory, executive function, verbal and visual memory, and motor speed compared to neuronormative controls (35). Our results add to this body of literature, indicating that cognitive impairments in the supra-second interval timing domain, do not differ by BD disorder sub-type. Additionally, our work suggests supra-second interval timing abilities do not differ between depressed and euthymic BD patients. The lack of distinction between these two groups is interesting given the well-established link between depression and a slowing of time perception in the supra-second domain (36). However, previous work suggests that 40–60% of euthymic BD patients may present with some sort of neurocognitive impairment (8). Indeed, work by Martino and colleagues (37) assessed six cognitive domains (attention, verbal memory, language, psychomotor speed, executive function, and facial emotional recognition) in BD and found that 62% of euthymic BD patients showed cognitive impairments, with 40% of patients showing 1 or 2 impaired domains, and 22% of patients showing impairments in 3 to 5 domains. Our findings thus suggest an additional domain – supra-second interval timing - where BD patients show impairments even when in the euthymic state, adding to the growing literature indicating that cognitive markers are fundamental characteristics of the disorder. Finally, our work suggests that antipsychotic medication-use does not alter supra-second timing in BD patients. Past work suggests a negative association between antipsychotic medication-use and IQ in BD (38). Specifically, tests of working memory, set-shifting, and response initiation/inhibition are negatively affected by antipsychotic medication-use (6). However, not all cognitive measures in BD are affected by antipsychotic medication-use, including measures of response planning and general working memory. In the context of this literature, our negative findings concerning antipsychotic use and interval timing are surprising. However, the dependence of timing abilities on working memory and response planning, two cognitive features not altered by antipsychotic use, could explain these findings. Another possibility is that participants were on low antipsychotic doses. Because antipsychotic dose is related to the degree of cognitive impairment, this could explain the lack of group difference. However, dose information was not collected, thus this analysis cannot be conducted leaving space for future work. Frontal theta during the ITT The present work identified abnormalities in frontal (Fz) theta oscillations during the ITT in BD patients compared to CT participants. Previous work suggests that ITT performance increases frontal theta power compared to rest. This pattern of activity was detected for the CT group, where visual inspection of ITT vs. resting state graphs suggests that frontal theta power was higher during the ITT (Figures 2E vs. 3A). However, this pattern was not detected for the BD group, where average power stayed approximately the same during task and rest. This suggests a failure in the mechanisms subserving time perception in the BD group, expressed electrophysiologically as unaltered frontal theta power and behaviorally as impaired supra-second interval timing. Previous work suggests that, compared to neuronormative controls, patients with SCZ show abnormal frontal low frequency (delta + theta) activity during the ITT (23). Our work suggests that the relationship between abnormal frontal theta and impaired ITT performance may not be a characteristic of SCZ specifically, extending to BD as well. If this characteristic extends outside of the schizoaffective domain, however, remains to be determined. Finally, in SCZ patients, work suggest that abnormalities in theta power in the 500ms window following timing-cue presentation is related to abnormal supra-second ITT performance (23). Although the primary objective of the present work was to analyze ITT performance and theta power within BD sub-groups, not between BD and CT groups, because of this previous SCZ work, secondary analyses were added to identify specific epochs of altered theta power during the task. These analyses were time-locked to cue presentation and response. Results suggest that oscillatory abnormalities in BD were not time-locked to the post-cue interval as they were in SCZ (Supplemental Figures 2B-C and 3B-C). One surprising finding from the present dataset is that individuals with BD showed lower theta power surrounding the response for the long interval (Supplemental Figure 3F), but not the short interval (Supplemental Figure 2F). This parallels performance data where BD patients show altered precision estimates for the long interval (Figure 2D [right]), but not the short interval (Figure 2C [right]). These results could indicate that frontal theta power is more closely linked with precision than response accuracy. However, further work is necessary to substantiate this claim. Pathophysiology of bipolar disorder Using timing task performance to triangulate single regions which may be abnormal in BD presents a challenge, as the neuroanatomy of time processing is famously diffuse (17) involving the coordinated functioning of multiple brain regions and neurotransmitter systems. One mechanism underlying the altered ITT performance observed in the present work may be the abnormal functioning of the dopamine system in individuals with BD. Indeed, the dopamine hypothesis of BD, which proposes intrinsic dysregulation of dopamine receptor transporter homeostasis (3, 39), is widely used to explain the pathophysiology of this disorder. Additionally, in other disorders where dopaminergic pathway function is altered, such as SCZ, Parkinson’s, or Huntington’s disease, abnormalities in temporal processing have also been reported (17). However, the absence of an effect of antipsychotic treatment on ITT performance weighs against the interpretation of timing deficits being caused by dopaminergic system abnormalities, as this medication class primarily targets the dopamine system. Another possible mechanism subserving the ITT performance and frontal theta deficits identified in the present work is the well-characterized frontal cortical abnormalities observed in individuals with BD including reductions in cortical grey matter (18, 40). Indeed, compromised frontal cortical activity has been linked to abnormalities in supra-second interval timing (17). This suggests a suggests a failure in the frontal mechanisms subserving time perception in BD patients, expressed electrophysiologically as unaltered frontal theta power and behaviorally as impaired supra-second interval timing. Limitations The present sample is skewed towards BD patients, as the CT group comprises 20% (n = 6) of the total study population while the BD group comprises 80% (n = 24) of the population. Because of this, the precision of estimates where comparisons between CT and BD groups are made are limited. However, one indicator of the reliability of these findings, is that they are in agreement with previous work which has identified differences in supra-second ITT performance between CT and BD groups (12, 27). Finally, we were unable to assess how mania alters supra-second interval timing performance, as none of our BD patients were in a manic state. This question is of particular interest as results are not consistent within the literature: while some studies suggest that manic patients under-estimate supra-second intervals (12), other suggests that manic patients over-estimate such intervals (27). However, this remains an open question for future work. Conclusions Although previous work has established timing deficits in BD, it is unclear if these cognitive abnormalities are due to secondary characteristics associated with BD, such as medication and mood, or if they are a fundamental characteristic of the disorder. In this study, we assessed whether BD sub-type (BDI vs. BDII), mood, or antipsychotic medication-use differentially affected BD patients’ ITT performance and associated frontal theta. Results suggest that ITT performance and frontal theta do not differ between BD sub-types, mood, or antipsychotic medication status. Together with previous work assessing interval timing in BD, these results suggest that an altered ability to assess the passage of time may be a trait cognitive abnormality in BD. Acknowledgements: The authors would like to acknowledge Laren Garrett for her excellent technical assistance. Funding: This work was funded by the following sponsors: Bipolar Disorders Research Program of Excellence via the University of Iowa Roy J. Carver Charitable Trust and the University of Iowa Neuroscience Institute. K.L.P. was supported by the Baszucki Brain Research Fund Grant Award. J.A.W. was supported by NIH National Institute of Mental Health grant R01MH113325, NIH National Institute of Drug Abuse grant R01DA052953, the Roy J. Carver Charitable Trust, the Roy J. Carver Chair, a U.S. Department of Veterans Affairs Merit Review Award, and the U.S. Department of Veterans Affairs. Availability of data and material: The datasets analyzed for the current study will be made available on an individual basis upon reasonable request to the corresponding author. Figure 1. Propensity histogram for neuronormative control and bipolar patient groups. Distribution of control (grey) and bipolar (teal) propensity scores along X axis indicates similar distribution of demographic variables for both groups. Nearly all participants have an estimated probability of being in the BD group that is greater than 0.5 because the majority of our sample belongs to the BD group. This graph indicates that while our groups may not be perfectly balanced, the degree of imbalance between them is not indicative of significant sampling bias. Figure 2. Individuals with bipolar disorder show impairments in supra-second interval timing and abnormal frontal theta compared to neuronormative controls. A. Schematic diagram of supra-second interval timing task. Trials begin when participants are shown a 3s or a 12s timing cue. Participants press the spacebar to indicate their estimation of the target interval. B. Response distribution for neuronormative controls vs. individuals with bipolar disorder. C. Individuals with bipolar disorder over-estimate the short interval compared to controls [left]. No differences in response distribution were detected [right]. D. Individuals with bipolar disorder do not differ from controls in estimation of the long interval duration [left], however, individuals with bipolar disorder have a significantly wider response distribution compared to controls [right]. E. Individuals with bipolar disorder show lower theta power compared to individuals in the neuronormative control group during the supra-second interval timing task. Mean and standard error of the mean plotted in bar graphs. Dots represent values from individual subjects. * p < 0.05 Figure 3. Individuals with bipolar disorder and neuronormative controls do not differ in frontal theta power at rest. A. To assess resting-state differences in theta power between bipolar disorder and neuronormative control groups resting-state data were analyzed. No differences in resting-state theta power were identified between neuronormative control and bipolar groups. Mean and standard error of the mean plotted in bar graphs. Dots represent values from individual subjects. * p < 0.05 Figure 4. Interval timing performance and frontal theta power do not differ as a function of bipolar disorder sub-type. A. Schematic diagram of supra-second interval timing task. Trials begin when participants are shown a 3s or a 12s timing cue. Participants press a button to indicate their estimation of the target interval. B. Response distribution for individuals with bipolar I or bipolar II disorder. C-D. Groups do not differ in time estimation for the short [C] or the long [D] intervals. E. Frontal theta power during the ITT did not differ between groups. Mean and standard error of the mean plotted in bar graphs. Dots represent values from individual subjects. * p < 0.05 Figure 5. Interval timing performance and frontal theta power do not differ as a function of mood. A. To assess task-wide differences in oscillatory activity data from the whole interval-timing task were analyzed. B. Response distribution for individuals with bipolar disorder who were either euthymic or depressed at the time of data collection. C-D. Groups do not differ in time estimation for the short [C] or the long [D] intervals. E. Frontal theta power during the ITT did not differ between groups. * p < 0.05 Figure 6. Interval timing performance and frontal theta power do not differ as a function of antipsychotic medication status. A. Response distribution for individuals with bipolar disorder divided by anti-psychotic medication status. B-D. Groups do not differ in time estimation for the short [C] or the long [D] intervals. E. Frontal theta power during the ITT did not differ between groups. Mean and standard error of the mean plotted in bar graphs. Dots represent values from individual subjects. * p < 0.05 TABLE 1. Participant Demographics Baseline Characteristic Controls (n=6) Bipolar type I (n=16) Bipolar type II (n=8) p-value Age 0.948  Mean (SD) 37.2 (10.3) 37.2 (13.5) 35.5 (12.1) Sex 0.212  Female 3 (50%) 13 (81.25%) 7 (87.50%)  Male 3 (50%) 3 (18.75%) 1 (12.50%) Race 0.471  White 4 (66.67%) 12 (75%) 6 (75%)  Black/African American 0 (0%) 3 (18.75%) 0 (0%)  Asian 1 (16.67%) 0 (0%) 1 (12.50%)  Other 1 (16.67%) 1 (6.25%) 1 (12.50%) Education 0.138  No High School 0 (0%) 2 (12.50%) 0 (0%)  High School 0 (0%) 5 (31.25%) 3 (37.50%)  Associate’s/Bachelor’s 3 (50%) 7 (43.75%) 5 (62.50%)  Post-Graduate 3 (50%) 2 (12.50%) 0 (0%) Handedness 0.126  Right 5 (83.33%) 16 (100%) 8 (100%)  Left 1 (16.67%) 0 (0%) 0 (0%) A one-way ANOVA was used to assess differences in continuous variables, while a Chi-square analysis was used to assess differences between categorical variables. TABLE 2. Comorbidities reported at enrollment for bipolar group Comorbidity n (%) Generalized Anxiety Disorder 5 (20.8%) Post-Traumatic Stress Disorder 4 (16.67%) Panic Disorder 2 (8.33%) Borderline Personality Disorder 2 (8.33%) Attention-Deficit/Hyperactivity Disorder 2 (8.33%) Migraine 1 (4.17%) Fibromyalgia 1 (4.17%) Other Medical Condition 5 (20.8%) TABLE 3. Medications reported at enrollment for bipolar group Type of Medication n (%) Any antidepressant 18 (75%) SSRI 4 (16.7%) SNRI 6 (25%) Atypical antidepressant 11 (45.8%) Atypical antipsychotic 14 (58.3%) Lithium 6 (25%) Benzodiazepine 10 (41.67%) Stimulant 4 (16.7%) Anticonvulsant 11 (45.8%) Opioid 2 (8.3%) SSRI, selective serotonin reuptake inhibitor SNRI, selective norepinephrine reuptake inhibitor Competing interests: The authors declare that they have no competing interests. Declarations Ethics approval and consent to participate: In accordance with federal and institutional guidelines, all procedures including informed consent were approved by the University of Iowa Institutional Review Board and are in accordance with the Declaration of Helsinki. Before study commencement, researchers reviewed informed consent materials with participants clearly explaining risks and benefits involved in study participation. Consent for publication: Not applicable Supplementary Files This is a list of supplementary les associated with this preprint. Click to download. BDandtimimgSupplementf1.pdf SF1.tif SF2.tif SF3.tif BDandTimingTablesf1.pdf ==== Refs Works Cited 1. Schaffer A , Isometsa ET , Tondo L , Moreno DH , Sinyor M , Kessing LV , Epidemiology, neurobiology and pharmacological interventions related to suicide deaths and suicide attempts in bipolar disorder: Part I of a report of the International Society for Bipolar Disorders Task Force on Suicide in Bipolar Disorder. Aust N Z J Psychiatry. 2015;49 (9 ):785–802.26185269 2. Scaini G , Valvassori SS , Diaz AP , Lima CN , Benevenuto D , Fries GR , Neurobiology of bipolar disorders: a review of genetic components, signaling pathways, biochemical changes, and neuroimaging findings. Braz J Psychiatry. 2020;42 (5 ):536–51.32267339 3. Ashok AH , Marques TR , Jauhar S , Nour MM , Goodwin GM , Young AH , The dopamine hypothesis of bipolar affective disorder: the state of the art and implications for treatment. Mol Psychiatry. 2017;22 (5 ):666–79.28289283 4. Murray RM , Sham P , Van Os J , Zanelli J , Cannon M , McDonald C . A developmental model for similarities and dissimilarities between schizophrenia and bipolar disorder. Schizophr Res. 2004;71 (2–3 ):405–16.15474912 5. Zhou FC , Wang YY , Zheng W , Ungvari GS , Ng CH , Yuan Z , Prospective memory in bipolar disorder: A meta-analysis. Psychiatry Res. 2018;259 :184–90.29055798 6. Frangou S. The Maudsley Bipolar Disorder Project. Epilepsia 2005;46 :19–25.15938707 7. Goldberg TE . Some fairly obvious distinctions between schizophrenia and bipolar disorder. Schizophrenia Research 1999;39 :127–32.10507523 8. Sole B , Bonnin CM , Torrent C , Martinez-Aran A , Popovic D , Tabares-Seisdedos R , Neurocognitive impairment across the bipolar spectrum. CNS Neurosci Ther. 2012;18 (3 ):194–200.22128808 9. Hinrichs KH , Easter RE , Angers K , Pester B , Lai Z , Marshall DF , Influence of cognitive reserve on neuropsychological functioning in bipolar disorder: Findings from a 5-year longitudinal study. Bipolar Disord. 2017;19 (1 ):50–9.28263040 10. King S , Stone JM , Cleare A , Young AH . A systematic review on neuropsychological function in bipolar disorders type I and II and subthreshold bipolar disorders-something to think about. CNS Spectr. 2019;24 (1 ):127–43.30859934 11. Brissos SVD V. ; Kapczinski F. Cognitive performance and quality of life in bipolar disorder Can J Psychiatry 2008;53 (8 ):517–24.18801213 12. Mahlberg R , Kienast T , Bschor T , Adli M . Evaluation of time memory in acutely depressed patients, manic patients, and healthy controls using a time reproduction task. Eur Psychiatry. 2008;23 (6 ):430–3.18515048 13. Bolbecker AR , Westfall DR , Howell JM , Lackner RJ , Carroll CA , O’Donnell BF , Increased timing variability in schizophrenia and bipolar disorder. PLoS One. 2014;9 (5 ):e97964.24848559 14. Bolbecker AR , Hong SL , Kent JS , Forsyth JK , Klaunig MJ , Lazar EK , Paced finger-tapping abnormalities in bipolar disorder indicate timing dysfunction. Bipolar Disord. 2011;13 (1 ):99–110.21320257 15. Bolbecker ARM C. ; Johannesen J. K. ; Edwards C. R. ; O’Donnell B. F. ; Shekhar A. ; Nurnberger J. I. ; Steinmetz J. E. ; Hetrick W. P. Eyeblink conditioning anomalities in bipolar disorder suggest cerebellar dysfunction. Bipolar Disord. 2009;11 (1 ):19–32.19133963 16. Timing Grondin S. and time perception: a review of recent behavioral and neuroscience findings and theoretical directions. Atten Percept Psychophys. 2010;72 (3 ):561–82.20348562 17. Buhusi CV , Meck WH . What makes us tick? Functional and neural mechanisms of interval timing. Nat Rev Neurosci. 2005;6 (10 ):755–65.16163383 18. Moorhead TW , McKirdy J , Sussmann JE , Hall J , Lawrie SM , Johnstone EC , Progressive gray matter loss in patients with bipolar disorder. Biol Psychiatry. 2007;62 (8 ):894–900.17617385 19. Shaffer JJ Jr ., Willour V , Fiedorowicz JG , Christensen GE , Long , Johnson CP , Distinct patterns of altered quantitative T1rho and functional BOLD response associated with history of suicide attempts in bipolar disorder. Brain Imaging Behav. 2022;16 (2 ):820–33.34601647 20. Shinn AK . Aberrant cerebellar connectivity in bipolar diorder with psychosis. Biological Psychiatry Cognitive Neuroscience Neuroimaging. 2017. 21. Soontornniyomkij B , Everall IP , Chana G , Tsuang MT , Achim CL , Soontornniyomkij V . Tyrosine kinase B protein expression is reduced in the cerebellum of patients with bipolar disorder. J Affect Disord. 2011;133 (3 ):646–54.21612826 22. Hibar DP , Westlye LT , van Erp TG , Rasmussen J , Leonardo CD , Faskowitz J , Subcortical volumetric abnormalities in bipolar disorder. Mol Psychiatry. 2016;21 (12 ):1710–6.26857596 23. Singh A , Trapp NT , De Corte B , Cao S , Kingyon J , Boes AD , Cerebellar Theta Frequency Transcranial Pulsed Stimulation Increases Frontal Theta Oscillations in Patients with Schizophrenia. Cerebellum. 2019;18 (3 ):489–99.30825131 24. Parker KL , Kim YC , Kelley RM , Nessler AJ , Chen KH , Muller-Ewald VA , Delta-frequency stimulation of cerebellar projections can compensate for schizophrenia-related medial frontal dysfunction. Mol Psychiatry. 2017;22 (5 ):647–55.28348382 25. Tham AE K. ; Mathe A. A. ; Johnson L. ; Olsson E. ; Aberg-Wistedt A. Impaired Neuropsychological Performance in Euthymic Patients with Recurring Mood Disorders. Journal of Clinical Psychiatry 1996;58 (1 ):26–9. 26. Lewis A. The Experience of Time in Mental Disorder. Proc R Soc Med. 1932;25 (5 ):611–20.19988601 27. Bschor TI M. ; Bauer M. ; Lewitzka U. ; Skerstupeit M. ; Muller-Oerlinghausen B. ; Baethge C. Time experience and time judgement in major depression, mania and healthy subjects. A controlled study of 93 subjects. Acta Psychiatr Scand. 2004;109 :222–9.14984395 28. Clark L , Iversen SD , Goodwin GM . Sustained attention deficit in bipolar disorder. Br J Psychiatry. 2002;180 :313–9.11925353 29. American Psychiatric Association., American Psychiatric Association. DSM-5 Task Force. Diagnostic and statistical manual of mental disorders : DSM-5. 5th ed. Washington, D.C.: American Psychiatric Association; 2013. xliv , 947 p. p. 30. Quilty LC , Robinson JJ , Rolland JP , De Fruyt F , Rouillon F , Bagby RM . The structure of the Montgomery–Åsberg depression rating scale over the course of treatment for depression. International Journal of Methods in Psychiatric Research. 2013;22 (3 ):175–84.24038301 31. Montgomery SA , Asberg M . A new depression scale designed to be sensitive to change. Br J Psychiatry. 1979;134 :382–9.444788 32. Klem GH , Lüders HO , Jasper HH , Elger C . The ten-twenty electrode system of the International Federation. The International Federation of Clinical Neurophysiology. Electroencephalography and clinical neurophysiology Supplement. 1999;52 :3–6. 33. Shinomia SU Y. , Nagata K. , Takahashi N. ; Inoue R. Frontal midline theta rhythm: differentiating the physiological theta rhythm from the abnormal discharge. Clin Electroencephalography. 1994;25 (1 ):30–5. 34. Delorme A , Makeig S . EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J Neurosci Methods. 2004;134 (1 ):9–21.15102499 35. Andersson S , Barder HE , Hellvin T , Lovdahl H , Malt UF . Neuropsychological and electrophysiological indices of neurocognitive dysfunction in bipolar II disorder. Bipolar Disord. 2008;10 (8 ):888–99.19594504 36. Kent L , Van Doorn G , Klein B . Time dilation and acceleration in depression. Acta Psychol (Amst). 2019;194 :77–86.30798221 37. Martino DJ , Strejilevich SA , Scapola M , Igoa A , Marengo E , Ais ED , Heterogeneity in cognitive functioning among patients with bipolar disorder. J Affect Disord. 2008;109 (1–2 ):149–56.18234352 38. Abrams R , Redfield J , Taylor MA . Cognitive dysfunction in schizophrenia, affective disorder and organic brain disease. Br J Psychiatry. 1981;139 :190–4.7317699 39. Wittenborn JR . Deductive approaches to the catecholamine hypothesis of affective disorders. Journal of Nervous and Mental Disease. 1974;158 (5 ):320–4.4823932 40. Michel CM , Koenig T . EEG microstates as a tool for studying the temporal dynamics of whole-brain neuronal networks: A review. Neuroimage. 2018;180 (Pt B ):577–93.29196270

PMC010xxxxxx/PMC10312940.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398424 10.21203/rs.3.rs-3012879/v1 10.21203/rs.3.rs-3012879 preprint 1 Article Bayesian estimation of gene constraint from an evolutionary model with gene features http://orcid.org/0000-0002-6509-9879 Zeng Tony 1*† http://orcid.org/0000-0002-3199-1447 Spence Jeffrey P. 1*† http://orcid.org/0000-0002-1060-2844 Mostafavi Hakhamanesh 1 http://orcid.org/0000-0002-8828-5236 Pritchard Jonathan K. 12† 1 Department of Genetics, Stanford University, Stanford CA 2 Department of Biology, Stanford University, Stanford CA * Equal contribution † Correspondence to: tkzeng@stanford.edu, jspence@stanford.edu, pritch@stanford.edu 13 6 2023 rs.3.rs-3012879https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs3012879v1.pdf Measures of selective constraint on genes have been used for many applications including clinical interpretation of rare coding variants, disease gene discovery, and studies of genome evolution. However, widely-used metrics are severely underpowered at detecting constraint for the shortest ~25% of genes, potentially causing important pathogenic mutations to be overlooked. We developed a framework combining a population genetics model with machine learning on gene features to enable accurate inference of an interpretable constraint metric, shet. Our estimates outperform existing metrics for prioritizing genes important for cell essentiality, human disease, and other phenotypes, especially for short genes. Our new estimates of selective constraint should have wide utility for characterizing genes relevant to human disease. Finally, our inference framework, GeneBayes, provides a flexible platform that can improve estimation of many gene-level properties, such as rare variant burden or gene expression differences. ==== Body pmc1 Introduction Identifying the genes important for disease and fitness is a central goal in human genetics. One particularly useful measure of importance is how much natural selection constrains a gene [1–4]. Constraint has been used to prioritize de novo and rare variants for clinical followup [5, 6], predict the toxicity of drugs [7], link GWAS hits to genes [8], and characterize transcriptional regulation [9, 10], among many other applications. To estimate the amount of constraint on a gene, several metrics have been developed using loss-of-function variants (LOFs), such as protein truncating or splice disrupting variants. If a gene is important, then natural selection will act to remove LOFs from the population. Several metrics of gene importance have been developed based on this intuition to take advantage of large exome sequencing studies. In one line of research, the number of observed unique LOFs is compared to the expected number under a model of no selective constraint. This approach has led to the widely-used metrics pLI [11] and LOEUF [12]. While pLI and LOEUF have proved useful for identifying genes intolerant to LOF mutations, they have important limitations [3]. First, they are uninterpretable in that they are only loosely related to the fitness consequences of LOFs. Their relationship with natural selection depends on the study’s sample size and other technical factors [3]. Second, they are not based on an explicit population genetics model so it is impossible to compare a given value of pLI or LOEUF to the strength of selection estimated for variants other than LOFs [3, 4]. Another line of research has solved these issues of interpretability by estimating the fitness reduction for heterozygous carriers of an LOF in any given gene [1, 2, 4]. Throughout, we will adopt the notation of Cassa and colleagues and refer to this reduction in fitness as shet [1, 2], although the same population genetic quantity has been referred to as hs [4, 13]. In [1], a deterministic approximation was used to estimate shet, which was relaxed to incorporate the effects of genetic drift in [2]. This model was subsequently extended by Agarwal and colleagues to include the X chromosome and applied to a larger dataset, with a focus on the interpretability of shet [4]. A major issue for most previous methods is that thousands of genes have few expected unique LOFs under neutrality, as they have short protein-coding sequences. For example, there are >5,000 genes that cannot be called as constrained by LOEUF, as they have too few expected unique LOFs to fall under the recommended LOEUF cutoff of 0.35 [14]. This problem is not limited to LOEUF, however, and all of these methods are severely underpowered to detect selection for this ~25% of genes. Here, we present an approach that can accurately estimate shet even for genes with few expected LOFs, while maintaining the interpretability of previous population-genetics based estimates [1, 2, 4]. Our approach has two main technical innovations. First, we use a novel population genetics model of LOF allele frequencies. Previous methods have either only modeled the number of unique LOFs, throwing away frequency information [11,12,15], or considered the sum of LOF frequencies across the gene [1,2,4], an approach that is not robust to misannotated LOFs. In contrast, we model the frequencies of individual LOF variants, allowing us to not only use the information in such frequencies but also to model the possibility that any given LOF variant has been misannotated, making our estimates more robust. Our approach uses new computational machinery, described in a companion paper [16], to accurately obtain the likelihood of observing an LOF at a given frequency without resorting to simulation [2, 4] or deterministic approximations [1]. Second, our approach uses thousands of gene features, including gene expression patterns, protein structure information, and evolutionary constraint, to improve estimates for genes with few expected LOFs. By using these features, we can share information across similar genes. Intuitively, this allows us to improve estimates for genes with few expected LOFs by leveraging information from genes with similar features that do have sufficient LOF data. Adopting a similar approach, a recent preprint [15] used gene features in a deep learning model to improve estimation of constraint for genes with few expected LOFs, but did not use an explicit population genetics model, resulting in the same issues with interpretability faced by pLI and LOEUF. We applied our method to a large exome sequencing cohort [12]. Our estimates of shet are substantially more predictive than previous metrics at prioritizing essential and disease-associated genes. We also interrogated the relationship between gene features and natural selection, finding that evolutionary conservation, protein structure, and expression patterns are more predictive of shet than co-expression and protein-protein interaction networks. Expression patterns in the brain and expression patterns during development are particularly predictive of shet. Finally, we use shet to highlight differences in selection on different categories of genes and consider shet in the context of selection on variants beyond LOFs. Our approach, GeneBayes, is extremely flexible and can be applied to improve estimation of numerous gene properties beyond shet. Our implementation is available at https://github.com/tkzeng/GeneBayes. 2 Results 2.1 Model Overview Using LOF data to infer gene constraint is challenging for genes with few expected LOFs, with metrics like LOEUF considering almost all such genes to be unconstrained (Figures 1A,B). We hypothesized that it would be possible to improve estimation using auxiliary information that may be predictive of LOF constraint, including gene expression patterns across tissues, protein structure, and evolutionary conservation. Intuitively, genes with similar features should have similar levels of constraint. By pooling information across groups of similar genes, constraint estimated for genes with sufficient LOF data may help improve estimation for underpowered genes. However, while the frequencies of LOFs can be related to shet through models from population genetics [1, 2, 4], we lack an understanding of how other gene features relate to constraint a priori. To address this problem, we developed a flexible empirical Bayes framework, GeneBayes, that learns the relationship between gene features and shet (Figure 1C). Our model consists of two main components. First, we model the prior on shet for each gene as a function of its gene features (Figure 1C, left). Specifically, we train gradient-boosted trees using NGBoost [17] to predict the parameters of each gene’s prior distribution from its features. Our gene features include gene expression levels, Gene Ontology terms, conservation across species, neural network embeddings of protein sequences, gene regulatory features, co-expression and protein-protein interaction features, sub-cellular localization, and intolerance to missense mutations (see Methods and Supplementary Note C for a full list). Second, we use a model from population genetics to relate shet to the observed LOF data (Figure 1C, right). This model allows us to fit the gradient-boosted trees for the prior by maximizing the likelihood of the LOF data. Specifically, we use the discrete-time Wright Fisher model with genic selection, a standard model in population genetics that accounts for mutation and genetic drift [13, 18]. In our model, shet is the reduction in fitness per copy of an LOF, and we infer shet while keeping the mutation rates and demography fixed to values taken from the literature (Supplementary Note B). Likelihoods are computed using new methods described in a companion paper [16]. Previous methods use either the number of unique LOFs or the sum of the frequencies of all LOFs in a gene, but we model the frequency of each individual LOF variant. We used LOF frequencies from the gnomAD consortium, which consists of exome sequences from ~125,000 individuals for 18,563 genes after filtering. Combining these two components—the learned priors and the likelihood of the LOF data— we obtained posterior distributions over shet for every gene. Throughout, we use the posterior mean value of shet for each gene as a point estimate. See Methods for more details and Supplementary Table 2 for estimates of shet. 2.2 Population genetics model and gene features both affect the estimation of shet First, we explored how LOF frequency and mutation rate relate to shet in our population genetics model (Figure 2A). Invariant sites with high mutation rates are indicative of strong selection shet>10-2, consistent with [19], while such sites with low mutation rates are consistent with essentially any value of shet for the demographic model considered here. Regardless of mutation rate, singletons are consistent with most values of shet but can rule out extremely strong selection, and variants observed at a frequency of >10% rule out even moderately strong selection shet>10-3. To assess how informative gene features are about shet, we trained our model on a subset of genes and evaluated the model on held-out genes (Figure 2B, Methods). We computed the Spearman correlation between shet estimates from the prior and shet estimated from the LOF data only. The correlation is high and comparable between train and test sets (Spearman ρ=0.83 and 0.78 respectively), indicating the gene features alone are highly predictive of shet and that this is not a consequence of overfitting. To further characterize the impact of features on our estimates of shet, we removed all features from our model and recalculated posterior distributions (Figure 2C). For most genes, posteriors are substantially more concentrated when using gene features. Next, we compared our estimates of shet using GeneBayes to LOEUF and to selection coefficients estimated by [4] (Figure 2D). To facilitate comparison, we use the posterior modes of shet reported in [4] as point estimates, but we note that [4] emphasizes the value of using full posterior distributions. While the correlation between our estimates is high for genes with sufficient LOFs (for genes with more LOFs than the median, Spearman ρ with LOEUF=0.94;ρ with shet from [4] = 0.88), it is lower for genes with few expected LOFs (for genes with fewer LOFs than the median, Spearman ρ with LOEUF=0.71;ρ with shet from [4] = 0.71). We further explored the reduced correlations for genes with few expected LOFs. For example, TBC1D3 and PLN have few expected LOFs, and their likelihoods are consistent with any level of constraint (Figure 2E). Due to the high degree of uncertainty, LOEUF considers both genes to be unconstrained, while the shet point estimates from [4] err in the other direction and consider both genes to be constrained (Figure 2D). This uncertainty arises from use of the LOF data alone, and is captured by the wide posterior distributions for the shet estimates from [4]. In contrast, by using gene features, our posterior distributions of shet indicate that PLN is strongly constrained but TBC1D3 is not, consistent with the observation that heterozygous LOFs in PLN cause severe cardiac dilation and heart failure [20]. In contrast to estimates of shet, LOEUF further ignores information about allele frequencies by considering only the number of unique LOFs, resulting in a loss of information. For example, AARD and TWIST1 have almost the same numbers of observed and expected unique LOFs, so LOEUF is similar for both (LOEUF = 1.1 and 1.06 respectively). However, while TWIST1’s observed LOF is present in only 1 of 246,192 alleles, AARD’s is ~40× more frequent. Consequently, the likelihood rules out the possibility of strong constraint at AARD (Figure 2F), causing the two genes to differ in their estimated selection coefficients (Figure 2D). In contrast, TWIST1 has a posterior mean shet of 0.11 when using gene features, indicating very strong selection. Consistent with this, TWIST1 is a transcription factor critical for specification of the cranial mesoderm, and heterozygous LOFs in the gene are associated with Saethre-Chotzen syndrome, a disorder characterized by congenital skull and limb abnormalities [21, 22]. Besides PLN and TWIST1, many genes are considered constrained by shet but not by LOEUF, which is designed to be highly conservative. In Table 1, we list 15 examples with shet>0.1 and LOEUF>0.5, selected based on their clinical significance and prominence in the literature (Methods). One notable example is a set of 16 ribosomal protein genes for which heterozygous disruption causes Diamond-Blackfan anemia—a rare genetic disorder characterized by an inability to produce red blood cells [23] (Supplementary Table 1). All are considered strongly constrained by shet (minimum shet=0.26). In contrast, only 6 are considered constrained by LOEUF (LOEUF<0.35), as many of these genes have few expected unique LOFs. 2.3 Utility of shet in prioritizing phenotypically important genes To assess the accuracy of our shet estimates and evaluate their ability to prioritize genes, we first used these estimates to classify genes essential for survival of human cells in vitro. Genome-wide CRISPR growth screens have measured the effects of gene knockouts on cell survival or proliferation, quantifying the in vitro importance of each gene for fitness [37, 38]. We find that our estimates of shet outperform other constraint metrics at classifying essential genes (Figure 3A, left; bootstrap p<2×10-5 for pairwise differences in AUPRC between our estimates and other metrics). The difference is largest for genes with few expected LOFs, where shet (GeneBayes) retains similar precision and recall while other metrics lose performance (Figure 3A, right). In addition, our estimates of shet outperform other metrics at classifying nonessential genes (Supplementary Figure 2A). DeepLOF [15], the only other method that combines information from both LOF data and gene features, outperforms methods that rely exclusively on LOF data, highlighting the importance of using auxiliary information. Yet, DeepLOF uses only the number of unique LOFs, discarding frequency information. As a result, it is outperformed by our method, indicating that careful modeling of LOF frequencies also contributes to the performance of our approach. Next, we performed further comparisons of our estimates of shet against LOEUF, as LOEUF and its predecessor pLI are extremely popular metrics of constraint. To evaluate the ability of these methods to prioritize disease genes, we first used shet and LOEUF to classify curated developmental disorder genes [39]. Here, shet outperforms LOEUF (Figure 3B; bootstrap p=2×10-9 for the difference in AUPRC) and performs favorably compared to additional constraint metrics (Supplementary Figure 2B). Next, we considered a broader range of phenotypic abnormalities annotated in the Human Phenotype Ontology (HPO) [40]. For each HPO term, we calculated the enrichment of the 10% most constrained genes and depletion of the 10% least constrained genes, ranked using shet or LOEUF. Genes considered constrained by shet are 1.9-fold enriched in HPO terms, compared to 1.5-fold enrichment for genes considered constrained by LOEUF (Figure 3C, left). Additionally, genes considered unconstrained by shet are 3.0-fold depleted in HPO terms, compared to 2.1-fold depletion for genes considered constrained by LOEUF (Figure 3C, right). X-linked inheritance is one of the terms with the largest enrichment of constrained genes (6.6-fold enrichment for shet and 4.2-fold enrichment for LOEUF). The ability of shet to prioritize X-linked genes may prove particularly useful, as many disorders are enriched for X-chromosome genes [41] and the selection on losing a single copy of such genes is stronger on average [4]. Yet, population-scale sequencing alone has less power to detect a given level of constraint on X-chromosome genes, as the number of X chromosomes in a cohort with males is smaller than the number of autosomes. We next assessed if de novo disease-associated variants are enriched in constrained genes, similar to the analyses in [4,5]. To this end, we used data from 31,058 trios to calculate for each gene the enrichment of de novo missense and LOF mutations in offspring with DDs relative to unaffected parents [5]. We found that for both classes of variants, enrichment is higher for genes considered constrained by shet, with the highest enrichment observed for LOF variants (Figure 3D; enrichment of shet and LOEUF respectively, for missense mutations = 2.2, 1.9; splice site mutations = 6.3, 4.6; and nonsense mutations = 9.5, 6.7). Consistent with previous findings, the excess burden of de novo variants is predominantly in highly constrained genes (Supplementary Figure 2C, left). Notably, this difference in enrichment remains after removing known DD genes (Supplementary Figure 2C, right). Together, these results indicate that shet not only improves identification of known disease genes but may also facilitate discovery of novel DD genes [5]. Finally, constraint can also be related to longer-term evolutionary processes that give rise to the variation among individuals or species, including variation in gene expression levels. We expect constrained genes to maintain expression levels closer to their optimal values across evolutionary time scales, as each LOF can be thought of as a ~50% reduction in expression. Consistent with this expectation, we find that less constrained genes have larger absolute differences in expression between human and chimpanzee in cortical cells [42], with a stronger correlation for shet than for LOEUF (Figure 3E). This pattern should also hold when considering the variation in expression within a species. We quantified variance using the normalized standard deviation of gene expression levels estimated from RNA-seq samples in GTEx [43] and found that the variance decreases with increased constraint, again with a stronger correlation for shet (Figure 3E). 2.4 Interpreting the learned relationship between gene features and shet Our framework allows us to learn the relationship between gene features and shet in a statistically principled way. In particular, by fitting a model with all of the features jointly, we can account for dependencies between the features. To interrogate the relationship between features and shet, we divided our gene features into 10 distinct categories (Figure 4A) and trained a separate model per category using only the features in that category. We found that missense constraint, gene expression patterns, evolutionary conservation, and protein embeddings are the most informative categories. Next, we further divided the expression features into 24 subgroups, representing tissues, cell types, and developmental stage (Table 6). Expression patterns in the brain, digestive system, and during development are the most predictive of constraint (Figure 4B). Notably, a study that matched Mendelian disorders to tissues through literature review found that a sizable plurality affect the brain [44]. Meanwhile, most of the top digestive expression features are also related to development (e.g., expression component loadings in a fetal digestive dataset [45]). The importance of developmental features is consistent with the severity of many developmental disorders and the expectation that selection is stronger on early-onset phenotypes [46], supported by the findings of [4]. To quantify the relationship between constraint and individual features, we changed the value of one feature at a time and used the variation in predicted shet over the feature values as the score for each feature (Methods). We first explored some of the individual Gene Ontology (GO) terms most predictive of constraint (Figure 4C). Consistent with the top expression features, the top GO features highlight developmental and brain-specific processes as important for selection. Next, we analyzed network (Figure 4D), gene regulatory (Figure 4E), and gene structure (Figure 4F) features. Protein-protein interaction (PPI) and gene co-expression networks have highlighted “hub” genes involved in numerous cellular processes [47,48], while genes linked to GWAS variants have more complex enhancer landscapes [49]. Consistent with these studies, we find that connectedness in PPI and co-expression networks as well as enhancer and promoter count are positively associated with constraint (Figure 4D,E). In addition, gene structure affects gene function—for example, UTR length and GC content affect RNA stability, translation, and localization [50, 51]—and likewise, several gene structure features are predictive of constraint (Figure 4F). Our results indicate that more complex genes—genes that are involved in more regulatory connections, that are more central to networks, and that have more complex gene structures—are generally more constrained. 2.5 Contextualizing the strength of selection against gene loss-of-function A major benefit of shet over LOEUF and pLI is that shet has a precise, intrinsic meaning in terms of fitness [1–4]. This facilitates comparison of shet between genes, populations, species, and studies. For example, shet can be compared to selection estimated from mutation accumulation or gene deletion experiments performed in model organisms [52,53]. More broadly, selection applies beyond LOFs. While we focused on estimating changes in fitness due to LOFs, consequences of non-coding, missense, and copy number variants can be understood through the same framework, as we expect such variants to also be under negative selection [19] due to ubiquitous stabilizing selection on traits [54]. Quantifying differences in the selection on variants will deepen our understanding of the evolution and genetics of human traits (see Discussion). To contextualize our shet estimates, we compared the distributions of shet for different gene sets (Figure 5A) and genes (Figure 5B), and analyzed them in terms of selection regimes. To define such regimes, we first conceptualized selection on variants as a function of their effects on expression (Figure 5C), where heterozygous LOFs reduce expression by ~50% across all contexts relevant to selection. Under this framework, we can directly compare shet to selection on other variant types—for the hypothetical genes in Figure 5C, a GWAS hit affecting Gene 1 has a stronger selective effect than a LOF affecting Gene 2, despite having a smaller effect on expression. Next, we divided the range of possible shet values into four regimes determined by theoretical considerations [55] and comparisons to other types of variants [56, 57]—nearly neutral (9% of genes), weak selection (22%), strong selection (54%), and extreme selection (15%). LOFs in nearly neutral genes shet<10-4 have minimal effects on fitness—the frequency of such variants is dominated by genetic drift rather than selection [55]. Under the weak selection regime (shet from 10−4 to 10−3), gene LOFs have similar effects on fitness as typical GWAS hits, which usually have small or context-specific effects on gene expression or function [56]. Under the strong selection regime (shet from 10−3 to 10−1), gene LOFs have fitness effects on par with the strongest selection coefficients measured for common variants, such as the selection estimated for adaptive mutations in LCT [57]. Finally, for genes in the extreme selection regime shet>10-1, LOFs have an effect on fitness equivalent to a >2% chance of embryonic lethality, indicating that such LOFs have an extreme effect on survival or reproduction. Gene sets vary widely in their constraint. For example, genes known to be haploinsufficient for severe diseases are almost all under extreme selection. In contrast, genes that can tolerate homozygous LOFs are generally under weak selection. One notable example of such a gene is LPA—while high expression levels are associated with cardiovascular disease, low levels have minimal phenotypic consequences [58, 59], consistent with limited conservation in the sequence or gene expression of LPA across species and populations [60, 61] Other gene sets have much broader distributions of shet values. For example, manually curated recessive genes are under weak to strong selection, indicating that many such genes are either not fully recessive or have pleiotropic effects on other traits under selection. For example, homozygous LOFs in PROC can cause life-threatening congenital blood clotting [62], yet shet for PROC is non-negligible (Figure 5B), consistent with observations that heterozygous LOFs can also increase blood clotting and cause deep vein thrombosis [63]. Similarly, shet values for ClinVar disease genes [64] span the range from weak to extreme selection, with only moderate enrichment for greater constraint relative to all genes. Consistent with this, the effects of disease on fitness depend on disease severity, age-of-onset, and prevalence throughout human history. For example, even though heterozygous loss of BRCA1 greatly increases risk of breast and ovarian cancer [65], BRCA1 is under strong rather than extreme selection. Possible partial explanations are that these cancers have an age-of-onset past reproductive age and are less prevalent in males, or that BRCA1 is subject to some form of antagonistic pleiotropy [14, 66]. 3 Discussion Here, we developed an empirical Bayes approach to accurately infer shet, an interpretable metric of gene constraint. Our approach uses powerful machine learning methods to leverage vast amounts of functional and evolutionary information about each gene while coupling them to a population genetics model. There are two advantages of this approach. First, the additional data sources result in substantially better performance than LOEUF across tasks, from classifying essential genes to identifying pathogenic de novo mutations. These improvements are especially pronounced for the large fraction of genes with few expected LOFs, where LOF data alone is underpowered for estimating constraint. Second, by inferring shet, our estimates of constraint are interpretable in terms of fitness, and we can directly compare the impact of a loss-of-function across genes, populations, species, and studies. As a selection coefficient, shet can also be directly compared to other selection coefficients, even for different types of variants [3, 4]. In general, we believe genes are close to their optimal levels of expression and experience stabilizing selection [54], in which case expression-altering variants decrease fitness, with larger perturbations causing greater decreases (Figure 5C). Estimating the fitness consequences of other types of expression-altering variants, such as duplications or eQTLs, will allow us to map the relationship between genetic variation and fitness in detail, deepening our understanding of the interplay of expression, complex traits, and fitness [10, 56, 67, 68]. A recent method, DeepLOF [15], uses a similar empirical Bayes approach, but by estimating constraint from the number of observed and expected unique LOFs, it inherits the same difficulties regarding interpretation as pLI and LOEUF, and loses information by not considering variant frequencies. On the other hand, another line of work [1, 2], culminating in [4], solved the issues with interpretability by directly estimating shet. Yet, by relying exclusively on LOFs, these estimates are underpowered for ~25% of genes. Furthermore, by using the aggregate frequencies of all LOF variants, previous shet estimates [1, 2, 4] are not robust to misannotated LOF variants. Our approach eliminates this tradeoff between power and interpretability present in existing metrics. Our estimates of shet will be useful for many applications. For example, by informing gene-level priors, LOEUF, pLI, and previous estimates of shet have been used to increase the power of association studies based on rare or de novo mutations [5,6,69]. In such contexts, our shet estimates can be used as a drop-in replacement. Additionally, extremely constrained and unconstrained genes may be interesting to study in their own right. Genes of unknown function with particularly high values of shet should be prioritized for further study. Investigating highly constrained genes may give insights into the mechanisms by which cellular and organism-level phenotypes affect fitness [70]. While we primarily used the posterior means of shet here, our approach provides the entire posterior distribution per gene, similar to [4]. In some applications, different aspects of the posterior may be more relevant than the mean. For example, when prioritizing rare variants for followup in a clinical setting, the posterior probability that shet is high enough for the variant to severely reduce fitness may be more relevant. As more exomes are sequenced, one might expect that we would be better able to more accurately estimate shet. Yet, in a companion paper [16], we show that increasing the sample size used for estimating LOF frequencies will provide essentially no additional information for the ~85% of genes with the lowest values of shet. This fundamental limit on how much we can learn about these genes from LOF data alone highlights the importance of approaches like ours that can leverage additional data types. By sharing information across genes, we can overcome this fundamental limit on how accurately we can estimate constraint. Here we focused on estimating shet, but our empirical Bayes framework, GeneBayes, can be used in any setting where one has a model that ties a gene-level parameter to gene-level observable data (Supplementary Note D). For example, GeneBayes can be used to find trait-associated genes using variants from case/control studies [71, 72], or to improve power to find differentially expressed genes in RNA-seq experiments [73]. We provide a graphical overview of how GeneBayes can be applied more generally in Figure 6. Briefly, GeneBayes requires users to specify a likelihood model and the form of a prior distribution for their parameter of interest. Then, using empirical Bayes and a set of gene features, it improves power to estimate the parameter by flexibly sharing information across similar genes. In summary, we developed a powerful framework for estimating a broadly applicable and readily interpretable metric of constraint, shet. Our estimates provide a more informative ranking of gene importance than existing metrics, and our approach allows us to interrogate potential causes and consequences of natural selection. 4 Methods Empirical Bayes overview Many genes have few observed loss-of-function variants, making it challenging to infer constraint without additional information. Bayesian approaches that specify a prior distribution for each gene can provide such information to improve constraint estimates, but specifying prior distributions is challenging as we have limited prior knowledge about the selection coefficients shet. Empirical Bayes procedures allow us to learn a prior distribution for each gene by combining information across genes. To use the information contained in the gene features, we learn a mapping from a gene’s features to a prior specific for that gene. We parameterize this mapping using gradient-boosted trees, as implemented in NGBoost [17]. Intuitively, this approach learns a notion of “similarity” between genes based on their features, and then shares information across similar genes to learn how shet relates to the gene features. This approach has two major benefits. First, by sharing information between similar genes, it can dramatically improve the accuracy of the predicted shet values, particularly for genes with few expected LOFs. Second, by leveraging the LOF data, this approach allows us to learn about how the various gene features relate to fitness, which cannot be modeled from first principles. For a more in-depth description of our approach along with mathematical and implementation details, see Supplementary Note A. Population genetic likelihood To model how shet relates to the frequency of individual LOF variants, we used the discrete-time Wright-Fisher model, with an approximation of diploid selection with additive fitness effects. We used a composite likelihood approach, assuming independence across individual LOF variants to obtain gene-level likelihoods. Within this composite likelihood, we model each individual variant as either having a selection coefficient of shet with probability 1-pmiss, or having a selection coefficient of 0 with probability pmiss. That is, pmiss acts as the prior probability that a given variant is misannotated, and we assume that misannotated variants evolve neutrally regardless of the strength of selection on the gene. All likelihoods were computed using new machinery developed in a companion paper [16]. Our model depends on a number of parameters—a demographic model of past population sizes, mutation rates for each site, and the probability of misannotation. The demographic model is taken from the literature [75] with modifications as described in [4]. The mutation rates account for trinucleotide context as well as methylation status at CpGs [12]. Finally, we estimated the probability of misannotation from the data. For additional technical details and intuition see Supplementary Note B. Curation of LOF variants We obtained annotations for the consequences of all possible single nucleotide changes to the hg19 reference genome from [76]. The effects of variants on protein function were predicted using Variant Effect Predictor (VEP) version 85 [77] using GENCODE v19 gene annotations [78] as a reference. We defined a variant as a LOF if it was predicted by VEP to be a splice acceptor, splice donor, or stop gain variant. In addition, predicted LOFs were further annotated using LOFTEE [12], which implements a series of filters to identify variants that may be misannotated (for example, LOFTEE considers predicted LOFs near the ends of transcripts as likely misannotations). For our analyses, we only kept predicted LOFs labelled as High Confidence by LOFTEE, which are LOFs that passed all of LOFTEE’s filters. Next, we considered potential criteria for further filtering LOFs: cutoffs for the median exome sequencing read depth, cutoffs for the mean pext (proportion expressed across transcripts) score [76], whether to exclude variants that fall in segmental duplications or regions with low mappability [79], and whether to exclude variants flagged by LOFTEE as potentially problematic but that passed LOFTEE’s primary filters. We trained models with these filters one at a time and in combination, and chose the model that had the best AUPRC in classifying essential from nonessential genes in mice. The filters we evaluated and chose for the final model are reported in Table 2. Since we used mouse gene essentiality data to choose the filters, we do not further evaluate shet on these data. We considered genes to be essential in mice if they are heterozygous lethal, as determined by [12] using data from heterozygous knockouts reported in Mouse Genome Informatics [80]. We classify genes as nonessential if they are reported as “Viable with No Phenotype” by the International Mouse Phenotyping Consortium [81] (annotations downloaded on 12/08/22 from https://www.ebi.ac.uk/mi/impc/essential-genes-search/). Finally, we annotated each variant with its frequency in the gnomAD v2.1.1 exomes [12], a dataset of 125,748 uniformly-analyzed exomes that were largely curated from case–control studies of common adult-onset diseases. gnomAD provides precomputed allele frequencies for all variants that they call. For potential LOFs that are not segregating, gnomAD does not release the number of individuals that were genotyped at those positions. For these sites, we used the median number of genotyped individuals at the positions for which gnomAD does provide this information. We performed this separately on the autosomes and X chromosome. Data sources for the variant annotations, filters, and frequencies, as well as additional information used to compute likelihoods are listed in Table 3. Feature processing and selection We compiled 10 types of gene features from several sources: Gene structure (e.g., number of transcripts, number of exons, GC content) Gene expression across tissues and cell lines Biological pathways and Gene Ontology terms Protein-protein interaction networks Co-expression networks Gene regulatory landscape (e.g., number and properties of enhancers and promoters) Conservation across species Protein embeddings Subcellular localization Missense constraint Additionally, we included an indicator variable that is 1 if the gene is on the non-pseudoautosomal region of the X chromosome and 0 otherwise. For a description of the features within each category and where we acquired them, see Supplementary Note C. Training and validation We fine-tuned a set of hyperparameters for our full empirical Bayes approach, using the best hyperparameters from an initial feature selection step (described in Supplementary Note C) as a starting point. To minimize overfitting, we split the genes into three sets—a training set (chromosomes 7–22, X), a validation set for hyperparameter tuning (chromosomes 2, 4, 6), and a test set to evaluate overfitting (chromosomes 1, 3, 5). During each training iteration, one or more trees were added to the model to fit the natural gradient of the loss on the training set. We stopped model training once the loss on the validation set did not improve for 10 iterations in a row (or the maximum number of iterations, 1,000, was reached). Using this approach, we performed a grid search over the hyperparameters listed in Table 4 and used the combination that minimized the validation loss. For Figure 2B, we reported results from the best model learned using the training set. For all other results, we trained a model on all genes using the hyperparameters and number of training iterations learned during this hyperparameter fine-tuning step. Choosing genes for Table 1 To identify genes that are considered constrained by shet but not by LOEUF, we filtered for genes with shet>0.1 (top ~17% most constrained genes, analogous to the recommended LOEUF cutoff of 0.35 [14], which corresponds to the top ~16% of genes) and LOEUF>0.5 (least constrained ~73% of genes). Of these, we identified genes where heterozygous or hemizygous mutations that decrease the amount of functional protein (e.g. LOF mutations) are associated with Mendelian disorders in the Online Mendelian Inheritance in Man (OMIM) database [36]. We chose genes for Table 1 primarily based on their prominence in the existing literature. Evaluation on additional datasets Definition of human essential and nonessential genes We obtained data from 1,085 CRISPR knockout screens quantifying the effects of genes on cell survival or proliferation from the DepMap portal (22Q2 release) [37, 38]. Scores from each screen are normalized such that nonessential genes identified by [82] have a median score of 0 and that common essential genes identified by [82, 83] have a median score of −1. In classifying essential genes (Figure 3A), we define a gene as essential if its score is < − 1 in at least 25% of screens, and as not essential if its score is > − 1 in all screens. In classifying nonessential genes, we define a gene as nonessential if it has a minimal effect on growth in most cell lines (score > − 0.25 and <0.25 in at least 99% of screens), and as not nonessential if its score is <0 in all screens. Definition of developmental disorder genes Through the Deciphering Developmental Disorders (DDD) study [39], clinicians have annotated a subset of genes with the strength and nature of their association with developmental disorders. We classify genes as developmental disorder genes if they are annotated by the DDD study with confidence_category = definitive and allelic_requirement = monoallelic_autosomal, monoallelic_X_hem (hemizygous), or monoallelic_X_het (heterozygous). We classify genes as not associated with developmental disorders if they are annotated by the DDD study, do not meet the above criteria, and are not annotated with confidence_category = strong or moderate and allelic_requirement = monoallelic_autosomal, monoallelic_X_hem, or monoallelic_X_het. We downloaded genes with DDD annotations from https://www.deciphergenomics.org/ddd/ddgenes on 05/06/2023 . Enrichment/depletion of Human Phenotype Ontology (HPO) genes The Human Phenotype Ontology (HPO) provides a structured organization of phenotypic abnormalities and the genes associated with them, with each HPO term corresponding to a phenotypic abnormality. We calculated the enrichment of constrained genes in each HPO term with at least 200 genes as the ratio (fraction of HPO genes under constraint)/(fraction of background genes under constraint). We defined genes under constraint to be the decile of genes considered most constrained by shet or LOEUF. To choose background genes, we sampled from the set of all genes to match each HPO term’s distribution of expected unique LOFs. Similarly, we calculated the depletion of unconstrained genes in each HPO term as the ratio (fraction of HPO genes not under constraint)/(fraction of background genes not under constraint), where we define genes not under constraint to be the decile of genes considered least constrained by shet or LOEUF. We downloaded HPO phenotype-to-gene annotations from http://purl.obolibrary.org/obo/hp/hpoa/phenotype_to_genes.txt on 01/27/2023 . Enrichment of de novo mutations in developmental disorder patients We used the enrichment metric developed by [5] in their analysis of de novo mutations (DNMs) identified from exome sequencing of 31,058 developmental disorder patients and their unaffected parents. Enrichment of DNMs in developmental disorder patients was calculated as the ratio of observed DNMs in patients over the expected number under a null mutational model that accounts for the study sample size and triplet mutation rate at the mutation sites [84]. For Figure 3D, we calculated the enrichment of DNMs in constrained genes, defined as the decile of genes considered most constrained by shet or LOEUF. For Supplementary Figure 2C, we calculated the enrichment of DNMs in constrained genes with and without known associations with development disorders. We defined a gene as having a known association if it is annotated by the DDD study (see Methods section “Definition of developmental disorder genes”) with confidence_category = definitive or strong and allelic_requirement = monoallelic_autosomal, monoallelic_X_hem (hemizygous), or monoallelic_X_het (heterozygous). For each set of genes, we computed the mean enrichment over sites and 95% Poisson confidence intervals for the mean using the code provided by [5]. Expression variability across species To understand the variability in expression between humans and other species, we focused on gene expression differences between human and chimpanzee as estimated from RNA sequencing of an in vitro model of the developing cerebral cortex for each species [42]. As a metric of variability between the two species, we used the absolute log-fold change (LFC) in gene expression between human and chimpanzee cortical spheroids, which was calculated from samples collected at several time points throughout differentiation of the spheroids. LFC estimates were obtained from Supplementary Table 9 of [42]. To visualize the relationship between constraint and absolute LFC, we plotted a LOESS curve between the constraint on a gene (gene rank from least to most constrained using either shet or LOEUF as the constraint metric) and the absolute LFC for the gene. Curves were calculated using the LOWESS function from the statsmodels package with parameters frac = 0.15 and delta = 10. Expression variability across individuals We used the coefficient of variance (CV) as a metric for gene expression variability across individuals, defined as CV=σi/μi where σi and μi are the standard deviation and mean of the expression level of gene i respectively. Here, expression is in units of Transcripts Per Million. We calculated CV using 17,398 RNA-seq samples in the GTEx v8 release [43], with data from 838 donors and 52 tissues/cell lines. Another potential metric for gene expression variability is the standard deviation for a gene, σi. However, as the mean expression for a gene, μi, is strongly correlated with σi (Spearman ρ=0.73 in GTEx), the relation between σi and shet(i) may be confounded by the relation between μi and shet(i). In contrast, we found that CV is only slightly correlated with μi (Spearman ρ=-0.06 in GTEx). LOESS curves were computed as in “Expression variability across species.” Feature interpretation Training models on feature subsets We grouped features into categories (see Supplementary Table 4 for the features in each category), and trained a model for each category to predict shet from the corresponding features. For each model, we tuned hyperparameters over a subset of the values we considered for the full model (Table 5), and chose the combination of hyperparameters that minimized the loss over genes in the validation set. As a baseline, we trained a model with no features, such that all genes have a shared prior distribution that is learned from the LOF data—this model is analogous to a standard empirical Bayes model. Definition of expression feature subsets We grouped gene expression features into 24 categories representing tissues, cell types, and developmental stage using terms present in the feature names (Table 6). Scoring individual features To score individual gene features, we varied the value of one feature at a time and calculated the variance in predicted shet as a feature score. In more detail, we fixed each feature to values spanning the range of observed values for that feature (0th, 2nd, ..., 98th, and 100th percentile), such that all genes shared the same feature value. Then, for each of these 51 feature values, we averaged the shet values predicted by the learned priors over all genes, where the predicted shet for each gene is the mean of its prior. We denote this averaged prediction by shet(f){p} for some feature f and percentile p. Finally, we define the score for feature f as score f=sd(shet(f){0},shet(f){2},…,shet(f){98},shet(f){100}), where sd is a function computing the sample standard deviation. In other words, a feature with a high score is one for which varying its value causes high variance in the predicted shet. For the lineplots in Figures 4C–4F, we scale the predictions shet(f){p} for each feature f by subtracting (shet(f){0}+shet(f){100})/2 from each prediction. Pruning features before computing feature scores While investigating the effects of features on predicted shet, we found that including highly correlated features in the model could produce unintuitive results, such as opposite correlations with shet for highly similar features. Therefore, for Figures 4C–4F, we first pruned the set of features to minimize pairwise correlations between the remaining features. To do this, we randomly kept one feature in each group of correlated features, where such a group is defined as a set of features where each feature in the set has an absolute Spearman ρ>0.7 to some other feature in the set. For Figures 4C–4F, we trained models on the relevant features in this pruned set (gene ontology, network, gene regulatory, and gene structure features for Figures 4C, 4D, 4E, and 4F respectively). After feature pruning, we found the directions of effect for the features were consistent with their marginal directions of effect. Supplementary Material Supplement 1 Acknowledgements We would like to thank Ipsita Agarwal, Molly Przeworski, Jesse Engreitz, and members of the Pritchard Lab for valuable feedback and discussions. This work was supported by NIH grants R01HG011432 and R01HG008140. Data availability Posterior means and 95% credible intervals for shet are available in Supplementary Table 2. Posterior densities for shet are available in Supplementary Table 3. A description of the gene features is available in Supplementary Table 4. These supplementary tables are also available at [74], along with likelihoods for shet, LOF variants with misannotation probabilities, and gene feature tables. Figure 1: Limitations of LOEUF and schematic for inferring shet using GeneBayes. A) Stacked histogram of the expected number of unique LOFs per gene, where the distribution for genes considered unconstrained (respectively constrained) by LOEUF are colored in red (respectively blue). Genes with LOEUF<0.35 are considered constrained, while all other genes are unconstrained (Methods). The plot is truncated on the x-axis at 100 expected LOFs. B) Scatterplot of the observed against the expected number of unique LOFs per gene. The dashed line denotes observed = expected. Each point is a gene, colored by its LOEUF score; genes with LOEUF>1 are colored as LOEUF=1. C) Schematic for estimating shet using GeneBayes, highlighting the major components of the model: prior (blue boxes) and likelihood (red boxes). Parameters of the prior are learned by maximizing the likelihood (red arrow). Combining the prior and likelihood produces posteriors over shet (purple box). See Methods for details. Figure 2: Factors that contribute to our estimates of shet. A) Likelihood curves for different allele frequencies (f) and mutation rates. B) Scatterplot of shet estimated from LOF data (y-axis; posterior mean from a model without features) against the prior’s predictions of shet (x-axis; mean of learned prior). Dotted line denotes y=x. Each point is a gene, colored by the expected number of LOFs. C) Comparison of posterior distributions of shet (95% Credible Intervals) from a model with (blue lines) and without (orange lines) gene features. Genes are ordered by their posterior mean in the model with gene features. D) Top: scatterplot of LOEUF (y-axis) and our shet estimates (x-axis; posterior mean). Each point is a gene, colored by the expected number of LOFs. Bottom: scatterplot of shet estimates from [4] (y-axis; posterior mode) and our shet estimates (x-axis; posterior mean). Numbered points refer to genes in panels E and F. E) TBC1D3 and PLN are two example genes where the gene features substantially affect the posterior. We plot their posterior distributions (blue) and likelihoods (orange; rescaled so that the area under the curve = 1). F) AARD and TWIST1 are two example genes with the same LOEUF but different shet. Posteriors and likelihoods are plotted as in panel E. Figure 3: GeneBayes estimates of shet perform well at identifying constrained and unconstrained genes. A) Precision-recall curves comparing the performance of shet against other methods in classifying essential genes (left: all genes, right: quartile of genes with the fewest expected unique LOFs). B) Precision-recall curves comparing the performance of shet against LOEUF in classifying developmental disorder genes. C) Scatterplots showing the enrichment (respectively depletion) of the top 10% most (respectively least) constrained genes in HPO terms, with genes ranked by shet (y-axis) or LOEUF (x-axis). D) Enrichment of de novo mutations in patients with developmental disorders, calculated as the observed number of mutations over the expected number under a null mutational model. We plot the enrichment of missense, splice, and nonsense variants in the 10% most constrained genes, ranked by shet (blue) or LOEUF (orange). Bars represent 95% confidence intervals. E) Left: LOESS curve showing the relationship between constraint (gene rank, x-axis) and absolute log fold change in expression between chimp and human cortical cells (y-axis). Genes are ranked by shet (blue) or LOEUF (orange) Right: LOESS curve showing the relationship between constraint (gene rank, x-axis) and gene expression variation (normalized standard deviation) in GTEx samples. Figure 4: Breakdown of the gene features important for shet prediction. A) Ordered from highest to lowest, plot of the mean per-gene log likelihood over the test genes for models separately trained on categories of features. “All” and “Baseline” include all and no features respectively. B) Plot of the mean per-gene log likelihood, as in panel A, for models separately trained on expression features grouped by tissue, cell type, or developmental stage. C) Ordered from highest to lowest, feature scores for individual gene ontology (GO) terms. Inset: lineplot showing the change in predicted shet for a feature as the feature value is varied. D) Lineplot as in panel C (inset) for protein-protein interaction (PPI) and co-expression features, E) enhancer and promoter features, and F) gene structure features. Figure 5: Comparing selection on LOFs (shet) between genes and to selection on other variant types. A) Distributions of shet for gene sets, calculated by averaging the posterior distributions for the genes in each gene set. Gene sets are sorted by the mean of their distributions. Colors represent four general selection regimes. B) Posterior distributions of shet for individual genes, ordered by mean. Lines represent 95% credible intervals, with labeled genes represented by thick black lines. Colors represent the selection regimes in panel A. C) Schematic demonstrating the hypothesized relationship between changes in expression (x-axis, log2 scale) and selection (y-axis) against these changes for two hypothetical genes, assuming stabilizing selection. The shapes of the curves are not estimated from real data. Background colors represent the selection regimes in panel A. The red points and line represent the effects of heterozygous LOFs and deletions on expression and selection, while the blue points and line represent the potential effects of other types of variants. Figure 6: GeneBayes is a flexible framework for estimating gene-level properties. Schematic for how GeneBayes can be applied to estimate gene-level properties beyond shet, showing the key inputs and outputs and two example applications. See Supplementary Note D for more details. Table 1: OMIM genes constrained by shet but not by LOEUF. Gene shet LOEUF Obs. Exp. Condition and reference RPS15A * 0.61 0.56 0 5.4 Diamond-Blackfan anemia: Red blood cell aplasia resulting in growth, craniofacial, and other congenital defects [23] DCX 0.48 0.62 3 12.6 Lissencephaly: Migrational arrest of neurons resulting in mental re-tardation and seizures [24] SOX2 0.33 0.57 1 8.3 Syndromic microphthalmia: Missing or small eyes from birth [25] NDP 0.33 0.88 0 3.4 Norrie disease: Retinal dystrophy resulting in early childhood blindness, mental disorders, and deafness [26] EIF5A 0.32 0.54 1 8.7 Faundes-Banka syndrome: Developmental delay, microcephaly, and facial dysmorphisms [27] CDKN1C 0.27 0.53 0 5.7 Beckwith-Wiedemann syndrome: Pediatric overgrowth with predisposition to tumor development [28] TGIF1 0.25 0.91 5 11.5 Holoprosencephaly: Structural malformation of the forebrain during development [29] SH2D1A 0.23 0.96 1 4.9 Lymphoproliferative syndrome: Severe immune dysregulation due to improper lymphocyte apoptosis [30] CEBPA 0.17 1.18 0 2.4 Acute myeloid leukemia: Blood and bone marrow cancer with rapid progression [31] GATA4 0.15 0.53 3 14.7 Atrial septal defect: Congenital heart defect resulting in a hole between the atria [32] TIMP3 0.13 0.53 2 11.8 Sorsby fundus dystrophy: Retinal dystrophy that causes loss of vision [33] FOXC2 0.13 0.79 3 9.8 Lymphedema-distichiasis syndrome: Lymphedema of the limbs and double rows of eyelashes [34] IGF2 0.12 1.13 3 6.8 Silver-Russell syndrome: Growth retardation, relative macrocephaly, and feeding difficulties [35] PLN 0.12 1.56 0 1.5 Dilated cardiomyopathy: Enlarged heart chambers, decreased contrac-tile function, and heart failure [20] TWIST1 0.11 1.06 1 4.5 Saethre-Chotzen syndrome: Craniosynostosis, facial dysmorphism, and hand and foot abnormalities [21] [22] Mutations that disrupt the functions of these genes are associated with Mendelian diseases in the OMIM database [36]. Genes are ordered by shet (posterior mean). Obs. and Exp. are the unique number of observed and expected LOFs respectively. * RPS15A is associated with Diamond-Blackfan anemia along with nine other genes considered constrained by shet but not by LOEUF (Supplementary Table 1). Table 2: Filtering criteria for LOF curation Filtering criterion Tested values Best value Cutoff for sequencing read depth (median across exomes) 5×, 10×, 20× 20× Cutoff for mean pext across tissues 0.05, 0.1 0.05 Filter if variant falls in a segmental duplication or low mappability region True, False False Filter if variant is flagged as potentially problematic True, False True Table 3: Sources for LOF data Resource Link Annotations for possible LOFs gs://gnomad-public/papers/2019-tx-annotation/pre_computed/all.possible.snvs.tx_annotated.GTEx.v7.021520.tsv Mean methylation for CpG sites gs://gcp-public-data--gnomad/resources/methylation Exome sequencing coverage gs://gcp-public-data--1gnomad/release/2.1/coverage/exomes/gnomad.exomes.coverage.summary.tsv.bgz Variant frequencies gs://gcp-public-data--gnomad/release/2.1.1/vcf/exomes/gnomad.exomes.r2.1.1.sites.vcf.bgz Low mappability and segmental duplications https://ftp-trace.ncbi.nlm.nih.gov/ReferenceSamples/giab/release/genome-stratifications/v3.1/GRCh37/Union/GRCh37_alllowmapandsegdupregions.bed.gz Table 4: Parameters for fitting the gradient-boosted trees Parameter(s) Tested values Best value Learning rate 0.0125, 0.05, 0.2 0.0125 Maximum tree depth (max_depth) 3, 4, 5 3 Data subsampling ratio (subsample) 0.6, 0.8, 1 0.8 Minimum weight of a leaf node (min_child_weight) 1, 2, 4 1 L1 regularization (alpha) 0, 1, 2 2 L2 regularization (lambda) 1, 2, 4 1 Number of trees to fit per iteration (n_estimators) 1, 2, 4 4 Table 5: Parameters for feature subsets Parameter(s) Tested values Learning rate 0.0125, 0.05 Maximum tree depth (max_depth) 3 Data subsampling ratio (subsample) 0.8, 1 Minimum weight of a leaf node (min_child_weight) 1 L1 regularization (alpha) 0, 1, 2 L2 regularization (lambda) 1 Number of trees to fit per iteration (n_estimators) 1, 2, 4 Table 6: Terms used to define tissues for expression features Category Terms in the feature (not case sensitive) Brain brain, nerve, microglia, hippocampus Digestive digestive, gut, gutendoderm, intestine, colon, ileum Development development, gastrulation, embryo Lung lung, airway Eye eye, retina Endothelium endothelium Muscle muscle Hair follicle hairfollicle Kidney kidney Immune immune, monocytes, nk, tcell, pbmc Prostate prostate Blood blood, heme, fetalblood Adipocyte adipocyte Heart heart, aorta Thymus thymus Pancreas pancreas, islets, pancreasductal Liver liver Testis testis Synovial fibroblast synovialfibroblast Bladder bladder Placenta placenta Bone marrow bonemarrow CSF csf Lymph nodes lymphnodes Additional Declarations: There is NO Competing Interest. Code availability GeneBayes and code for estimating shet are available at https://github.com/tkzeng/GeneBayes. ==== Refs References [1] Cassa CA , Weghorn D , Balick DJ , Jordan DM , Nusinow D , Samocha KE , Estimating the selective effects of heterozygous protein-truncating variants from human exome data. Nature Genetics. 2017;49 (5 ):806–10.28369035 [2] Weghorn D , Balick DJ , Cassa C , Kosmicki JA , Daly MJ , Beier DR , Applicability of the Mutation–Selection Balance Model to Population Genetics of Heterozygous Protein-Truncating Variants in Humans. Molecular Biology and Evolution. 2019;36 (8 ):1701–10.31004148 [3] Fuller ZL , Berg JJ , Mostafavi H , Sella G , Przeworski M . Measuring intolerance to mutation in human genetics. Nature Genetics. 2019;51 (5 ):772–6.30962618 [4] Agarwal I , Fuller ZL , Myers SR , Przeworski M . Relating pathogenic loss-of function mutations in humans to their evolutionary fitness costs. eLife. 2023;12 :e83172.36648429 [5] Kaplanis J , Samocha KE , Wiel L , Zhang Z , Arvai KJ , Eberhardt RY , Evidence for 28 genetic disorders discovered by combining healthcare and research data. Nature. 2020;586 (7831 ):757–62.33057194 [6] Fu JM , Satterstrom FK , Peng M , Brand H , Collins RL , Dong S , Rare coding variation provides insight into the genetic architecture and phenotypic context of autism. Nature Genetics. 2022;54 (9 ):1320–31.35982160 [7] Whiffin N , Armean IM , Kleinman A , Marshall JL , Minikel EV , Goodrich JK , The effect of LRRK2 loss-of-function variants in humans. Nature Medicine. 2020;26 (6 ):869–77. [8] Gazal S , Weissbrod O , Hormozdiari F , Dey KK , Nasser J , Jagadeesh KA , Combining SNP-to-gene linking strategies to identify disease genes and assess disease omnigenicity. Nature Genetics. 2022;54 (6 ):827–36.35668300 [9] Wang X , Goldstein DB . Enhancer domains predict gene pathogenicity and inform gene discovery in complex disease. The American Journal of Human Genetics. 2020;106 (2 ):215–33.32032514 [10] Mostafavi H , Spence JP , Naqvi S , Pritchard JK . Limited overlap of eQTLs and GWAS hits due to systematic differences in discovery. bioRxiv. 2022:2022–05. [11] Lek M , Karczewski KJ , Minikel EV , Samocha KE , Banks E , Fennell T , Analysis of protein-coding genetic variation in 60,706 humans. Nature. 2016;536 (7616 ):285–91.27535533 [12] Karczewski KJ , Francioli LC , Tiao G , Cummings BB , Alföldi J , Wang Q , The mutational constraint spectrum quantified from variation in 141,456 humans. Nature. 2020;581 (7809 ):434–43.32461654 [13] Gillespie JH . Population genetics: a concise guide. JHU press; 2004. [14] Gudmundsson S , Singer-Berk M , Watts NA , Phu W , Goodrich JK , Solomonson M , Variant interpretation using population databases: Lessons from gnomAD. Human Mutation. 2022;43 (8 ):1012–30.34859531 [15] LaPolice TM , Huang YF . A deep learning framework for predicting human essential genes from population and functional genomic data. bioRxiv. 2021:2021–12. [16] Spence J , Zeng T , Mostafavi H , Pritchard J . Scaling the discrete-time Wright-Fisher model to biobank-scale datasets. bioRxiv. 2023. [17] Duan T , Anand A , Ding DY , Thai KK , Basu S , Ng A , Ngboost: Natural gradient boosting for probabilistic prediction. In: International Conference on Machine Learning. PMLR; 2020. p. 2690–700. [18] Ewens WJ . Mathematical population genetics: theoretical introduction. vol. 27 . Springer; 2004. [19] Agarwal I , Przeworski M . Mutation saturation for fitness effects at human CpG sites. eLife. 2021;10 :e71513.34806592 [20] Haghighi K , Kolokathis F , Pater L , Lynch RA , Asahi M , Gramolini AO , Human phospholamban null results in lethal dilated cardiomyopathy revealing a critical difference between mouse and human. The Journal of Clinical Investigation. 2003;111 (6 ):869–76.12639993 [21] Howard TD , Paznekas WA , Green ED , Chiang LC , Ma N , Luna RIOD, Mutations in TWIST, a basic helix–loop–helix transcription factor, in Saethre-Chotzen syndrome. Nature Genetics. 1997;15 (1 ):36–41.8988166 [22] Ghouzzi VE , Merrer ML , Perrin-Schmitt F , Lajeunie E , Benit P , Renier D , Mutations of the TWIST gene in the Saethre-Chotzene syndrome. Nature Genetics. 1997;15 (1 ):42–6.8988167 [23] Da Costa L , Leblanc T , Mohandas N . Diamond-Blackfan anemia. Blood. 2020;136 (11 ):1262–73.32702755 [24] des Portes V , Pinard JM , Billuart P , Vinet MC , Koulakoff A , Carrié A , A novel CNS gene required for neuronal migration and involved in X-linked subcortical laminar heterotopia and lissencephaly syndrome. Cell. 1998;92 (1 ):51–61.9489699 [25] Fantes J , Ragge NK , Lynch SA , McGill NI , Collin JRO , Howard-Peebles PN , Mutations in SOX2 cause anophthalmia. Nature Genetics. 2003;33 (4 ):462–3. [26] Berger W , de Pol Dv , Warburg M , Gal A , Bleeker-Wagemakers L , de Silva H , Mutations in the candidate gene for Norrie disease. Human Molecular Genetics. 1992;1 (7 ):461–5.1307245 [27] Faundes V , Jennings MD , Crilly S , Legraie S , Withers SE , Cuvertino S , Impaired eIF5A function causes a Mendelian disorder that is partially rescued in model systems by spermidine. Nature Communications. 2021;12 (1 ):833. [28] Hatada I , Ohashi H , Fukushima Y , Kaneko Y , Inoue M , Komoto Y , An imprinted gene p57 KIP2 is mutated in Beckwith–Wiedemann syndrome. Nature Genetics. 1996;14 (2 ):171–3.8841187 [29] Gripp KW , Wotton D , Edwards MC , Roessler E , Ades L , Meinecke P , Mutations in TGIF cause holoprosencephaly and link NODAL signalling to human neural axis determination. Nature Genetics. 2000;25 (2 ):205–8.10835638 [30] Coffey AJ , Brooksbank RA , Brandau O , Oohashi T , Howell GR , Bye JM , Host response to EBV infection in X-linked lymphoproliferative disease results from mutations in an SH2-domain encoding gene. Nature Genetics. 1998;20 (2 ):129–35.9771704 [31] Smith ML , Cavenagh JD , Lister TA , Fitzgibbon J . Mutation of CEBPA in familial acute myeloid leukemia. New England Journal of Medicine. 2004;351 (23 ):2403–7.15575056 [32] Garg V , Kathiriya IS , Barnes R , Schluterman MK , King IN , Butler CA , GATA4 mutations cause human congenital heart defects and reveal an interaction with TBX5. Nature. 2003;424 (6947 ):443–7.12845333 [33] Langton KP , McKie N , Curtis A , Goodship JA , Bond PM , Barker MD , A novel tissue inhibitor of metalloproteinases-3 mutation reveals a common molecular phenotype in Sorsby’s fundus dystrophy. Journal of Biological Chemistry. 2000;275 (35 ):27027–31.10854443 [34] Fang J , Dagenais SL , Erickson RP , Arlt MF , Glynn MW , Gorski JL , Mutations in FOXC2 (MFH-1), a forkhead family transcription factor, are responsible for the hereditary lymphedema-distichiasis syndrome. The American Journal of Human Genetics. 2000;67 (6 ):1382–8.11078474 [35] Begemann M , Zirn B , Santen G , Wirthgen E , Soellner L , Büttel HM , Paternally inherited IGF2 mutation and growth restriction. New England Journal of Medicine. 2015;373 (4 ):349–56.26154720 [36] Amberger JS , Bocchini CA , Schiettecatte F , Scott AF , Hamosh A . OMIM. org: Online Mendelian Inheritance in Man (OMIM®), an online catalog of human genes and genetic disorders. Nucleic Acids Research. 2015;43 (D1 ):D789–98.25428349 [37] Meyers RM , Bryan JG , McFarland JM , Weir BA , Sizemore AE , Xu H , Computational correction of copy number effect improves specificity of CRISPR–Cas9 essentiality screens in cancer cells. Nature Genetics. 2017;49 (12 ):1779–84.29083409 [38] Ghandi M , Huang FW , Jané-Valbuena J , Kryukov GV , Lo CC , McDonald III ER , Next-generation characterization of the cancer cell line encyclopedia. Nature. 2019;569 (7757 ):503–8.31068700 [39] Wright CF , Campbell P , Eberhardt RY , Aitken S , Perrett D , Brent S , Genomic Diagnosis of Rare Pediatric Disease in the United Kingdom and Ireland. New England Journal of Medicine. 2023. [40] Köhler S , Gargano M , Matentzoglu N , Carmody LC , Lewis-Smith D , Vasilevsky NA , The human phenotype ontology in 2021. Nucleic Acids Research. 2021;49 (D1 ):D1207–17.33264411 [41] Leitão E , Schröder C , Parenti I , Dalle C , Rastetter A , Kühnel T , Systematic analysis and prediction of genes associated with monogenic disorders on human chromosome X. Nature Communications. 2022;13 (1 ):6570. [42] Agoglia RM , Sun D , Birey F , Yoon SJ , Miura Y , Sabatini K , Primate cell fusion disentangles gene regulatory divergence in neurodevelopment. Nature. 2021;592 (7854 ):421–7.33731928 [43] Consortium G. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science. 2020;369 (6509 ):1318–30.32913098 [44] Basha O , Argov CM , Artzy R , Zoabi Y , Hekselman I , Alfandari L , Differential network analysis of multiple human tissue interactomes highlights tissue-selective processes and genetic disorder genes. Bioinformatics. 2020;36 (9 ):2821–8.31960892 [45] Gao S , Yan L , Wang R , Li J , Yong J , Zhou X , Tracing the temporal-spatial transcriptome landscapes of the human fetal digestive tract using single-cell RNA-sequencing. Nature Cell Biology. 2018;20 (6 ):721–34.29802404 [46] Charlesworth B , Evolution in age-structured populations. vol. 2 . Cambridge University Press Cambridge; 1994. [47] Barrio-Hernandez I , Schwartzentruber J , Shrivastava A , Del-Toro N , Gonzalez A , Zhang Q , Network expansion of genetic associations defines a pleiotropy map of human cell biology. Nature Genetics. 2023:1–10. [48] Van Dam S , Vosa U , van der Graaf A , Franke L , de Magalhaes JP . Gene co-expression analysis for functional classification and gene–disease predictions. Briefings in Bioinformatics. 2018;19 (4 ):575–92.28077403 [49] Nasser J , Bergman DT , Fulco CP , Guckelberger P , Doughty BR , Patwardhan TA , Genome-wide enhancer maps link risk variants to disease genes. Nature. 2021;593 (7858 ):238–43.33828297 [50] Mayr C. Regulation by 3’-untranslated regions. Annual Review of Genetics. 2017;51 :171–94. [51] Leppek K , Das R , Barna M . Functional 5’ UTR mRNA structures in eukaryotic translation regulation and how to find them. Nature Reviews Molecular Cell Biology. 2018;19 (3 ):158–74.29165424 [52] Agrawal AF , Whitlock MC . Inferences about the distribution of dominance drawn from yeast gene knockout data. Genetics. 2011;187 (2 ):553–66.21098719 [53] Mukai T , Chigusa SI , Mettler L , Crow JF . Mutation rate and dominance of genes affecting viability in Drosophila melanogaster. Genetics. 1972;72 (2 ):335–55.4630587 [54] Sella G , Barton NH . Thinking about the evolution of complex traits in the era of genome-wide association studies. Annual Review of Genomics and Human Genetics. 2019;20 :461–93. [55] Charlesworth B. Effective population size and patterns of molecular evolution and variation. Nature Reviews Genetics. 2009;10 (3 ):195–205. [56] Simons YB , Mostafavi H , Smith CJ , Pritchard JK , Sella G . Simple scaling laws control the genetic architectures of human complex traits. bioRxiv. 2022:2022–10. [57] Mathieson I , Terhorst J . Direct detection of natural selection in Bronze Age Britain. Genome Research. 2022;32 (11–12 ):2057–67.36316157 [58] Emdin CA , Khera AV , Natarajan P , Klarin D , Won HH , Peloso GM , Phenotypic characterization of genetically lowered human lipoprotein (a) levels. Journal of the American College of Cardiology. 2016;68 (25 ):2761–72.28007139 [59] Langsted A , Nordestgaard BG , Kamstrup PR . Low lipoprotein (a) levels and risk of disease in a large, contemporary, general population study. European Heart Journal. 2021;42 (12 ):1147–56.33724357 [60] Rausell A , Luo Y , Lopez M , Seeleuthner Y , Rapaport F , Favier A , Common homozygosity for predicted loss-of-function variants reveals both redundant and advantageous effects of dispensable human genes. Proceedings of the National Academy of Sciences. 2020;117 (24 ):13626–36. [61] Reyes-Soffer G , Ginsberg HN , Berglund L , Duell PB , Heffron SP , Kamstrup PR , Lipoprotein (a): a genetically determined, causal, and prevalent risk factor for atherosclerotic cardiovascular disease: a scientific statement from the American Heart Association. Arteriosclerosis, Thrombosis, and Vascular Biology. 2022;42 (1 ):e48–60.34647487 [62] Millar DS , Johansen B , Berntorp E , Minford A , Bolton-Maggs P , Wensley R , Molecular genetic analysis of severe protein C deficiency. Human Genetics. 2000;106 :646–53.10942114 [63] Romeo G , Hassan HJ , Staempfli S , Roncuzzi L , Cianetti L , Leonardi A , Hereditary thrombophilia: identification of nonsense and missense mutations in the protein C gene. Proceedings of the National Academy of Sciences. 1987;84 (9 ):2829–32. [64] Landrum MJ , Chitipiralla S , Brown GR , Chen C , Gu B , Hart J , ClinVar: improvements to accessing data. Nucleic Acids Research. 2020;48 (D1 ):D835–44.31777943 [65] Couch FJ , Nathanson KL , Offit K . Two decades after BRCA: setting paradigms in personalized cancer care and prevention. Science. 2014;343 (6178 ):1466–70.24675953 [66] Smith KR , Hanson HA , Hollingshaus MS . BRCA1 and BRCA2 mutations and female fertility. Current Opinion in Obstetrics & Gynecology. 2013;25 (3 ):207.23411475 [67] O’Connor LJ , Schoech AP , Hormozdiari F , Gazal S , Patterson N , Price AL . Extreme poly-genicity of complex traits is explained by negative selection. The American Journal of Human Genetics. 2019;105 (3 ):456–76.31402091 [68] Benton ML , Abraham A , LaBella AL , Abbot P , Rokas A , Capra JA . The influence of evolutionary history on human health and disease. Nature Reviews Genetics. 2021;22 (5 ):269–83. [69] Satterstrom FK , Kosmicki JA , Wang J , Breen MS , De Rubeis S , An JY , Large-scale exome sequencing study implicates both developmental and functional changes in the neurobiology of autism. Cell. 2020;180 (3 ):568–84.31981491 [70] Gardner EJ , Neville MD , Samocha KE , Barclay K , Kolk M , Niemi ME , Reduced reproductive success is associated with selective constraint on human genes. Nature. 2022;603 (7903 ):858–63.35322230 [71] He X , Sanders SJ , Liu L , De Rubeis S , Lim ET , Sutcliffe JS , Integrated model of de novo and inherited genetic variants yields greater power to identify risk genes. PLoS Genetics. 2013;9 (8 ):e1003671.23966865 [72] Zhu X , Stephens M . Bayesian large-scale multiple regression with summary statistics from genome-wide association studies. The Annals of Applied Statistics. 2017;11 (3 ):1561.29399241 [73] Boyeau P , Regier J , Gayoso A , Jordan MI , Lopez R , Yosef N . An empirical Bayes method for differential expression analysis of single cells with deep generative models. bioRxiv. 2022:2022–05. [74] Zeng T , Spence JP , Mostafavi H , Pritchard JK . s_het estimates from GeneBayes and other supplementary datasets. Zenodo; 2023. Available from: 10.5281/zenodo.7939768. [75] Schiffels S , Durbin R . Inferring human population size and separation history from multiple genome sequences. Nature Genetics. 2014;46 (8 ):919–25.24952747 [76] Cummings BB , Karczewski KJ , Kosmicki JA , Seaby EG , Watts NA , Singer-Berk M , Transcript expression-aware annotation improves rare variant interpretation. Nature. 2020;581 (7809 ):452–8.32461655 [77] McLaren W , Gil L , Hunt SE , Riat HS , Ritchie GR , Thormann A , The ensembl variant effect predictor. Genome Biology. 2016;17 (1 ):1–14.26753840 [78] Frankish A , Carbonell-Sala S , Diekhans M , Jungreis I , Loveland JE , Mudge JM , GENCODE: reference annotation for the human and mouse genomes in 2023. Nucleic Acids Research. 2023;51 (D1 ):D942–9.36420896 [79] Olson ND , Wagner J , McDaniel J , Stephens SH , Westreich ST , Prasanna AG , PrecisionFDA Truth Challenge V2: Calling variants from short and long reads in difficult-to-map regions. Cell Genomics. 2022;2 (5 ):100129.35720974 [80] Blake JA , Baldarelli R , Kadin JA , Richardson JE , Smith CL , Bult CJ . Mouse Genome Database (MGD): Knowledgebase for mouse–human comparative biology. Nucleic Acids Research. 2021;49 (D1 ):D981–7.33231642 [81] Groza T , Gomez FL , Mashhadi HH , Muñoz-Fuentes V , Gunes O , Wilson R , The International Mouse Phenotyping Consortium: comprehensive knockout phenotyping underpinning the study of human disease. Nucleic acids research. 2023;51 (D1 ):D1038–45.36305825 [82] Hart T , Brown KR , Sircoulomb F , Rottapel R , Moffat J . Measuring error rates in genomic perturbation screens: gold standards for human functional genomics. Molecular Systems Biology. 2014;10 (7 ):733.24987113 [83] Blomen VA , Májek P , Jae LT , Bigenzahn JW , Nieuwenhuis J , Staring J , Gene essentiality and synthetic lethality in haploid human cells. Science. 2015;350 (6264 ):1092–6.26472760 [84] Samocha KE , Robinson EB , Sanders SJ , Stevens C , Sabo A , McGrath LM , A framework for the interpretation of de novo mutation in human disease. Nature Genetics. 2014;46 (9 ):944–50.25086666 [85] Si Amari . Natural Gradient Works Efficiently in Learning. Neural Computation. 1998;10 (2 ):251–76. [86] Paszke A , Gross S , Massa F , Lerer A , Bradbury J , Chanan G , Pytorch: An imperative style, high-performance deep learning library. Advances in Neural Information Processing Systems. 2019;32. [87] Loshchilov I , Hutter F . Decoupled weight decay regularization. arXiv preprint arXiv:171105101. 2017. [88] Gómez P , Toftevaag HH , Meoni G . torchquad: Numerical Integration in Arbitrary Dimensions with PyTorch. Journal of Open Source Software. 2021;6 (64 ):3439. [89] Chen T , Guestrin C . Xgboost: A scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; 2016. p. 785–94. [90] Sawyer SA , Hartl DL . Population genetics of polymorphism and divergence. Genetics. 1992;132 (4 ):1161–76.1459433 [91] Harpak A , Bhaskar A , Pritchard JK . Mutation rate variation is a primary determinant of the distribution of allele frequencies in humans. PLoS Genetics. 2016;12 (12 ):e1006489.27977673 [92] Varin C , Reid N , Firth D . An overview of composite likelihood methods. Statistica Sinica. 2011:5–42. [93] Ramoni RB , Mulvihill JJ , Adams DR , Allard P , Ashley EA , Bernstein JA , The undiagnosed diseases network: accelerating discovery about health and disease. The American Journal of Human Genetics. 2017;100 (2 ):185–92.28157539 [94] Consortium GP , An integrated map of genetic variation from 1,092 human genomes. Nature. 2012;491 (7422 ):56.23128226 [95] Lee Y , Nelder JA . Hierarchical generalized linear models. Journal of the Royal Statistical Society: Series B (Methodological). 1996;58 (4 ):619–56. [96] Meng XL . Decoding the h-likelihood. Statistical Science. 2009;24 (3 ):280–93. [97] Weeks EM , Ulirsch JC , Cheng NY , Trippe BL , Fine RS , Miao J , Leveraging polygenic enrichments of gene features to predict genes underlying complex traits and diseases. medRxiv. 2020:2020–09. [98] Boukas L , Bjornsson HT , Hansen KD . Promoter CpG density predicts downstream gene loss-of-function intolerance. The American Journal of Human Genetics. 2020;107 (3 ):487–98.32800095 [99] Pers TH , Karjalainen JM , Chan Y , Westra HJ , Wood AR , Yang J , Biological interpretation of genome-wide association studies using predicted gene functions. Nature Communications. 2015;6 (1 ):5890. [100] The Gene Ontology resource: enriching a GOld mine. Nucleic acids research. 2021;49 (D1 ):D325–34.33290552 [101] Raina P , Guinea R , Chatsirisupachai K , Lopes I , Farooq Z , Guinea C , GeneFriends: gene co-expression databases and tools for humans and model organisms. Nucleic Acids Research. 2023;51 (D1 ):D145–58.36454018 [102] Consortium G , Ardlie KG , Deluca DS , Segrè AV , Sullivan TJ , Young TR , The Genotype-Tissue Expression (GTEx) pilot analysis: multitissue gene regulation in humans. Science. 2015;348 (6235 ):648–60.25954001 [103] DGT RPC , Consortium F , A promoter-level mammalian expression atlas. Nature. 2014;507 (7493 ):462–70.24670764 [104] Fulco CP , Nasser J , Jones TR , Munson G , Bergman DT , Subramanian V , Activity-by-contact model of enhancer–promoter regulation from thousands of CRISPR perturbations. Nature Genetics. 2019;51 (12 ):1664–9.31784727 [105] Roadmap EC , Kundaje A , Meuleman W , Ernst J , Bilenky M , Yen A , Integrative analysis of 111 reference human epigenomes. Nature. 2015;518 (7539 ):317–30.25693563 [106] Liu Y , Sarkar A , Kheradpour P , Ernst J , Kellis M . Evidence of reduced recombination rate in human regulatory domains. Genome Biology. 2017;18 (1 ):1–11.28077169 [107] Siepel A , Bejerano G , Pedersen JS , Hinrichs AS , Hou M , Rosenbloom K , Evolutionarily conserved elements in vertebrate, insect, worm, and yeast genomes. Genome Research. 2005;15 (8 ):1034–50.16024819 [108] Sullivan PF , Meadows JR , Gazal S , Phan BN , Li X , Genereux DP , Leveraging base-pair mammalian constraint to understand genetic variation and human disease. Science. 2023;380 (6643 ):eabn2937.37104612 [109] Pollard KS , Hubisz MJ , Rosenbloom KR , Siepel A . Detection of nonneutral substitution rates on mammalian phylogenies. Genome research. 2010;20 (1 ):110–21.19858363 [110] Elnaggar A , Heinzinger M , Dallago C , Rehawi G , Wang Y , Jones L , Prottrans: Toward understanding the language of life through self-supervised learning. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2021;44 (10 ):7112–27. [111] Stärk H , Dallago C , Heinzinger M , Rost B . Light attention predicts protein location from the language of life. Bioinformatics Advances. 2021;1 (1 ):vbab035.36700108 [112] Huang YF . Unified inference of missense variant effects and gene constraints in the human genome. PLoS Genetics. 2020;16 (7 ):e1008922.32667917

PMC010xxxxxx/PMC10312943.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398194 10.21203/rs.3.rs-2925426/v1 10.21203/rs.3.rs-2925426 preprint 1 Article Proposal of a new genomic framework for categorization of pediatric acute myeloid leukemia associated with prognosis Umeda Masayuki 18 Ma Jing 18 http://orcid.org/0000-0002-1000-6698 Westover Tamara 1 Ni Yonghui 2 Song Guangchun 1 Maciaszek Jamie L. 1 http://orcid.org/0000-0002-5363-1848 Rusch Michael 3 Rahbarinia Delaram 3 Foy Scott 3 http://orcid.org/0000-0001-6996-0833 Huang Benjamin J. 4 Walsh Michael P. 1 Kumar Priyadarshini 1 Liu Yanling 3 Fan Yiping 5 http://orcid.org/0000-0002-1678-5864 Wu Gang 15 Baker Sharyn D. 6 http://orcid.org/0000-0002-6233-2145 Ma Xiaotu 3 Wang Lu 1 http://orcid.org/0000-0001-9885-3527 Rubnitz Jeffrey E. 7 http://orcid.org/0000-0002-9167-2114 Pounds Stanley 2 http://orcid.org/0000-0003-2961-6960 Klco Jeffery M. 19 1 Department of Pathology, St. Jude Children’s Research Hospital, Memphis, TN, USA. 2 Department of Biostatistics, St. Jude Children’s Research Hospital, Memphis, TN, USA. 3 Department of Computational Biology, St. Jude Children’s Research Hospital, Memphis, TN, USA. 4 Department of Pediatrics, University of California San Francisco, San Francisco, CA, US 5 Center for Applied Bioinformatics, St. Jude Children’s Research Hospital, Memphis, TN, USA. 6 Division of Pharmaceutics and Pharmacology, College of Pharmacy, Comprehensive Cancer Center, The Ohio State University, Columbus, OH, USA. 7 Department of Oncology, St. Jude Children’s Research Hospital, Memphis, TN, USA. 8 These authors contributed equally 9 Corresponding author: Contact information: Jeffery M. Klco: Mail Stop 342, Room D4047B, St. Jude Children’s Research Hospital, 262 Danny Thomas Place, Memphis, TN 38105-3678, Phone: (901) 595-6807, Fax: (901) 595-5947, jeffery.klco@stjude.org 29 5 2023 rs.3.rs-2925426https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs2925426v1.pdf Recent studies on pediatric acute myeloid leukemia (pAML) have revealed pediatric-specific driver alterations, many of which are underrepresented in the current classification schemas. To comprehensively define the genomic landscape of pAML, we systematically categorized 895 pAML into 23 molecular categories that are mutually distinct from one another, including new entities such as UBTF or BCL11B, covering 91.4% of the cohort. These molecular categories were associated with unique expression profiles and mutational patterns. For instance, molecular categories characterized by specific HOXA or HOXB expression signatures showed distinct mutation patterns of RAS pathway genes, FLT3, or WT1, suggesting shared biological mechanisms. We show that molecular categories were strongly associated with clinical outcomes using two independent cohorts, leading to the establishment of a prognostic framework for pAML based on molecular categories and minimal residual disease. Together, this comprehensive diagnostic and prognostic framework forms the basis for future classification of pAML and treatment strategies. ==== Body pmcIntroduction Pediatric acute myeloid leukemia (pAML) is characterized by aberrant clonal expansion of hematopoietic progenitors with differentiation defects1–4. Although pAML shares many clinical and pathological characteristics with adult AML, genetic differences have also been appreciated5–7. Notably, t(11;x), resulting in KMT2A rearrangements, are more common in pAML, and adult AML frequently harbors mutations in DNMT3A and splicing factor genes, while core binding factor (CBF) AMLs are common across the age spectrum5. Additionally, progress in diagnostic technologies has led to the identification of cryptic fusions of NUP988 and GLIS family9 members and complex UBTF tandem duplications10 that are enriched in pAML. Recent updates in the WHO classification11 (WHO5th) and the International Consensus Classification (ICC)12 define AMLs with KMT2A and NUP98 rearrangements with various partners as distinct disease entities. However, a substantial number of more recently discovered recurrent driver alterations in pAML continue to be categorized as “acute myeloid leukemia with other defined genetic alterations” or “AML, not otherwise specified (NOS)”, confirming the need to understand both the biological features of pAMLs with these other driver alterations and their clinical significance. Accumulation of clinical outcomes associated with gene alterations enabled the risk stratification of adult AML according to detailed mutational profiling. The ELN2022 risk stratification13 incorporated various fusions, NPM1 mutations, FLT3-ITD, and myelodysplastic syndrome-related changes, including chromosomal alterations and somatic mutations. In contrast, risk stratification for pAML is still developing, and various strategies are utilized in clinical trials14,15. This is partly due to differences in the genetic background between adult and pediatric AML5, the rarity of the disease16, and a shortage of clinical outcome studies related to genetic alterations. To clarify the genomic landscape of pAML and its association with clinical outcomes, we characterized 895 cases of pAML by transcriptome and genome profiling. These analyses resulted in 23 molecular categories, defined by mutually exclusive gene alterations and specific expression profiles, that show unique biological characteristics and mutational backgrounds. We further determined that these molecular categories have predictive value regarding clinical outcomes that can be leveraged to establish a framework for diagnosis and outcome prediction by investigating an independent clinical study cohort. Results Comprehensive genetic characterization of pAML Pediatric AML samples were collected from previously published studies5,9,10,17–26 or clinical trials at St. Jude Children’s Research Hospital, resulting in a cohort of 895 unique pAMLs either at diagnosis (n=786, 87.8%) or at relapse (n=109, 12.2%) (Fig.1A and Fig.S1A, Table.S1). This pAML cohort showed a wide age distribution at diagnosis (range: 0–23.5, median 9.3), with peaks in infancy and adolescence (Fig.S1B). We first assessed the genetic landscape of these AMLs using RNA sequencing (RNA-Seq) data to detect fusions, internal or partial tandem duplications (ITD/PTD), copy number variants (CNV), as well as single nucleotide variants (SNV) and insertions and deletions (Indel) (Fig.1A–E, Table S2–9). For 671 cases (75.0%) with either whole genome sequencing (WGS, 59.2%) or whole exome sequencing (WES, 43.7%), we also collected processed data from publications or performed de novo calling for new cases included in this study (Fig.1A and Fig.S1C). Pathogenic fusions or structural variants (SV) were identified in 627 patients (70.1%). Most of these are recurrent and class-defining in pAML (e.g., KMT2Ar: 20.2%, RUNX1::RUNX1T1: 12.3%, Fig.1B, Fig.1E, Table.S6), whereas we also found fusions recurrent in other leukemias, such as SET::NUP21427 (n=1) or SFPQ::ZFP36L228 (n=1). Mutational profiling revealed 1,947 pathogenic or likely pathogenic somatic mutations in 757 (84.6%) patients, including class-defining NPM1 (68 patients: 7.6%) and CEBPA (49 patients: 5.5%) mutations (Fig.1C, Fig.1E, Table.S7–8). The majority of mutations were in genes involved in signaling pathways (n=874), epigenetics (n=313), and transcription factors (n=440). RAS pathway mutations were most frequent, with 37.5% (336/895) having at least one RAS-related mutation and 21.1% of those (71/336) having mutations in multiple RAS pathway genes. Among CNVs, we frequently observed gains of chromosome 8 (7.2%) or chromosome 21 (6.5%) and loss of the long arm of chromosome 5 (5q-: 1.5%) or chromosome 7 (3.9%) (Fig. 1D, Fig.S1E Table.S9). Enrichment of focal deletions involving RB126 (13q14: 2.9%), ETV629 (12p13: 2.1%), NF130 (17q11: 2.0%), and TP53 (17p13: 2.0%) were also observed in this cohort. In addition, underappreciated focal gains involving AKT331 and FH32 (1q43: 3.0%) or ABCA transporters (17q24: 2.3%) were identified, suggesting their possible importance in leukemogenesis. GRIN (genomic random interval) analysis33 identified 143 genes significantly altered in the entire cohort (Fig.1E, Fig.S2A-B, Table.S10). Consistent with previous reports, RAS-related mutations or FLT3-ITD with variable variant allele frequencies (VAFs) were highly co-occurring with class-defining alterations (Fig.1E, Fig.S2). In contrast, mutations in UBTF or CBFB genes were predominantly found in cases without a defining driver alteration, as previously shown10,34, suggesting that these alterations define subgroups with distinct molecular characteristics. Based on these collective data, we classified pAMLs using current WHO and ICC systems (Fig.1E–F, Fig.S1E-F, Table.S1), and the frequencies of major classifications are consistent with cytogenetic profiles of European pediatric AML cohorts35,36. In our pAML cohort, 68.3% of cases had specified genetic alterations in WHO5th, 10.7% of cases were defined as “acute myeloid leukemia, myelodysplasia-related” (AML-MR), and the remaining cases with rare fusions or no defining alteration were classified as “acute myeloid leukemia with other defined genetic alterations” (15.9%) or by differentiation stages (3.4%). In contrast, 95.0% of adult AMLs can be classified either by specific gene alteration (67.1%) or as AML-MR (27.8%)37, emphasizing the need for a more comprehensive classification of pAML based on the unique biology of childhood AML. Molecular categories of pAML defined by mutually exclusive gene alterations We and others have shown that class-defining driver alterations are associated with specific expression patterns10,38,39 or that allele-specific and outlier expression of MECOM40,41, BCL11B42, or MNX143 by SVs can define disease subtypes. We then integrated the mutational landscape with expression profiling to define granular molecular categories for pAML (Table.S11). UMAP analysis44,45 of transcriptional data revealed tight clustering of classes defined in WHO5th, including KMT2A rearrangement (KMT2Ar), NUP98 rearrangement (NUP98r), RUNX1::RUNX1T1, CBFB::MYH11, NPM1 mutation, CEBPA mutation, and DEK::NUP214, suggesting subtype-specific expression patterns (Fig.2A, Fig.S3A-B). We noted that the clustering is also driven in part by differentiation status represented by marker gene expression, FAB (French-American-British) classification, or cellular hierarchy46 (Fig.S3C-E), contributing to heterogeneity within large categories such as KMT2Ar or NUP98r (Fig.2A, Fig.S3A, S4A). Diffusion maps47 confirmed similar patterns of clustering and differentiation status (Fig.S3A-E). Cases with NPM1 fusions or indels outside the N-terminus48 clustered with canonical NPM1 mutations, and thus we assigned them to the NPM1 category (Fig.S4A), and similarly, TBL1XR1::RARB49 to the APL (acute promyelocytic leukemia) category. For the MECOM category, we noted a case showing outlier and allele-specific expression of MECOM without any evidence of rearrangements, leaving other possible dysregulation mechanisms, such as enhancer amplification42 (Table.S12). Among the remaining cases without class-defining alterations, we found that the following alterations were also mutually exclusive with each other and thus were defined as independent molecular categories: UBTF tandem duplications10, GLIS family (GLIS2–3) fusions9, fusions of FET and ETS family genes50,51 (e.g., FUS::ERG), BCL11B structural variants42, PICALM::MLLT1052, KAT6A rearrangements53, MNX1 structural variants54, RUNX1 fusion with CBFA2T2–355,56 (RUNX1::RUNX1T1-like), and newly reported CBFB insertions (CBFB-GDXY)34 (Fig.2A–C). GATA1 fusions (e.g., MYB::GATA1) or mutations57, rearrangements involving HOX cluster genes26, and PTD of KMT2A58 could rarely co-occur with the above-mentioned category-defining alterations (Fig.2B). However, they are still predominantly found in cases without category-defining alterations, and cases were assigned to these categories only with consistent expression patterns and without previously explained driver alterations (Fig.2A–C). In contrast, defining mutations of AML-MR in WHO5th were overall rare (range: 0.1–2.1%), frequently co-occurred with other defining alterations (e.g., EZH2 in PICALM::MLLT10), and could be found in various clusters rather than as a distinct group (Fig.S3A, F), leading to its exclusion as a defining category for pAML. These 23 molecular categories with defining driver alterations and characteristic expression profiles covered 91.4% (818/895) of the pediatric AML cohort (Fig.2D, Table.S1). Clinical, mutational, and transcriptional characterization of the molecular categories Establishing updated molecular categories for pAML allowed for the investigation of clinicopathological associations. Categories with acute megakaryoblastic or erythroid leukemia (AMKL/AEL) phenotypes are clearly enriched in infants, whereas CBF leukemias and mutation-defining leukemias (e.g., UBTF, NPM1, CEBPA) were enriched in adolescents and young adults (Fig.3A, Fig.S4A-B). Notably, among KMT2A fusion partners, MLLT3 and MLLT10 were found in both monocytic AML and AMKL; however, these fusions preferentially show AMKL phenotypes in infants (Fig.S4A), suggesting that AMKL phenotypes are defined both by types of driver alterations and by the developmental stages as discussed in GATA1-driven AMKL in Down syndrome patients59 or GLIS2::CBFA2T3-driven AMKL60. Overall, however, each molecular category showed variable morphological features represented by FAB classification except categories with APL (M3) or AMKL (M7) phenotypes (Fig.3A). Likewise, conventional cytogenetics also had a limited role. Complex karyotypes, which also define AML-MR11, were frequently observed in MNX1, HOXr, and PICALM::MLLT10 categories. Additionally, as many of these category-defining alterations are cytogenetically cryptic (e.g., NUP98r8 or GLIS family9,61) or somatic mutations (e.g., CEBPA, UBTF, or GATA1), sequencing approaches are needed for the appropriate molecular diagnosis of pAML. We next explored the association between initiating driver alterations and cooperating mutations, as some cooperating mutations co-occur and act synergistically with specific driver events,5,62–64. Signaling alterations were broadly found in 66.6% of patients, whereas each mutation showed distinct patterns among molecular categories with variable VAFs (Fig.1E, Fig.3B, Table.S7). Among RAS mutations, NRAS mutations were broadly found and enriched in CBFB::MYH11, whereas KRAS mutations were enriched in KMT2Ar. Similarly, FLT3-ITD showed strong enrichment in NUP98r, NPM1, UBTF, KMT2A-PTD, DEK::NUP214, APL, and BCL11B categories, accounting for 76.3% of FLT3-ITD+ cases, whereas 74.0% of FLT3-TKD (tyrosine kinase domain) were found in KMT2Ar, NPM1, and CBF-AMLs. Similarly, WT1 mutations were specifically enriched in NUP98r, UBTF, and BCL11B and highly co-occurring with FLT3-ITD (Fig.3B, Fig.S2A). We further evaluated gene expression signatures among molecular categories. Top variable genes across the cohort enriched genes in development (e.g., HOX genes), differentiation, or inflammation (Fig.S5A, Table.S13), consistent with previous reports that the heterogeneity of AML can be partly attributed to differentiation status1,46,65. Gene set enrichment analysis (GSEA) first confirmed that expression profiles of major categories were congruent with previous reports38,66,67 (Fig.3C, Table.S14). The new categories we propose in this study show similarities and differences with clearly defined categories. For example, UBTF showed expression signatures similar to NPM1 and DEK::NUP214, while KAT6Ar was similar to KMT2Ar, suggesting shared biological mechanisms. In addition, genes involved in signaling pathways, immunity, or drug resistance showed unique enrichment across categories (Fig.3C). Weighted gene co-expression network analysis (WGCNA)68 confirmed that each category showed characteristic patterns of active gene networks associated with specific biological functions (Fig.S5B, Table.S15). Given recent adult AML-focused studies using expression profiling to uncover the associations of cellular stemness69,70 or hierarchy46,71 with prognosis or drug response, we investigated these features in our pAML dataset. We observed unique patterns of stemness and cellular hierarchy scores in each category. Molecular categories known to have a good prognosis (RUNX1::RUNX1T135,36, CBFB::MYH1135,36, and CEBPA72) tended to have high GMP scores (median >0.20) (Fig.3D, Fig.S5C), with the exception of the low GMP scores (median: 0.047) and mid-high stemness-related scores in NPM1. Also, KMT2Ar, associated with poor prognosis35,73,74, showed low stemness-related scores and variable differentiation-related scores. Various prognostic scores (e.g., LSC1769, iScore65) also correlated with molecular categories (Fig.S5D). These data collectively demonstrate that molecular categories are associated with unique pathophysiological characteristics, whose clinical association needs to be assessed based on detailed molecular categories. Superfamilies defined by HOX gene expression profiles These molecular categories also showed inter-categorical similarities, forming large clusters of AMKL/AEL, immature AML, CBF leukemias, CEBPA, and two clusters demarcated by HOXA and HOXB cluster gene expression (Fig.2A, Fig.4A–B). The cluster with high HOXA gene expression and low HOXB expression consisted mainly of KMT2Ar and KAT6Ar (herein referred to as HOXA categories), and the other cluster characterized by high expression of both HOXA and HOXB genes included NPM1, NUP98r, UBTF, KMT2A-PTD, and DEK::NUP214 (HOXB group) (Fig.4A, Fig.S6A). Overall, HOXA and HOXB groups, not including those with AMKL features, account for 17.8% and 22.4% of the cohort, respectively. Differential gene expression analyses revealed that pAMLs in HOXB group had high expression of stemness-related genes (PRDM16 and NKX2–3) or differentiation genes (CD96 and WT1) (Fig.4C–D, Table.S16). In contrast, HOXA group cases showed high expression of monocyte or signaling-related genes. GRIN analysis also revealed striking differences in mutational patterns between HOXA and HOXB groups (Fig.4E–F, Table.S17). FLT3 was significantly altered in both HOX groups but with different mutation types; FLT3-TKD was dominant in HOXA group, and FLT3-ITD was prevalent in HOXB group, accounting for 67.3% of FLT3-ITD+ patients (Fig.4F, Fig.S6B). WT1 mutations were preferentially found in HOXB group (56.6%). FLT3-ITD75 and WT1 mutations17,76 have been associated with poor prognosis: however, these data suggest that these mutations highly confound with specific driver alterations that converge on a common expression signature. KRAS mutations were strongly associated with HOXA group and rare in HOXB group (22.2% and 4.2%, respectively). In comparison, NRAS mutations were prevalent in both HOXA and HOXB group (22.7% and 19.3%) (Fig.4F); however, among NRAS mutations, NRAS p.G12–13 mutations were comparable in both categories, while NRAS p.Q61 mutations were more frequent in HOXA group (Fig.4E, Table.S7). It is well-established that each RAS mutation has preferential distribution among cancer subtypes77. Expression levels and differences in the downstream signaling are postulated as the possible mechanisms, and similarly, between FLT3-ITD and TKD78, while at the RNA levels, these genes were homogenously expressed in this pAML cohort (Fig.S6B). These molecular category-dependent mutational patterns may reflect different signal dependencies, potentially offering less toxic targeted therapies guided by these biological insights. Along with the global distinction between HOXA and HOXB groups, we also noted heterogeneity within each HOX cluster. The HOXA cluster consisted of subclusters characterized by MECOM or LAMP5 expression (Fig.S7A-C, Table.S18), harboring most KMT2Ar cases (120/181; 66.3%). Notably, the larger subcluster expressed XAGE1 family genes specifically (Fig.S7B-C), which encode members of testis-specific proteins postulated as therapeutic targets in various tumor types79. Also, the remaining KMT2Ar cases were clustered with other categories with HOXB expression or AMKL less frequently (Fig.S4A and S7A). These clustering patterns were associated with fusion partners (e.g., KMT2A::ELL in the HOXB cluster) or age (younger age with AMKL and outliers), but the associations were not exclusive (Fig.S7D-E). Among KMT2Ar, fusion partners and MECOM expression have been reported to be prognostic73,74; however, our data suggest considerable heterogeneity in expression patterns not explained by only fusion partners or MECOM expression. The HOXB cluster showed similar heterogeneity represented by cellular hierarchies (Fig.S7F-G). These heterogeneities were occasionally associated with molecular categories or somatic mutations but were not exclusive (Fig.S7G-H), with possible factors, including cell-extrinsic factors65,80 to be further investigated. Molecular basis of AML without defining gene alterations Seventy-seven Unclassified cases remained after assignments into these 23 molecular categories. Twenty-one cases had recurrent driver alterations previously reported in the literature, whose sample sizes are insufficient to assess the biological characteristics (Fig.5A, Table.S19). These included rare or novel in-frame RUNX1 fusions (n=2: USP4281, n=1: EVX182 and ZEB2) and rare MLLT10 fusions (n=1: DDX3X83, TEC84, and MAP2K226), with the possibility of further categorization in a larger cohort. Also, in addition to hi-allelic burden JAK2 p.V617F mutation (n=1), we found candidate driver somatic mutations of MLLT1 p.C119>SPAR (n=1) and H3F3A p.K28M (n=1) in pAML with HOX gene expression (Fig.5A, Fig.S8A, Table.S7). These mutations resemble recurrent mutations in other pediatric cancer types with HOX gene expression and immature phenotypes (MLLT1 p.C118QPPG in Wilms tumor85 or H3F3A p.H28M in high-grade glioma86), postulating a shared mechanism of tumorigenesis among these pediatric neoplasms. This series of genomic characterization did not find any pathogenic alterations in 9 cases of the remaining 56 Unclassified cases, partly attributed to the lack of WGS data for 8 of these cases. The rest had at least one pathogenic but not subtype-defining alteration enriched in ETV6, RUNX1, TP53, and RAS pathway genes (Fig.5B–C, Table.S19–20), in addition to complex karyotypes or monosomy 7 (Fig. 3A). Of note, complex karyotypes or NRAS mutations were found broadly in various clusters, whereas non-canonical ETV6 and RUNX1 alterations were found preferentially in clusters comprised of FAB M0–1 cases. These clusters are associated with immature or T cell-like signatures (Fig.5D, Fig.S8B, Table.S21), consistent with a new entity of AMTL87. Although various ETV6 or RUNX1 alterations can co-occur with other defining alterations or be class-defining (e.g., RUNX1::RUNX1T1), the mutations in the Unclassified category are commonly loss-of-function (Fig.5E). Given that germline mutations of RUNX1 or ETV6 are associated with leukemia with incomplete penetrance88,89, these data suggest somatic alterations of these genes also require additional mutations for leukemia development, which may cooperatively define the immature leukemic phenotypes in the absence of other defining alterations. Further accumulation of genomic data and experimental models will be necessary to understand the stepwise leukemogenesis and the clinicopathological features of immature pAML with these mutations. Clinical association of molecular categories Although the association between KMT2Ar or NUP98r with poor outcomes is well-appreciated, the clinical associations of new molecular categories have been discussed only in separate studies10,26. To address this deficiency and translate them into a clinical framework, we investigated the outcomes of these molecular categories using the COG AAML1031 clinical study15 (n=1,034, Table.S22). Analyses of the AAML1031 RNA-Seq data using the same pipeline revealed similar clustering of molecular categories (Fig. 6A) and the overall category frequencies (Fig. 6B). The AAML1031 cohort confirmed the association of molecular categories with age and FLT3-ITD status (Fig.6C) and showed variable MRD (minimal residual disease) positivity among molecular categories. We first assessed the clinical association of the categories using recursive partitioning models90 for censored event time data, which revealed three groups with distinctive prognoses (Fig.6D, Fig.S9A-B). The grouping of major categories aligns with previous reports (e.g., RUNX1::RUNX1T1 (n=141), CBFB::MYH11 (n=102), and CEBPA (n=63) in low-risk35,36,72) except DEK::NUP214 (n=17) in low-risk, historically regarded as high-risk36,91. We confirmed the known association of GLISr61 (n=20), MECOM92 (n=11), PICALM::MLLT1093 (n=8), and KAT6Ar93 (n=7) with poor outcomes, while new categories of MNX1 (n=4), RUNX1::RUNX1T1-like (n=4), and CBFB-GDXY (n=4) were assigned to low-risk. Univariate analyses of other risk factors revealed that age and FLT3-ITD were not prognostic, whereas MRD positivity and cellular hierarchy scores (GMP-like, cDC-like, and cycling LSPC) were associated with the overall survival (Fig.6E, Fig.S9C-D, Table.S23). A Cox proportional hazards model using risk groups and prognostic factors showed that hierarchy scores did not significantly contribute to prognosis (Fig.S9E), whereas risk groups and MRD positivity were independently prognostic (Table.S23). These data led us to establish a simple predictive framework solely based on molecular categories and MRD positivity, resulting in six risk strata with granular outcome prediction (Fig.6F and Fig.S9F-G), whose prognostic values were validated using the AML08 trial14 (n=211, Fig.S9H-J, Table.S24). With these transcriptional and outcome data, we also investigated the clinical association of transcriptional heterogeneity within major molecular categories. Among KMT2Ar, fusion partners or MECOM expression73,74 also confound in the AAML1031 cohort (Fig.6G–H). Cox hazard models showed that both fusion partners and expression clusters are prognostic (P =0.00052 and 0.0015, respectively), with fusions with SEPTIN family and ELL or immature expression patterns associated with favorable outcomes (Fig.6I). Bootstrapping showed that the association of fusion partners or expression clusters with prognosis did not significantly differ (difference in C-index of 95% bootstrap interval for fusions and expression clusters: −0.025–0.093). Although HOXB categories of NUP98r, NPM1, and UBTF also showed heterogeneity of expression patterns, their outcomes were not associated with UMAP clusters or FLT3-ITD status (Fig.S10), and further mutational profiling outside the clinical standard (e.g., WT1 mutations) may be required to risk-stratify within these categories further. Discussion In addition to known enrichment of chromosomal events like t(11,x) in pediatric patients with AML, advances in sequencing technology have identified additional pediatric-specific driver alterations9,10,34. This prompted us to comprehensively investigate the increasingly complex genomic landscape of pAML in the context of the latest classification systems for hematological malignancies (WHO5th,11 and ICC12) and to develop a pAML-focused categorization schema based on the unique disease biology of childhood AML. In this study, we systematically categorized our pAML cohort of 895 patients using an RNA-Seq-based approach, resulting in 23 molecular categories defined by mutually-exclusive driver alterations, covering 91.4% of the entire cohort. Of these 23 categories, 12 are not currently defined by WHO5th. This includes common categories like UBTF, GLISr and GATA1, which were otherwise categorized as “acute myeloid leukemia, myelodysplasia-related” or “acute myeloid leukemia, other defined gene alteration” in the current WHO classification. Notably, myelodysplasia-related (MR) mutations or chromosomal alterations often co-occur with many pAML category-defining alterations and override them in WHO5th despite these MR alterations not driving consistent patterns of gene expression. Considering that the current classification systems are mainly based on evidence from adult AML94,95 and pediatric myelodysplasia syndrome (MDS) is rare22, we propose an alternative framework for pAML to better reflect the unique genetic and clinical landscape. These molecular categories show unique expression and mutational profiles, whereas some categories also show critical similarities, which can suggest common molecular mechanisms and potentially therapeutic options. In particular, we noticed two large clusters characterized by unique HOXA-B expression profiles. Molecular categories with HOXB signatures were strongly associated with FLT3-ITD and WT1 mutations, whereas those with HOXA signatures were associated with RAS mutations. Considering that AMLs with KMT2Ar, NUP98r, and NPM1 have been shown to be dependent on KMT2A/Menin96–98, and that a Menin inhibitor (SNDX-5613) targeting KMT2Ar and NPM1 AML is in a clinical trial99 (NCT04065399), our data suggest that other subtypes marked by HOX expression, such as UBTF or DEK::NUP214, may also be candidates for Menin inhibitors. Thus, this approach may cover nearly 50% of pediatric AML. Also, the high frequency of FLT3-ITD in categories with HOXB expression implies that FLT3 signaling is closely related to biology and that a data-driven implementation of FLT3 inhibitors to HOXB subtypes can be effective. Some cases without category-defining alterations could be characterized by rare fusion or mutations, which need further evidence to establish as a disease entity, including MLLT1 and H3F3A mutations that are frequent and class-defining in Wilms tumor85 and glioma86, respectively. Considering that AML and Ewing sarcoma also share ETS family fusions51 (e.g., EWSR1::ERG), it would be intriguing to incorporate knowledge of these solid tumors to understand the biology behind pAML with these rare alterations. Also, enrichment of RUNX1 or ETV6 loss of function alterations in immature AML implies that these can be class-defining in the absence of other defining alterations and likely with specific cooperating mutations. These findings further suggest a continuum with other immature leukemias, such as early T-cell precursor (ETP)-ALL and mixed phenotype acute leukemias (T/My) which have similar mutational features82,100. We further investigated the clinical outcomes of these molecular categories using two independent cohorts --the COG AAML1031 study and the St. Jude AML08 study. Using both cohorts, we show a strong association of new molecular categories with outcomes (e.g., PICALM::MLLT10 and KAT6Ar as high-risk, CBFB-GDXY as low-risk). These analyses also revealed confounding variables like molecular categories and known prognostic factors like FLT3-ITD status or cellular hierarchy scores. With this comprehensive profiling recognizing new pAML subtypes, we established a simple risk stratification using molecular categories and MRD. This strategy, however, heavily relies on the analysis of next-generation sequencing data. While the WHO classification requires targeted sequencing or WGS, we propose a diagnostic pipeline utilizing RNA-Seq which is highly sensitive for canonical and cryptic fusion calling, allows for categorization based on gene expression signatures, including outlier and allele-specific expression (MECOM, BCL11B, and MNX1), and provides limited but sufficiently sensitive mutation calling to enable our comprehensive molecular categorization strategy to newly diagnosed pAML.This approach is favored over current commercial panels commonly used for pAML, which either lack coverage of all the defining genes (e.g., UBTF) or are not suitable to detect complex structural variations that drive aberrant expression of MECOM or BCL11B. Given that clinical sequencing is not readily available globally and these molecular analyses require substantial expertise, robust and easy pipelines are needed for future and broad application of this framework for pAML in the general clinical setting. Online Methods Subject cohorts and sample details. Tumor samples from patients with AML from the St. Jude Children’s Research Hospital tissue resource core facility were obtained with written informed consent using a protocol approved by the St. Jude Children’s Research Hospital institutional review board (IRB). Studies were conducted in accordance with the International Ethical Guidelines for Biomedical Research Involving Human Subjects. Samples for RNA sequencing (RNA-Seq: n=221), whole genome sequencing (WGS: n=54), and whole exome sequencing (WES: n=6) are newly sequenced in this study, and the rest of the data were obtained from previous publications5,9,10,17–26 or public databases (see details in Data availability and Table S1). For samples with multiple available data points, we included one representative time point with a high tumor purity and good RNA-Seq data quality. Cases were assigned to current WHO and ICC by board-certified hematopathologists (PK and JMK). Sample processing, library preparation, and sequencing. For newly sequenced samples with low tumor purity (below 60%), the leukemic cell population was enriched either by flow cytometric sorting or T cell depletion by magnetic beads (EasySep Human CD3 Positive Selection Kit II, 17851, StemCell Technologies). For flow cytometric sorting, CD45dimCD33dim positive population was sorted using anti-CD45 PerCP-Cyanine5.5 (eBioscience cat# 8045–9459-120) and anti-CD33 APC (eBioscience cat# 17–0338-42). CD34 gating using anti-CD34 PE (Beckman cat# IM1459U) was added depending on the positivity of each patient sample. Enrichment of the tumor population was confirmed flow cytometric analysis of the post-sorting samples (generally > 90%). Libraries were constructed using the TruSeq Stranded Total RNA Kit, with Ribozero Gold (20020598, Illumina) for RNA-Seq, the TruSeq DNA PCR-Free Library Prep Kit (20015963, Illumina) for WGS, and the TruSeq Exome Kit v1 (20020614, Illumina) for WES according to the manufacturer’s instructions. After library quality and quantity assessment, samples were sequenced on HiSeq2000 or 2500 (Illumina, RRID:SCR_020132, RRID:SCR_016383) instruments with paired-end (2 × 101 bp, 2 × 126 bp, or 2 × 151 bp) sequencing using TruSeq SBS Kit v3-HS (FC-401–3001, Illumina) or TruSeq Rapid SBS Kit (FC-402–4023, Illumina). RNA-Seq mapping, fusion detection, and large-scale copy number variant calling. RNA reads from newly sequenced samples and from publications were mapped to the GRCh37/hg19 human genome assembly using the StrongARM pipeline101. Chimeric fusion detection was carried out using CICERO102 (v0.3.0). For the cases with only RNA-Seq data, RNAseqCNV103 (v1.2.1) was used to call large-scale copy number variants (CNV). Somatic mutation calling from RNA-Seq. We initially performed RNA-based variant calling methods for detecting Single-nucleotide variants (SNV) and Insertions and deletions (Indel). Specifically, we applied the following approach to simultaneously account for germline polymorphisms (without germline control) and sequencing artifacts specific to RNA-Seq on a panel of 86 predefined genes previously reported to be significantly mutated in pediatric AML5 and myelodysplastic syndrome (MDS; Table.S5). Briefly, candidate SNVs/Indels were called by Bambino104 (v1.07) or RNAindel105,106 (v3.0.4), annotated by VEP107 (v95), and in turn, classified for putative pathogenicity with PeCanPie/MedalCeremony108. Candidate variants with putative pathogenicity were considered germline or artifacts if present in >5% of the cases. Candidate variants were further filtered if the number of supporting reads was ≤5 or if the variant allele fraction (VAF) was ≤5%. UBTF tandem duplications were detected by soft-clipped read counting in addition to CICERO and RNAindel as we previously described10. Whole genome and whole exome sequencing data analysis. The previous genomic lesion calls for the cases (WGS; n=394, WES; n=284) from published studies5,9,10,17,19–21,24,26 were collected from their respective publications. For the unpublished cases with DNA data (WGS; n=136, WES; n=107), DNA reads were mapped using BWA109,110(WGS: v0.7.15-r1140 and v0.5.9-r26-dev; WES: v0.5.9-r26-dev and v0.5.9, RRID:SCR_010910) to the GRCh37/hg19 human genome assembly. Aligned files were merged, sorted, and de-duplicated using Picard tools 1.65 (broadinstitute.github.io/picard/). SNVs and Indels were called using Bambino104. Candidate SNVs and Indels were similarly classified for putative pathogenicity with PeCanPie/MedalCeremony as in somatic mutation calling from RNA-Seq. The counting of somatic mutations included all the pathogenic or likely pathogenic mutations detected by WGS, whereas mutation detection from cases with only RNA-Seq data is limited to the 86 preselected genes. Structural variations (SV) were analyzed using CREST111 (v1.0), and CNVs were analyzed using CONSERTING112 on the WGS data. CNVs were also called on cases with only WES DNA data using the following methods. Briefly, Samtools113 mpileup command was used to generate a mpileup file from matched germline and tumor BAM files with duplicates removed. If a matched germline was not available, a high-quality normal sample was used to pair with the tumor sample. VarScan114 (v2.3.5) was then used to take the mpileup file to call somatic CNVs after adjusting for normal/tumor sample read coverage depth and GC content. Circular Binary Segmentation algorithm115 implemented in the DNAcopy R package (v1.52.0) was used to identify the candidate CNVs for each sample. B-allele frequency info was also used to assess allelic imbalance. GRIN analysis for significantly mutated genes. For the 895 AML cases, the genomic random interval (GRIN; v2.0) model33 was used to evaluate the statistical significance of the number of subjects with each type of lesion: fusions, CNVs (amplifications and deletions), copy neutral loss of heterozygosity (CN-LOH), SNV/indels, and tandem duplications in each gene. For each type of lesion, robust false discovery estimates were computed from P values using Storey’s q value116 with the Pounds-Cheng estimator of the proportion of hypothesis tests with a true null hypothesis117. FDR cutoff of <0.05 for the number of subjects with any one type of lesion overlapping the gene locus was used to obtain significantly mutated genes, where we focused on protein-coding genes and genes that are known or likely to be pathogenic in leukemia. We also excluded genes that are part of a large chromosomal gain, loss, or CN-LOH but not the target of the CNVs based on the GISTIC (Genomic Identification of Significant Targets in Cancer) analysis. Subgroup GRIN analyses for HOXA categories (n=167), HOXB categories (n=211) categories and the Unclassified category (n=77) were also performed using the same methods. GISTIC analysis for significant recurring copy-number alterations. We used GISTIC (v2.0.23, RRID:SCR_000151) 118,119 to identify genomic regions that are significantly amplified or deleted across our 895 samples. Each aberration was assigned a G-score that considered the amplitude of the aberration as well as the frequency of its occurrence across samples. False discovery rate q values were then calculated for the aberrant regions, and regions with q values ≤0.25 were considered significant. A “peak region” was identified for each significant region with the greatest amplitude and frequency of alteration. In addition, a “wide peak” was determined using a leave-one-out algorithm to allow for errors in the boundaries in a single sample. The “wide peak” boundaries were more robust for identifying the most likely gene targets in the region. Each significantly aberrant region was also tested to determine whether it resulted primarily from broad or focal events (A broad event was set as >90% of the chromosome arm, whereas a focal event was ≤90%). Allele-specific expression estimation for MNX1, BCL11B, and MECOM categories. For cases with both WGS and RNA-Seq available, SNP (single-nucleotide polymorphism) markers in the respective gene locus with ≥10x coverage that are heterozygous (defined as 0.2≤VAF≤0.8) in WGS and also present in RNA-Seq were extracted and a two-sided binomial test (with probability of success P=0.5) was performed on each marker for allelic imbalance in RNA expression. The median of binomial P values was used to assess allele-specific expression (ASE). For RNA-Seq only cases, SNP markers in the respective gene locus with ≥10x coverage and allelic imbalance (VAF≤0.2 or VAF≥0.8) support ASE. Germline variant curation methods. We focused on 15 candidate genes relevant to AML that define specific categories in WHO5th (Table.S25) and scanned for germline mutations in the cases with WGS or WES germline BAM files available (WGS n=367; WES n=354). For cases with germline mutation called in previously published studies10,22, we collected calls from the studies. For the remaining cases, the putative germline variants were called using Bambino104, annotated by VEP107, and in turn, classified for putative pathogenicity with PeCanPie/MedalCeremony108. We then used the following criteria to obtain the candidate germline variants: gnomAD (v2.1.1, RRID: SCR_014964)120 population allele frequency ≤0.001; read coverage SNV≥20 and Indel≥15; for SNV, variant allele frequency between 0.2 and 0.8; for Indel, ≥3 reads supporting the alternative allele. All candidate germline variants were comprehensively reviewed and classified as pathogenic, likely pathogenic, of uncertain significance, likely benign, or benign based on recommendations from the American College of Medical Genetics and Genomics and the Association for Molecular Pathology121 and the Clinical Genome Resource122–125 by a variant scientist (JLM). Gene expression data summarization, batch correction, dimension reduction, and clustering. Reads from aligned RNA-Seq BAM files were assigned to genes and counted using HTSeq126 (v0.11.2, RRID: SCR_005514) with the GENCODE (RRID: SCR_014966) human release 19 gene annotation. The gene count matrix was generated, and for a gene to be considered as expressed, we required that at least 5 samples should have ≥10 read counts per million reads sequenced. The count data were transformed to log2-counts per million (log2CPM) using Voom127 available from R package Limma128(v3.50.3, RRID: SCR_010943). We corrected for library strand (stranded total RNA vs. unstranded mRNA) and batch effect between St. Jude and TARGET cases using the ComBat method available from R package SVA129(v3.42.0, RRID:SCR_012836). The R package Seurat130–133(v4.1.0, RRID:SCR_016341) was used for dimension reduction and sample clustering. Briefly, the top variable genes were selected using the “vst” method. The expression data were then scaled, and PCA (Principal Component Analysis) was performed on the scaled data using the top 320 variable genes. Dimension reduction was performed using UMAP44,45 (Uniform Manifold Approximation and Projection, RRID:SCR_018217) with the top 100 principal components, n_neighbors =15 and min_dist =0.2. Samples were clustered using the top 100 principal components by first constructing a K nearest-neighbor graph and then iteratively optimizing the modularity using Louvain algorithm with resolution=3.5. Dimension reduction was also performed by Diffusion maps 47,134 algorithm available in the R package destiny135 (v 3.10.0) using the same 320 genes with the default setting except for number of principal components n_pcs=50. Differential gene expression analysis was performed by Limma128, and we set Log2 CPM = 0 if it is < 0 based on the Log2 CPM data distribution. P values were adjusted by the Benjamini-Hochberg method to calculate the false discovery rate (FDR) using R function p.adjust. Genes with absolute fold change > 2 and FDR < 0.05 were regarded as significantly differentially expressed. Gene Set Enrichment Analysis (GSEA)136 was performed by GSEA (v4.2.3, RRID: SCR_003199) using MSigDB gene sets c2.all (v7.5.1), comparing each category with the rest of the categories. Permutations were done 1000 times among gene sets with sizes between 15 and 1500 genes. Normalized enrichment scores (NES) and FDR for arbitrary gene sets representing hematopoiesis, leukemia phenotype, biological processes, and drug responses were extracted from the entire results and shown in a heatmap. Weighted correlation network analysis (WGCNA) was done by R package WGCNA68 (v1.70–3, RRID:SCR_003302) using top 2000 variable genes and default setting with the exception of block-wide module calculation with reassignThreshold = 0 and mergeCutHeight = 0.25. Functional annotation of top 320 variable genes, differentially expressed genes, and genes in WGCNA modules were performed with DAVID137 (v6.8), and results for GO term, biological process (GOTERM_BP_DIRECT) were exported. Inference of cellular hierarchy by CIBERSORT138 (RRID:SCR_016955) was performed by the web interface of CIBERSORTx in absolute mode with S-mode batch correction without a permutation as previously reported46. TPM values were used as input data, and Malignant Signature Matrix and Malignant Single Cell Reference Samples were used as previously described46, and the malignant cell populations were normalized to 1 to calculate the relative fraction scores, which were shown in UMAP space or violin plots. Prognostic scores of LSC1769, pLSC670, ADE-RS139, and iScore65 were calculated as reportedly. Hierarchical clustering (RRID:SCR_014673) of expression data, mutual-exclusivity matrix, and GSEA scores were performed using the Euclidian distance and Ward method. Statistical test. For discrete values of the molecular category and the mutation frequency in cohorts, statistical significance and mutual exclusivity were assessed by two-sided Fisher’s exact test and Pearson’s correlation. Adjustment of multiple testing was performed by the Benjamini-Hochberg method using p.adjust function on R when appropriate. For survival data, decision trees were established by a recursive partitioning method using R library rpart90 (v4.1.19, RRID:SCR_021777). Kaplan–Meier curves for the probability of overall survival (OS) and event-free survival (EFS) were constructed using R package survival (v3.3–1, RRID:SCR_021137). Events in the probability of EFS calculations were defined as relapse, death in remission by any cause, and non-response, which was included as an event at the date of diagnosis. The Cox proportional hazards model was used to calculate the statistical significance of individual prognostic factors by univariate analyses first, and significant factors were included in a multivariate analysis. Clinical association of the molecular categories were first assessed using the AAML1031 study (n=1034), and the results were validated using the AML08 cohort (n=211, independent from the AAML1031, a part of this study cohort). R statistical environment (R v4.0.2, RRID:SCR_001905) was used for statistical tests. Visualization. Mutational heatmaps and mutations on individual genes were visualized using ProteinPaint (https://proteinpaint.stjude.org/). Heatmaps of expression data, mutual-exclusivity matrix, and GSEA scores were created by pheatmap function of R library pheatmap (v1.0.12, RRID:SCR_016418). Other data visualizations were performed by ggplot function of R library ggplot2 (v3.3.6, RRID:SCR_014601), survminer (v0.4.9), and base plot function in R statistical environment. Figures are incorporated and edited using Adobe Illustrator (2021, RRID:SCR_010279). Annotation of genes in mutational heatmaps depend on common knowledges, and the definition of RAS pathway genes included causative genes of Noonan or Noonan-like syndrome140 (NRAS, KRAS, PTPN11, NF1, CBL, LZTR1, RIT1, BRAF, SOS1, and HRAS). Data availability. The genomic data and expression data newly generated in this study (RNA-Seq: n=221, WGS: n=54, WES: n=6) have been deposited in the European Genome-Phenome Archive (EGA, RRID:SCR_004944), which is hosted by the European Bioinformatics Institute (EBI), under accession EGAS00001005760. For the remaining RNA-Seq data for 588 cases, 401 are St. Jude cases, of which 274 cases with data from the publications are available either on EGA or St. Jude Cloud9,10,18,20–24,26 or on the original publication25. For the other 127 published cases19, we downloaded the BAM files from EGA (EGAS00001004701). Unpublished data (n=86) are also available on St. Jude Cloud under the PCGP study (https://permalinks.stjude.cloud/permalinks/PCGP, n=8) and the RTCG study (https://platform.stjude.cloud/data/cohorts?dataset_accession=SJC-DS-1007, n=78). Of the remaining WGS data for 394 cases, 207 are St. Jude cases, of which 115 cases with data from the original publications9,10,20,21,24,26 are available on either EGA or St. Jude Cloud, and for the other 92 published cases19, we downloaded the BAM files from EGA (EGAS00001004701). Unpublished WGS data (n=82) are also available on St. Jude Cloud under the RTCG study. For the remaining WES data for 314 cases, 275 are St. Jude cases, of which 155 with data from the original publications9,10,18,20–24,26 are available either on St. Jude Cloud or EGA, and for the other 120 published cases19, we downloaded the BAM files from EGA (EGAS00001004701). Unpublished WES data (n=101) are also available on St. Jude Cloud under the PCGP study (n=2) and the RTCG study (n=99). The data generated by the TARGET initiative5,17 (n=187) is also available under accession phs000218 (TARGET-AML) and phs000465 (TARGET sub-study, data is available as a part of phs000218), managed by the NCI. Information about TARGET can be found at http://ocg.cancer.gov/programs/target. Other data generated in this study are available in the Supplemental tables or upon request to the corresponding author. Acknowledgements We thank all the patients and their families at St. Jude Children’s Research Hospital (SJCRH) for their contribution of the biological specimens used in this study. We also thank the Biorepository, the Flow Cytometry and Cell Sorting Core, and the Hartwell Center for Bioinformatics and Biotechnology at SJCRH for their essential services. This work was funded by the American Lebanese and Syrian Associated Charities of St. Jude Children’s Research Hospital and grants from the NIH (P30 CA021765, Cancer Center Support Grant and a Developmental Fund Award, to J.M. Klco and X. Ma). The content, however, does not necessarily represent the official views of the NIH and is solely the responsibility of the authors. This work was also supported in part by the Fund for Innovation in Cancer Informatics (www.the-ici-fund.org, to X. Ma and J.M. Klco). J.M. Klco holds a Career Award for Medical Scientists from the Burroughs Wellcome Fund and is a previous recipient of the V Foundation Scholar Award (Pediatric). Figure 1: Comprehensive genetic characterization of pediatric acute myeloid leukemia (pAML) A. Study cohort of pediatric AML (n=895) and study design B. Recurrent pathogenic or likely pathogenic in-frame fusions (blue) and structural variants (SV: gray) detected in the entire cohort. Fusions included only in-frame fusions, and SVs included out-of-frame fusions resulting loss of the C-terminus of the protein and alterations detected from WGS data using CREST. C. Recurrent pathogenic or likely pathogenic somatic mutations. Colors represent types of mutations. Bars in Fig.1B–C represent the total number of alterations in the cohort. D. Results of GISTIC analysis for focal chromosomal events (shorter than 90% of the chromosome arm). The left panel shows the enrichment of focal gains, and the right panel shows the enrichment of focal loses. Green lines show a significance threshold for q values (0.25). Representative genes in enriched regions are highlighted. E. The genomic landscape and the WHO classification of pAML. Representative genes from GRIN analysis or defining alterations are shown. F. Summary of the WHO classification of the entire cohort. Figure 2: Molecular categories defined by mutually exclusive gene alterations A. UMAP plot of the entire pAML cohort (n=895) and cord blood CD34+ cells (normal controls: n=5) using top 320 variable genes. The colors of each dot denote the molecular categories of the samples. Representative category names are shown, and large clusters are highlighted in circles. B. A heatmap showing frequencies of defining gene alterations represented by the color. Statistical significance was assessed by two-sided Fisher’s exact test to calculate p values of co-occurrence, followed by the Benjamini-Hochberg adjustment for multiple testing to calculate q values (*P<0.05, **q<0.05). C. Definition of molecular categories and diagnostic flow. Molecular categories not defined in WHO5th are highlighted in red. D. A ribbon plot showing the association between WHO classification and molecular categories. Colors represent molecular categories of samples Figure 3: Clinical and molecular profiles of molecular categories A. Clinical background of molecular categories. Upper row. Violin plots showing age distribution within each molecular category. Large dots and bars represent the median and the range of 2.5~97.5 percentiles, respectively. Small dots represent individual patients’ ages. Bottom row. Frequency of FAB and karyotype in individual categories. B. Mutational heatmap showing mutation frequencies in each molecular category. The color of each panel represents the frequency of a mutation in each molecular category, and the statistical significance was assessed by two-sided Fisher’s exact test to calculate p values of co-occurrence followed by the Benjamini-Hochberg adjustment for multiple testing to calculate q values (*P<0.05, **q<0.05 after the adjustment). Bars on the top panel show the frequency of mutations in the entire cohort, and the colors represent mutation types. Molecular categories are clustered according to Ward clustering using the Euclidean distance of the frequency matrix. Genes are grouped according to the functional annotations. C. A heatmap showing normalized enrichment scores (NES) and false discovery rates (FDR) of gene set enrichment analysis (GSEA) of each molecular category. Colors denote NES, and asterisks show FDR (*FDR<0.05, **FDR<0.01, ***FDR<0.001) D. Violin plots showing cellular hierarchy scores in each molecular category inferred by CIBERSORT. Lines of the box represent 25% quantile, median, and 75% quantile. The upper whisker represents the higher value of maxima or 1.5 × interquartile range (IQR), and the lower whisker represents the lower value of minima or 1.5 × interquartile range (IQR). Dots show outliers. LSPC stands for leukemic stem and progenitor cells. Figure 4: Categories demarcated by HOXA and HOXB cluster expression A. UMAP plot showing groups of molecular categories based on UMAP clustering and HOX cluster gene expression profiles. B. HOXA9 and HOXB5 expression on UMAP plot. The dot colors represent the relative expression of the genes. C. A volcano plot showing differentially expressed genes (DEG) between HOXA and HOXB groups. Genes with absolute fold change > 2 and FDR < 0.05 are considered DEGs. Representative gene names are shown. D. GO term analyses of genes with significantly high expression in each HOX group by DAVID. Bars represent logged FDR. E. Plots showing results of GRIN analyses in HOXA group (horizontal axis) and HOXB group (vertical axis). Genes with FDR<0.1 in either HOXA or HOXB groups are shown. Red or blue dots show genes enriched only in either HOXA or HOXB groups, respectively. The dotted lines represent thresholds for statistical significance (FDR<0.05). F. A mutational heatmap comparing patterns between HOXA and HOXB groups. Colors represent mutation types, and molecular categories are annotated on the top. Bar plots on the right show frequencies of mutations in HOXA and HOXB groups. Statistical significance of GRIN analysis in HOXA and HOXB groups (*FDR<0.05) and two-sided Fisher’s exact test between HOXA and HOXB groups (*P<0.05, **q<0.05 after the Benjamini-Hochberg adjustment) are also shown. GRIN results for FLT3 are for the entire gene, while Fisher’s tests were performed separately for ITD, TKD, and non-TKD mutations. Figure 5: Characterization of cases without category-defining alterations A. UMAP plot showing cases without category-defining alterations. Red dots represent cases with rare recurrent gene alterations, blue dots represent cases for which no pathogenic alteration was found, and black dots represent cases with at least one gene alteration not defining the phenotype. B. A plot showing the FDR of GRIN analysis for the Unclassified category (horizontal axis) and relative enrichment of the alteration in the Unclassified category (vertical axis). The dot sizes and colors denote the Unclassified category’s frequency, which included fusions, mutations, copy number loss and gain, and copy-neutral heterozygosity. C. A mutational heatmap of the Unclassified cases, including complex karyotypes and monosomy 7. Patients’ age, FAB, and UMAP clustering are annotated on the top. Colors represent mutation types. D. UMAP plots showing FAB (top-left), CD34 or CD3D expression (bottom-left), and cases with ETV6 alterations (top-right) and RUNX1 alteration (bottom-right). E. Patterns of alteration in ETV6 (left) and RUNX1 (right). Category-defining fusions are shown in the top row, alterations co-occurring with category-defining alterations in the middle row, and alterations in the Unclassified category in the bottom row. Bars represent a relative fraction of alteration in each group; the colors denote the alteration types. Figure 6: Clinical association of molecular categories A. UMAP plot of transcriptome data of the AAML1031 cohort (n=1,034) using top340 variable genes. The dot colors denote molecular categories assigned to the samples according to genomic profiling using the same pipeline as this study cohort. Representative category names are shown, and large clusters are highlighted in circles. B. Frequency of molecular categories in the AAML1031 cohort. Asterisks denote the statistical significance of the frequency of each category assessed by two-sided Fisher’s exact test followed by the Benjamini-Hochberg adjustment (*P<0.05, **q<0.05, blue: fewer and black: more in the AAML1031). C. Clinical features of molecular categories showing age at diagnosis (left), FLT3-ITD status (mid), and MRD (minimal residual disease) positivity at the end of induction (right). Molecular category names associated with megakaryocytic phenotypes are highlighted in red. Lines of the box represent 25% quantile, median, and 75% quantile. The upper whisker represents the higher value of maxima or 1.5 × interquartile range (IQR), and the lower whisker represents the lower value of minima or 1.5 × interquartile range (IQR). D. Grouping of molecular categories into Low, Intermediate, and High-risk groups by recursive partitioning (top) and Kaplan-Meier curves of overall survival of patients in each risk group (bottom). E. Kaplan-Meier curves and statistical significance of overall survival of patients with known prognostic factors (FLT3-ITD status: top-left, age: bottom-left, MRD positivity at the end of the induction I: top-right). F. Kaplan-Meier curves of overall survival of patients in six risk strata using risk groups (Low-Intermediate-High) and MRD positivity. G. Distribution of KMT2Ar cases among transcriptional clusters on UMAP plot, colors representing fusion partners (left) and XAGE1A and MECOM expression, colors representing relative expression (right) on UMAP plot. H. The association of fusion partners of KMT2Ar among different clusters. I. Kaplan-Meier curves of overall survival of patients with each fusion (left) and in each cluster (right). For survival curves in D, E, F, and I, statistical significance was assessed by Cox Proportional-Hazards models, and P values are shown in the plot. For the validity of prediction by KMT2Ar fusion partners and clusters in I, c-index scores assessed by bootstrapping were shown below the plots. For I, statistical significance of the enrichment and exclusivity were assessed by two-sided Fisher’s exact test followed by the Benjamini-Hochberg adjustment (*P<0.05, **q<0.05, blue: exclusive, black: enriched). Additional Declarations: There is NO Competing Interest. ==== Refs References 1 Miles L. A. Single-cell mutation analysis of clonal evolution in myeloid malignancies. Nature 587 , 477–482, doi:10.1038/s41586-020-2864-x (2020).33116311 2 Cicconi L. & Lo-Coco F. Current management of newly diagnosed acute promyelocytic leukemia. Ann Oncol 27 , 1474–1481, doi:10.1093/annonc/mdw171 (2016).27084953 3 Tenen D. G. Disruption of differentiation in human cancer: AML shows the way. Nat Rev Cancer 3 , 89–101, doi:10.1038/nrc989 (2003).12563308 4 Morita K. Clonal evolution of acute myeloid leukemia revealed by high-throughput single-cell genomics. Nat Commun 11 , 5327, doi:10.1038/s41467-020-19119-8 (2020).33087716 5 Bolouri H. The molecular landscape of pediatric acute myeloid leukemia reveals recurrent structural alterations and age-specific mutational interactions. Nat Med 24 , 103–112, doi:10.1038/nm.4439 (2018).29227476 6 Cancer Genome Atlas Research, N. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med 368 , 2059–2074, doi:10.1056/NEJMoa1301689 (2013).23634996 7 Papaemmanuil E. Genomic Classification and Prognosis in Acute Myeloid Leukemia. N Engl J Med 374 , 2209–2221, doi:10.1056/NEJMoa1516192 (2016).27276561 8 Jaju R. J. A novel gene, NSD1, is fused to NUP98 in the t(5;11)(q35;p15.5) in de novo childhood acute myeloid leukemia. Blood 98 , 1264–1267, doi:10.1182/blood.v98.4.1264 (2001).11493482 9 Gruber T. A. An Inv(16)(p13.3q24.3)-encoded CBFA2T3-GLIS2 fusion protein defines an aggressive subtype of pediatric acute megakaryoblastic leukemia. Cancer Cell 22 , 683–697, doi:10.1016/j.ccr.2012.10.007 (2012).23153540 10 Umeda M. Integrated Genomic Analysis Identifies UBTF Tandem Duplications as a Recurrent Lesion in Pediatric Acute Myeloid Leukemia. Blood Cancer Discov 3 , 194–207, doi:10.1158/2643-3230.BCD-21-0160 (2022).35176137 11 Khoury J. D. The 5th edition of the World Health Organization Classification of Haematolymphoid Tumours: Myeloid and Histiocytic/Dendritic Neoplasms. Leukemia 36 , 1703–1719, doi:10.1038/s41375-022-01613-1 (2022).35732831 12 Arber D. A. International Consensus Classification of Myeloid Neoplasms and Acute Leukemias: integrating morphologic, clinical, and genomic data. Blood 140 , 1200–1228, doi:10.1182/blood.2022015850 (2022).35767897 13 Mrozek K. Outcome prediction by the 2022 European LeukemiaNet genetic-risk classification for adults with acute myeloid leukemia: an Alliance study. Leukemia, doi:10.1038/s41375-023-01846-8 (2023). 14 Rubnitz J. E. Clofarabine Can Replace Anthracyclines and Etoposide in Remission Induction Therapy for Childhood Acute Myeloid Leukemia: The AML08 Multicenter, Randomized Phase III Trial. J Clin Oncol 37 , 2072–2081, doi:10.1200/JCO.19.00327 (2019).31246522 15 Pollard J. A. Sorafenib in Combination With Standard Chemotherapy for Children With High Allelic Ratio FLT3/ITD+ Acute Myeloid Leukemia: A Report From the Children’s Oncology Group Protocol AAML1031. J Clin Oncol 40 , 2023–2035, doi:10.1200/JCO.21.01612 (2022).35349331 16 Puumala S. E. , Ross J. A. , Aplenc R. & Spector L. G. Epidemiology of childhood acute myeloid leukemia. Pediatr Blood Cancer 60 , 728–733, doi:10.1002/pbc.24464 (2013).23303597 17 McNeer N. A. Genetic mechanisms of primary chemotherapy resistance in pediatric acute myeloid leukemia. Leukemia 33 , 1934–1943, doi:10.1038/s41375-019-0402-3 (2019).30760869 18 Iacobucci I. Genomic subtyping and therapeutic targeting of acute erythroleukemia. Nat Genet 51 , 694–704, doi:10.1038/s41588-019-0375-1 (2019).30926971 19 Fornerod M. Integrative Genomic Analysis of Pediatric Myeloid-Related Acute Leukemias Identifies Novel Subtypes and Prognostic Indicators. Blood Cancer Discov 2 , 586–599, doi:10.1158/2643-3230.BCD-21-0049 (2021).34778799 20 Newman S. Genomes for Kids: The Scope of Pathogenic Mutations in Pediatric Cancer Revealed by Comprehensive DNA and RNA Sequencing. Cancer Discov 11 , 3008–3027, doi:10.1158/2159-8290.CD-20-1631 (2021).34301788 21 Rusch M. Clinical cancer genomic profiling by three-platform sequencing of whole genome, whole exome and transcriptome. Nat Commun 9 , 3962, doi:10.1038/s41467-018-06485-7 (2018).30262806 22 Schwartz J. R. The genomic landscape of pediatric myelodysplastic syndromes. Nat Commun 8 , 1557, doi:10.1038/s41467-017-01590-5 (2017).29146900 23 Andersson A. K. The landscape of somatic mutations in infant MLL-rearranged acute lymphoblastic leukemias. Nat Genet 47 , 330–337, doi:10.1038/ng.3230 (2015).25730765 24 Faber Z. J. The genomic landscape of core-binding factor acute myeloid leukemias. Nat Genet 48 , 1551–1556, doi:10.1038/ng.3709 (2016).27798625 25 Buelow D. R. Uncovering the Genomic Landscape in Newly Diagnosed and Relapsed Pediatric Cytogenetically Normal FLT3-ITD AML. Clin Transl Sci 12 , 641–647, doi:10.1111/cts.12669 (2019).31350825 26 de Rooij J. D. Pediatric non-Down syndrome acute megakaryoblastic leukemia is characterized by distinct genomic subsets with varying outcomes. Nat Genet 49 , 451–456, doi:10.1038/ng.3772 (2017).28112737 27 Van Vlierberghe P. The recurrent SET-NUP214 fusion as a new HOXA activation mechanism in pediatric T-cell acute lymphoblastic leukemia. Blood 111 , 4668–4680, doi:10.1182/blood-2007-09-111872 (2008).18299449 28 Liu Y. The genomic landscape of pediatric and young adult T-lineage acute lymphoblastic leukemia. Nat Genet 49 , 1211–1218, doi:10.1038/ng.3909 (2017).28671688 29 Andreasson P. Deletions of CDKN1B and ETV6 in acute myeloid leukemia and myelodysplastic syndromes without cytogenetic evidence of 12p abnormalities. Genes Chromosomes Cancer 19 , 77–83, doi:10.1002/(sici)1098-2264(199706)19:2<77::aid-gcc2>3.0.co;2-x (1997).9171997 30 Balgobind B. V. Leukemia-associated NF1 inactivation in patients with pediatric T-ALL and AML lacking evidence for neurofibromatosis. Blood 111 , 4322–4328, doi:10.1182/blood-2007-06-095075 (2008).18172006 31 Turner K. M. Genomically amplified Akt3 activates DNA repair pathway and promotes glioma progression. Proc Natl Acad Sci U S A 112 , 3421–3426, doi:10.1073/pnas.1414573112 (2015).25737557 32 Zhang H. , Ju Q. , Ji J. & Zhao Y. Pan-Cancer Analysis Reveals FH as a Potential Prognostic and Immunological Biomarker in Lung Adenocarcinoma. Dis Markers 2021 , 8554844, doi:10.1155/2021/8554844 (2021).34737838 33 Pounds S. A genomic random interval model for statistical analysis of genomic lesion data. Bioinformatics 29 , 2088–2095, doi:10.1093/bioinformatics/btt372 (2013).23842812 34 Ryland G. L. Description of a novel subtype of acute myeloid leukemia defined by recurrent CBFB insertions. Blood 141 , 800–805, doi:10.1182/blood.2022017874 (2023).36179268 35 von Neuhoff C. Prognostic impact of specific chromosomal aberrations in a large group of pediatric patients with acute myeloid leukemia treated uniformly according to trial AML-BFM 98. J Clin Oncol 28 , 2682–2689, doi:10.1200/JCO.2009.25.6321 (2010).20439630 36 Harrison C. J. Cytogenetics of childhood acute myeloid leukemia: United Kingdom Medical Research Council Treatment trials AML 10 and 12. J Clin Oncol 28 , 2674–2681, doi:10.1200/JCO.2009.24.8997 (2010).20439644 37 Huber S. AML classification in the year 2023: How to avoid a Babylonian confusion of languages. Leukemia, doi:10.1038/s41375-023-01909-w (2023). 38 Ross M. E. Gene expression profiling of pediatric acute myelogenous leukemia. Blood 104 , 3679–3687, doi:10.1182/blood-2004-03-1154 (2004).15226186 39 Cheng W. Y. Transcriptome-based molecular subtypes and differentiation hierarchies improve the classification framework of acute myeloid leukemia. Proc Natl Acad Sci U S A 119 , e2211429119, doi:10.1073/pnas.2211429119 (2022).36442087 40 Groschel S. A single oncogenic enhancer rearrangement causes concomitant EVI1 and GATA2 deregulation in leukemia. Cell 157 , 369–381, doi:10.1016/j.cell.2014.02.019 (2014).24703711 41 Schwartz J. R. The acquisition of molecular drivers in pediatric therapy-related myeloid neoplasms. Nat Commun 12 , 985, doi:10.1038/s41467-021-21255-8 (2021).33579957 42 Montefiori L. E. Enhancer Hijacking Drives Oncogenic BCL11B Expression in Lineage-Ambiguous Stem Cell Leukemia. Cancer Discov 11 , 2846–2867, doi:10.1158/2159-8290.CD-21-0145 (2021).34103329 43 Tosi S. Paediatric acute myeloid leukaemia with the t(7;12)(q36;p13) rearrangement: a review of the biological and clinical management aspects. Biomark Res 3 , 21, doi:10.1186/s40364-015-0041-4 (2015).26605042 44 McInnes L , H. J. , & Melville J (2018). UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction. Retrieved from http://arxiv.org/abs/1802.03426. 45 Becht E. Dimensionality reduction for visualizing single-cell data using UMAP. Nat Biotechnol, doi:10.1038/nbt.4314 (2018). 46 Zeng A. G. X. A cellular hierarchy framework for understanding heterogeneity and predicting drug response in acute myeloid leukemia. Nat Med 28 , 1212–1223, doi:10.1038/s41591-022-01819-x (2022).35618837 47 Haghverdi L. , Buettner F. & Theis F. J. Diffusion maps for high-dimensional single-cell analysis of differentiation data. Bioinformatics 31 , 2989–2998, doi:10.1093/bioinformatics/btv325 (2015).26002886 48 Martelli M. P. Novel NPM1 exon 5 mutations and gene fusions leading to aberrant cytoplasmic nucleophosmin in AML. Blood 138 , 2696–2701, doi:10.1182/blood.2021012732 (2021).34343258 49 Zhang X. , Sun J. , Yu W. & Jin J. Current views on the genetic landscape and management of variant acute promyelocytic leukemia. Biomark Res 9 , 33, doi:10.1186/s40364-021-00284-x (2021).33957999 50 Panagopoulos I. Fusion of the FUS gene with ERG in acute myeloid leukemia with t(16;21)(p11;q22). Genes Chromosomes Cancer 11 , 256–262, doi:10.1002/gcc.2870110408 (1994).7533529 51 Thomsen C. , Grundevik P. , Elias P. , Stahlberg A. & Aman P. A conserved N-terminal motif is required for complex formation between FUS, EWSR1, TAF15 and their oncogenic fusion proteins. FASEB J 27 , 4965–4974, doi:10.1096/fj.13-234435 (2013).23975937 52 Dreyling M. H. The t(10;11)(p13;q14) in the U937 cell line results in the fusion of the AF10 gene and CALM, encoding a new member of the AP-3 clathrin assembly protein family. Proc Natl Acad Sci U S A 93 , 4804–4809, doi:10.1073/pnas.93.10.4804 (1996).8643484 53 Borrow J. The translocation t(8;16)(p11;p13) of acute myeloid leukaemia fuses a putative acetyltransferase to the CREB-binding protein. Nat Genet 14 , 33–41, doi:10.1038/ng0996-33 (1996).8782817 54 von Bergh A. R. High incidence of t(7;12)(q36;p13) in infant AML but not in infant ALL, with a dismal outcome and ectopic expression of HLXB9. Genes Chromosomes Cancer 45 , 731–739, doi:10.1002/gcc.20335 (2006).16646086 55 Gamou T. The partner gene of AML1 in t(16;21) myeloid malignancies is a novel member of the MTG8(ETO) family. Blood 91 , 4028–4037 (1998).9596646 56 Guastadisegni M. C. CBFA2T2 and C20orf112: two novel fusion partners of RUNX1 in acute myeloid leukemia. Leukemia 24 , 1516–1519, doi:10.1038/leu.2010.106 (2010).20520637 57 Hitzler J. K. , Cheung J. , Li Y. , Scherer S. W. & Zipursky A. GATA1 mutations in transient leukemia and acute megakaryoblastic leukemia of Down syndrome. Blood 101 , 4301–4304, doi:10.1182/blood-2003-01-0013 (2003).12586620 58 Caligiuri M. A. Molecular rearrangement of the ALL-1 gene in acute myeloid leukemia without cytogenetic evidence of 11q23 chromosomal translocations. Cancer Res 54 , 370–373 (1994).8275471 59 Li Z. Developmental stage-selective effect of somatically mutated leukemogenic transcription factor GATA1. Nat Genet 37 , 613–619, doi:10.1038/ng1566 (2005).15895080 60 Lopez C. K. Ontogenic Changes in Hematopoietic Hierarchy Determine Pediatric Specificity and Disease Phenotype in Fusion Oncogene-Driven Myeloid Leukemia. Cancer Discov 9 , 1736–1753, doi:10.1158/2159-8290.CD-18-1463 (2019).31662298 61 de Rooij J. D. Recurrent abnormalities can be used for risk group stratification in pediatric AMKL: a retrospective intergroup study. Blood 127 , 3424–3430, doi:10.1182/blood-2016-01-695551 (2016).27114462 62 Patel J. P. Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med 366 , 1079–1089, doi:10.1056/NEJMoa1112304 (2012).22417203 63 Vassiliou G. S. Mutant nucleophosmin and cooperating pathways drive leukemia initiation and progression in mice. Nat Genet 43 , 470–475, doi:10.1038/ng.796 (2011).21441929 64 Yun H. Mutational synergy during leukemia induction remodels chromatin accessibility, histone modifications and three-dimensional DNA topology to alter gene expression. Nat Genet 53 , 1443–1455, doi:10.1038/s41588-021-00925-9 (2021).34556857 65 Lasry A. An inflammatory state remodels the immune microenvironment and improves risk stratification in acute myeloid leukemia. Nat Cancer 4 , 27–42, doi:10.1038/s43018-022-00480-0 (2023).36581735 66 Takeda A. , Goolsby C. & Yaseen N. R. NUP98-HOXA9 induces long-term proliferation and blocks differentiation of primary human CD34+ hematopoietic cells. Cancer Res 66 , 6628–6637, doi:10.1158/0008-5472.CAN-06-0458 (2006).16818636 67 Mullighan C. G. Pediatric acute myeloid leukemia with NPM1 mutations is characterized by a gene expression profile with dysregulated HOX gene expression distinct from MLL-rearranged leukemias. Leukemia 21 , 2000–2009, doi:10.1038/sj.leu.2404808 (2007).17597811 68 Langfelder P. & Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics 9 , 559, doi:10.1186/1471-2105-9-559 (2008).19114008 69 Ng S. W. A 17-gene stemness score for rapid determination of risk in acute leukaemia. Nature 540 , 433–437, doi:10.1038/nature20598 (2016).27926740 70 Elsayed A. H. A six-gene leukemic stem cell score identifies high risk pediatric acute myeloid leukemia. Leukemia 34 , 735–745, doi:10.1038/s41375-019-0604-8 (2020).31645648 71 Bottomly D. Integrative analysis of drug response and clinical outcome in acute myeloid leukemia. Cancer Cell 40 , 850–864 e859, doi:10.1016/j.ccell.2022.07.002 (2022).35868306 72 Matsuo H. Prognostic implications of CEBPA mutations in pediatric acute myeloid leukemia: a report from the Japanese Pediatric Leukemia/Lymphoma Study Group. Blood Cancer J 4 , e226, doi:10.1038/bcj.2014.47 (2014).25014773 73 Groschel S. Deregulated expression of EVI1 defines a poor prognostic subset of MLL-rearranged acute myeloid leukemias: a study of the German-Austrian Acute Myeloid Leukemia Study Group and the Dutch-Belgian-Swiss HOVON/SAKK Cooperative Group. J Clin Oncol 31 , 95–103, doi:10.1200/JCO.2011.41.5505 (2013).23008312 74 Bill M. Mutational landscape and clinical outcome of patients with de novo acute myeloid leukemia and rearrangements involving 11q23/KMT2A. Proc Natl Acad Sci U S A 117 , 26340–26346, doi:10.1073/pnas.2014732117 (2020).33020282 75 Meshinchi S. Clinical implications of FLT3 mutations in pediatric AML. Blood 108 , 3654–3661, doi:10.1182/blood-2006-03-009233 (2006).16912228 76 Ho P. A. Prevalence and prognostic implications of WT1 mutations in pediatric acute myeloid leukemia (AML): a report from the Children’s Oncology Group. Blood 116 , 702–710, doi:10.1182/blood-2010-02-268953 (2010).20413658 77 Prior I. A. , Lewis P. D. & Mattos C. A comprehensive survey of Ras mutations in cancer. Cancer Res 72 , 2457–2467, doi:10.1158/0008-5472.CAN-11-2612 (2012).22589270 78 Takahashi S. Downstream molecular pathways of FLT3 in the pathogenesis of acute myeloid leukemia: biology and therapeutic implications. J Hematol Oncol 4 , 13, doi:10.1186/1756-8722-4-13 (2011).21453545 79 Mahmoud A. M. Cancer testis antigens as immunogenic and oncogenic targets in breast cancer. Immunotherapy 10 , 769–778, doi:10.2217/imt-2017-0179 (2018).29926750 80 Chen P. Single-Cell and Spatial Transcriptomics Decodes Wharton’s Jelly-Derived Mesenchymal Stem Cells Heterogeneity and a Subpopulation with Wound Repair Signatures. Adv Sci (Weinh) 10 , e2204786, doi:10.1002/advs.202204786 (2023).36504438 81 Jeandidier E. A cytogenetic study of 397 consecutive acute myeloid leukemia cases identified three with a t(7;21) associated with 5q abnormalities and exhibiting similar clinical and biological features, suggesting a new, rare acute myeloid leukemia entity. Cancer Genet 205 , 365–372, doi:10.1016/j.cancergen.2012.04.007 (2012).22867997 82 Zhang J. The genetic basis of early T-cell precursor acute lymphoblastic leukaemia. Nature 481 , 157–163, doi:10.1038/nature10725 (2012).22237106 83 Nilius-Eliliwi V. Broad genomic workup including optical genome mapping uncovers a DDX3X: MLLT10 gene fusion in acute myeloid leukemia. Front Oncol 12 , 959243, doi:10.3389/fonc.2022.959243 (2022).36158701 84 Chisholm K. M. Pathologic, cytogenetic, and molecular features of acute myeloid leukemia with megakaryocytic differentiation: A report from the Children’s Oncology Group. Pediatr Blood Cancer, e30251 , doi:10.1002/pbc.30251 (2023). 85 Perlman E. J. MLLT1 YEATS domain mutations in clinically distinctive Favourable Histology Wilms tumours. Nat Commun 6 , 10013, doi:10.1038/ncomms10013 (2015).26635203 86 Schwartzentruber J. Driver mutations in histone H3.3 and chromatin remodelling genes in paediatric glioblastoma. Nature 482 , 226–231, doi:10.1038/nature10833 (2012).22286061 87 Gutierrez A. & Kentsis A. Acute myeloid/T-lymphoblastic leukaemia (AMTL): a distinct category of acute leukaemias with common pathogenesis in need of improved therapy. Br J Haematol 180 , 919–924, doi:10.1111/bjh.15129 (2018).29441563 88 Brown A. L. RUNX1-mutated families show phenotype heterogeneity and a somatic mutation profile unique to germline predisposed AML. Blood Adv 4 , 1131–1144, doi:10.1182/bloodadvances.2019000901 (2020).32208489 89 Feurstein S. & Godley L. A. Germline ETV6 mutations and predisposition to hematological malignancies. Int J Hematol 106 , 189–195, doi:10.1007/s12185-017-2259-4 (2017).28555414 90 Breiman L. , Friedman J. H. & Olshen R. A. Classification and regression trees. (New York (N.Y.) : Chapman and Hall, 1984). 91 Tarlock K. Significant Improvements in Survival for Patients with t(6;9)(p23;q34)/DEK-NUP214 in Contemporary Trials with Intensification of Therapy: A Report from the Children’s Oncology Group. Blood 138 , 519, doi:10.1182/blood-2021-147576 (2021). 92 Lugthart S. Clinical, molecular, and prognostic significance of WHO type inv(3)(q21q26.2)/t(3;3)(q21;q26.2) and various other 3q abnormalities in acute myeloid leukemia. J Clin Oncol 28 , 3890–3898, doi:10.1200/JCO.2010.29.2771 (2010).20660833 93 Creutzig U. Diagnosis and management of acute myeloid leukemia in children and adolescents: recommendations from an international expert panel. Blood 120 , 3187–3205, doi:10.1182/blood-2012-03-362608 (2012).22879540 94 Lindsley R. C. Acute myeloid leukemia ontogeny is defined by distinct somatic mutations. Blood 125 , 1367–1376, doi:10.1182/blood-2014-11-610543 (2015).25550361 95 Gao Y. Distinct Mutation Landscapes Between Acute Myeloid Leukemia With Myelodysplasia-Related Changes and De Novo Acute Myeloid Leukemia. Am J Clin Pathol 157 , 691–700, doi:10.1093/ajcp/aqab172 (2022).34664628 96 Krivtsov A. V. A Menin-MLL Inhibitor Induces Specific Chromatin Changes and Eradicates Disease in Models of MLL-Rearranged Leukemia. Cancer Cell 36 , 660–673 e611, doi:10.1016/j.ccell.2019.11.001 (2019).31821784 97 Uckelmann H. J. Therapeutic targeting of preleukemia cells in a mouse model of NPM1 mutant acute myeloid leukemia. Science 367 , 586–590, doi:10.1126/science.aax5863 (2020).32001657 98 Heikamp E. B. The menin-MLL1 interaction is a molecular dependency in NUP98-rearranged AML. Blood 139 , 894–906, doi:10.1182/blood.2021012806 (2022).34582559 99 Issa G. C. The menin inhibitor revumenib in KMT2A-rearranged or NPM1-mutant leukaemia. Nature 615 , 920–924, doi:10.1038/s41586-023-05812-3 (2023).36922593 100 Alexander T. B. The genetic basis and cell of origin of mixed phenotype acute leukaemia. Nature 562 , 373–379, doi:10.1038/s41586-018-0436-0 (2018).30209392 101 Wu G. The genomic landscape of diffuse intrinsic pontine glioma and pediatric non-brainstem high-grade glioma. Nat Genet 46 , 444–450, doi:10.1038/ng.2938 (2014).24705251 102 Tian L. CICERO: a versatile method for detecting complex and diverse driver fusions using cancer RNA sequencing data. Genome Biol 21 , 126, doi:10.1186/s13059-020-02043-x (2020).32466770 103 Jakubek Y. A. Large-scale analysis of acquired chromosomal alterations in non-tumor samples from patients with cancer. Nat Biotechnol 38 , 90–96, doi:10.1038/s41587-019-0297-6 (2020).31685958 104 Edmonson M. N. Bambino: a variant detector and alignment viewer for next-generation sequencing data in the SAM/BAM format. Bioinformatics 27 , 865–866, doi:10.1093/bioinformatics/btr032 (2011).21278191 105 Hagiwara K. RNAIndel: discovering somatic coding indels from tumor RNA-Seq data. Bioinformatics 36 , 1382–1390, doi:10.1093/bioinformatics/btz753 (2020).31593214 106 Hagiwara K. , Edmonson M. N. , Wheeler D. A. & Zhang J. indelPost: harmonizing ambiguities in simple and complex indel alignments. Bioinformatics, doi:10.1093/bioinformatics/btab601 (2021). 107 McLaren W. The Ensembl Variant Effect Predictor. Genome Biol 17 , 122, doi:10.1186/s13059-016-0974-4 (2016).27268795 108 Edmonson M. N. Pediatric Cancer Variant Pathogenicity Information Exchange (PeCanPIE): a cloud-based platform for curating and classifying germline variants. Genome Res 29 , 1555–1565, doi:10.1101/gr.250357.119 (2019).31439692 109 Li H. & Durbin R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25 , 1754–1760, doi:10.1093/bioinformatics/btp324 (2009).19451168 110 Li H. & Durbin R. Fast and accurate long-read alignment with Burrows-Wheeler transform. Bioinformatics 26 , 589–595, doi:10.1093/bioinformatics/btp698 (2010).20080505 111 Wang J. CREST maps somatic structural variation in cancer genomes with base-pair resolution. Nat Methods 8 , 652–654, doi:10.1038/nmeth.1628 (2011).21666668 112 Chen X. CONSERTING: integrating copy-number analysis with structural-variation detection. Nat Methods 12 , 527–530, doi:10.1038/nmeth.3394 (2015).25938371 113 Li H. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25 , 2078–2079, doi:10.1093/bioinformatics/btp352 (2009).19505943 114 Koboldt D. C. VarScan 2: somatic mutation and copy number alteration discovery in cancer by exome sequencing. Genome Res 22 , 568–576, doi:10.1101/gr.129684.111 (2012).22300766 115 Olshen A. B. , Venkatraman E. S. , Lucito R. & Wigler M. Circular binary segmentation for the analysis of array-based DNA copy number data. Biostatistics 5 , 557–572, doi:10.1093/biostatistics/kxh008 (2004).15475419 116 Storey J. D. A direct approach to false discovery rates. Journal of the Royal Statistical Society: Series B (Statistical Methodology) 64 , 479–498, doi:10.1111/1467-9868.00346 (2002). 117 Pounds S. & Cheng C. Robust estimation of the false discovery rate. Bioinformatics 22 , 1979–1987, doi:10.1093/bioinformatics/btl328 (2006).16777905 118 Beroukhim R. Assessing the significance of chromosomal aberrations in cancer: methodology and application to glioma. Proc Natl Acad Sci U S A 104 , 20007–20012, doi:10.1073/pnas.0710052104 (2007).18077431 119 Mermel C. H. GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers. Genome Biol 12 , R41, doi:10.1186/gb-2011-12-4-r41 (2011).21527027 120 Karczewski K. J. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature 581 , 434–443, doi:10.1038/s41586-020-2308-7 (2020).32461654 121 Richards S. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med 17 , 405–424, doi:10.1038/gim.2015.30 (2015).25741868 122 Abou Tayoun A. N. Recommendations for interpreting the loss of function PVS1 ACMG/AMP variant criterion. Hum Mutat 39 , 1517–1524, doi:10.1002/humu.23626 (2018).30192042 123 Lee K. Specifications of the ACMG/AMP variant curation guidelines for the analysis of germline CDH1 sequence variants. Hum Mutat 39 , 1553–1568, doi:10.1002/humu.23650 (2018).30311375 124 Luo X. ClinGen Myeloid Malignancy Variant Curation Expert Panel recommendations for germline RUNX1 variants. Blood Adv 3 , 2962–2979, doi:10.1182/bloodadvances.2019000644 (2019).31648317 125 Gelb B. D. ClinGen’s RASopathy Expert Panel consensus methods for variant interpretation. Genet Med 20 , 1334–1345, doi:10.1038/gim.2018.3 (2018).29493581 126 Anders S. , Pyl P. T. & Huber W. HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics 31 , 166–169, doi:10.1093/bioinformatics/btu638 (2015).25260700 127 Law C. W. , Chen Y. , Shi W. & Smyth G. K. voom: Precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol 15 , R29, doi:10.1186/gb-2014-15-2-r29 (2014).24485249 128 Ritchie M. E. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43 , e47, doi:10.1093/nar/gkv007 (2015).25605792 129 Leek J. T. , Johnson W. E. , Parker H. S. , Jaffe A. E. & Storey J. D. The sva package for removing batch effects and other unwanted variation in high-throughput experiments. Bioinformatics 28 , 882–883, doi:10.1093/bioinformatics/bts034 (2012).22257669 130 Satija R. , Farrell J. A. , Gennert D. , Schier A. F. & Regev A. Spatial reconstruction of single-cell gene expression data. Nat Biotechnol 33 , 495–502, doi:10.1038/nbt.3192 (2015).25867923 131 Butler A. , Hoffman P. , Smibert P. , Papalexi E. & Satija R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol 36 , 411–420, doi:10.1038/nbt.4096 (2018).29608179 132 Stuart T. Comprehensive Integration of Single-Cell Data. Cell 177 , 1888–1902 e1821, doi:10.1016/j.cell.2019.05.031 (2019).31178118 133 Hao Y. Integrated analysis of multimodal single-cell data. Cell 184 , 3573–3587 e3529, doi:10.1016/j.cell.2021.04.048 (2021).34062119 134 Haghverdi L. , Buttner M. , Wolf F. A. , Buettner F. & Theis F. J. Diffusion pseudotime robustly reconstructs lineage branching. Nat Methods 13 , 845–848, doi:10.1038/nmeth.3971 (2016).27571553 135 Angerer P. destiny: diffusion maps for large-scale single-cell data in R. Bioinformatics 32 , 1241–1243, doi:10.1093/bioinformatics/btv715 (2016).26668002 136 Subramanian A. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 102 , 15545–15550, doi:10.1073/pnas.0506580102 (2005).16199517 137 Huang da W. , Sherman B. T. & Lempicki R. A. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4 , 44–57, doi:10.1038/nprot.2008.211 (2009).19131956 138 Steen C. B. , Liu C. L. , Alizadeh A. A. & Newman A. M. Profiling Cell Type Abundance and Expression in Bulk Tissues with CIBERSORTx. Methods Mol Biol 2117 , 135–157, doi:10.1007/978-1-0716-0301-7_7 (2020).31960376 139 Elsayed A. H. A 5-Gene Ara-C, Daunorubicin and Etoposide (ADE) Drug Response Score As a Prognostic Tool to Predict AML Treatment Outcome. Blood 134 , 1429, doi:10.1182/blood-2019-128787 (2019). 140 Tartaglia M. , Gelb B. D. & Zenker M. Noonan syndrome and clinically related disorders. Best Pract Res Clin Endocrinol Metab 25 , 161–179, doi:10.1016/j.beem.2010.09.002 (2011).21396583

PMC010xxxxxx/PMC10312944.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398052 10.21203/rs.3.rs-2968661/v1 10.21203/rs.3.rs-2968661 preprint 1 Article The Importance of Facilitating Goal-Concordant Care (GCC) in a Pandemic: The MD Anderson Experience with hospitalized COVID-19 positive patients Muthu Mayoora Dalal Shalini George Marina Clavijo Cesar Simbaqueba Lenz Caitlin Nortje Nico The University of Texas MD Anderson Cancer Center Author Contributions: The listed authors have made substantial contribution to various aspects of this article, including its concept and design, as well as the acquisition, analysis, and interpretation of data for the article. The listed authors have either drafted the article or helped revise it for important intellectual content and approved the final version for manuscript submission. Concept and design: MM, SD, MG, CS, CL, NN Collection and assembly of data: MM, SD, CL, NN Data analysis and interpretation: MM, SD, CL, NN Manuscript writing: MM, SD, MG, CS, CL, NN Final approval of manuscript: MM, SD, MG, CS, CL, NN ✉ mmuthu@mdanderson.org 12 6 2023 rs.3.rs-2968661https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs2968661v1.pdf Purpose Provider-patient communication (PPC) about goals of care (GOC) facilitates goal-concordant care (GCC) delivery. Hospital resource limitations imposed during the pandemic made it vital to deliver GCC to a patient cohort with COVID-19 and cancer. Our aim was to understand the population and adoption of GOC-PPC along with structured documentation in the form of an Advance Care Planning (ACP) note. Methods A multidisciplinary GOC task force developed processes for ease of conducting GOC-PPC and implemented structured documentation. Data were obtained from multiple electronic medical record elements, with each source identified, data integrated and analyzed. We looked at PPC and ACP documentation pre and post implementation alongside demographics, length of stay (LOS), 30-day readmission rate and mortality. Results 494 unique patients were identified, 52% male, 63% Caucasian, 28% Hispanic, 16% African American and 3% Asian. Active cancer was identified in 81% patients, of which 64% were solid tumors and 36% hematologic malignancies. LOS was 9 days with a 30-day readmission rate of 15% and inpatient mortality of 14%. Inpatient ACP note documentation was significantly higher post-implementation as compared to pre-implementation (90% vs 8%, P < 0.05). We saw sustained ACP documentation throughout the pandemic suggesting effective processes. Conclusions The implementation of institutional structured processes for GOC-PPC resulted in rapid sustainable adoption of ACP documentation for COVID-19 positive cancer patients. This was highly beneficial for this population during the pandemic, as it demonstrated the role of agile processes in care delivery models, which will be beneficial in the future when rapid implementation is needed. goals of care goal concordant care advance care planning covid-19 coronavirus National Institutes of Health through MD Anderson’s CancerPC30CA016672 ==== Body pmcBACKGROUND The Covid-19 pandemic has stressed healthcare systems across the globe, with hospitals facing resource challenges such as low bed capacity, ventilator shortages, staffing inadequacies and scarcity of personal protective equipment (PPE), to name a few. It has been found that there is futility in cardiopulmonary resuscitation in the Covid-19 patient population, with one retrospective multi-hospital study showing that patients with Covid-19 who suffered from in-hospital cardiac arrest had 100% mortality regardless of their baseline comorbidities, illness severity, and location of arrest, with 81% of these patients being on a mechanical ventilator prior to arrest and a majority of the cardiac arrests (84.1%) occurring in the ICU setting.[1] While it has always been of paramount importance to prioritize timely goals of care (GOC) conversations, the gravity of the pandemic and the known futility of resuscitation in this setting, placed further urgency on timely delivery of goal concordant care for our unique population of patients with cancer plus Covid-Effective and empathetic communication about disease prognosis, patient values and preferences, and treatment options is vital in delivering goal-concordant care (GCC). While appropriate discussions about advance care planning (ACP) are best initiated in the outpatient setting by primary oncologists, an admission to the hospital presents an important opportunity to re-evaluate and continue GOC discussions, as it signals a change in the trajectory of the patient’s illness, giving increased relevance to these conversations. It is recorded that 99% of clinicians believe that GCC discussions are important[2], however only 29% of clinicians report having such conversations.[3] It is also worthwhile to note that only roughly 11% of patients report having GOC conversations with their providers[4], though 92% of Americans indicated they would be comfortable having GOC and End-of-Life (EoL) discussions with their provider.[3] Inconsistencies with care preferences has been associated with higher medical costs and lower quality of care for the patient.[4, 5] Literature indicates that timely GOC contributes to better care experience by the patient,[6, 7] longer survival,[8] better quality of life,[8, 9, 10] and fewer depressive symptoms by patients.[8, 11] Now, more than ever, prioritizing timely GOC conversations and ensuring delivery of goal-concordant care is important, as we strive to respect the wishes of patients who do not prefer higher levels of care at EoL, while efficiently navigating potential shortages in resources and effectively steering resource allocation. Our primary aim is thus to give a global overview of our experience in delivering goal concordant care to the Covid-19 patient population within a cancer institution. PROCESS At the direction of institutional leadership, a multidisciplinary GOC task force was created to accelerate the ongoing work of engaging patients with timely GOC conversations on March 17, 2020. This taskforce included medical oncologists, intensivists, ethicists, palliative care physicians, internal medicine hospitalists, nursing, case managers, and social workers. The task force convened daily to create appropriate criteria and workflow for the inpatient cancer population, to develop virtual training and allocating resources to support primary oncologists in initiating these sensitive yet essential conversations. Additionally, the task force was responsible for creation of standardized ACP note templates, to capture essential information related to goal-concordant care. A day later, March 18, 2020, a national emergency was announced due to the rapid spread of Covid-19. The institution set up a designated Covid-19 unit and our first Covid-positive patient was admitted on March 24, 2020. This unique turn of global events prompted the initiation of a separate work stream for GOC on the Covid-19 unit. Following initial review, the Covid-19 GOC team assessed challenges in the current process, strategized and proposed an updated workflow to tailor delivery of GCC to our distinctive population of Covid-19 patients with cancer. This new workflow included daily multidisciplinary virtual rounds/discussions with team members including nursing, oncologists, hospitalists, ethicist, physical therapy/occupational therapy, social worker and case management. This multidisciplinary method was taken to ensure that a holistic approach was utilized in determining each patient’s clinical condition, performance status, and severity of cancer and Covid-19 illness, and urgency for GOC conversation. A workflow process included a 3-tiered model for GOC conversations in the Covid-19 unit (Table I), which included the new GOC-Rapid Response Team (RRT). The RRT included the attending physician, palliative care physician and an ethicist, with the ability to respond within thirty minutes, if needed. On April 24, 2020 the GOC team for the Covid-19 unit was formalized. All patients admitted to the Covid-19 unit were required to have a GOC conversation documented at some point during hospital admission, with preference given to documentation within first 24 hours of admission to the Covid-19 unit. After the initial GOC conversation, any acute change in condition would appropriately necessitate a follow-up GOC conversation with either the patient or family members (medical Power of Attorney [mPOA]/surrogate/legal next of kin). We instituted this workflow during a pilot period from April 24, 2020 through May 24, 2020 and continued the efforts from May 25, 2020 onwards to present day, making efforts to measure sustainability of this care model through January 24, 2021 (Figure I). This research was performed as part of the institutional Data-Driven Determinants for COVID-19 Oncology Discovery Effort (D3CODE), IRB-approved protocol 2020–0348. Data were obtained from structured and unstructured electronic medical record elements, clinical note text, and ACP note documentation. Each source was identified, data integrated and analyzed using the Palantir Foundry platform (Syntropy), part of the Context Engine Data Management System at the MD Anderson Cancer Center (MDACC). Additionally, for some areas of our research, which required manual data analysis, we utilized data that were collected and managed using REDCap electronic data capture tools hosted at MDACC.[12, 13] REDCap (Research Electronic Data Capture) is a secure, web-based software platform designed to support data capture for research studies, providing 1) an intuitive interface for validated data capture; 2) audit trails for tracking data manipulation and export procedures; 3) automated export procedures for seamless data downloads to common statistical packages; and 4) procedures for data integration and interoperability with external sources. RESULTS In our cancer institution, 494 unique patients who required hospitalization to the Covid-19 unit were identified from March 24, 2020 through January 24, 2021. 81% of patients admitted had an active cancer diagnosis, while the other 19% either had non-active cancer or cancer of indeterminate/unspecified origin. Of the 81% active cancers, 36% of patients had underlying active hematologic malignancies, and 64% had active solid tumor malignancies. 4.5% of total admitted patients were identified as having a cancer involving the respiratory tract. Other high-risk comorbidities identified included hypertension (72%), chronic kidney disease/end-stage renal disease (45%), diabetes mellitus (44%), chronic obstructive pulmonary disease (17%), congestive heart failure (16%), asthma (13%), venous thromboembolism (12%) and obesity (12%). Mean patient age was 59, with median being 61. Gender distribution showed 52% of patients being male and 48% of patients being female. Race and ethnicity demographics showed 63% of patients identified as Caucasian, 28.6% as Hispanic/Latino, 16.2% as African American and 2.7% as Asian. Inpatient average length of stay (LOS) was 9 days, and 30-day readmission rate was 15%. Inpatient Covid-19 mortality during this time was 14%. Of the patients that expired during their hospitalization for Covid-19 in this timeframe, 90.4% were Do Not Resuscitate (DNR), 82.2% opted for comfort care, and 9.6% remained full code status, expiring after a terminal code blue event (Figure II). Referral to social worker was 53.4%, supportive care service was 15.4%, to spiritual services it was 12.6%, and to psychiatry it was 0.6% (Figure III). During the timeframe of our study, a mean of 90% of patient encounters had ACP note documentation, with 6 out of 11 of the study period months having greater than 90% ACP note documentation (Figure IV). We noted that this practice sustained even past our pilot period and through our peak census times. During the pre-implementation period (March 24, 2020 through April 23, 2020), only 8% of Covid-19 patient encounters had ACP note documentation. Comparatively, on non-Covid hospitalized patients within our institution, ACP note documentation was recorded to be a mean of 58% for the same post-implementation time-period (Figure V). We also found that there was a correlation between age of patient and provider ACP note documentation within the first 24 hours of hospitalization of the Covid Unit, with the highest ACP note documentation rate being in patients greater than or equal to 81 years of age (51.85%) and the lowest ACP note documentation rate being in patients less than or equal to 30 years of age (22.58%).(Figure VI) DISCUSSION It is challenging to make conclusive statements regarding pre- and post-GOC algorithm implementation outcomes for the Covid-19 patients, given that pre-implementation patient cohort consisted of patients (n = 29) admitted from March 24, 2020 through April 23, 2020 and post-implementation cohort included patients admitted from April 24, 2020 through January 24, 2021 (n = 465). However, our experience showed that with implementation of a daily multidisciplinary goal-concordant approach on the Covid-19 unit, a significant proportion of physicians had routine GOC conversations with patients and/or caregivers and documented their outcomes in the format of a templated ACP note (90%), which identified goals of cancer care as well as goals of Covid-19 care specifically. Our benchmark goal for ACP note documentation during this study period was 70%. Our benchmark goal as well as achieved ACP documentation rate of 90% substantially exceeds the 11% of patients reported as having GOC conversations with their providers in literature.[4] This is further highlighted by our analysis showing 90.4% of those patients (or caregivers of patients) who expired opted for DNR status leading up to EoL, along with 82.2% of those patients electing to go the comfort care route. Additionally, we found through literature search [14–17] that our inpatient mortality rate of 14% was amongst the lowest published hospitalized Covid-19 patient mortality rate, during a time when Covid-19 vaccination was not yet widely available or robustly implemented. We were able to extract data on illness severity for our Covid-19 cancer patient population during the study time period and found that 67.4% of patients required some degree of supplemental oxygen support, while 19.8% of patients required higher levels of non-invasive oxygen support (i.e., high-flow nasal cannula, non-rebreather mask, or non-invasive positive pressure ventilation), and 8.3% of patients ultimately required mechanical ventilation (Figure VII). Additionally, 22% of these patients were noted to have worsening oxygen requirements within the first seventy-two hours of hospital admission (Figure VIII). Maintaining a low inpatient mortality rate in patients with such high illness severity furthermore emphasizes the vital significance of utilizing an adept multidisciplinary care team for complex patient populations. These figures demonstrate that early initiation of conversations regarding goal concordant care between patients, caregivers and providers have significant impact on EoL outcomes. The more traditional model of care in cancer medicine previously has been dichotomous, with curative or disease-modifying treatment offered primarily and palliative options only being discussed later in disease course. Including a selected team of experts in having these discussions, not only lowers the burden of responsibility of the primary treating physician, but also increases the support system for the patient/family/caregiver. Within our Covid-19 patient cohort for the study period, we found that 53.4% of patients were referred to social work for either medical power of attorney identification, living will documentation, hospice education, out-of-hospital DNR documentation, or other social/financial issues. 15.4% of patients were referred to our supportive/palliative care consultants for either pain/symptom management, assistance with GOC, or psychologic services. 12.6% of patients were referred to spiritual services. This type of multidisciplinary approach affords patients/family/caregivers the opportunity to look at their current situation from more than just the medical perspective. Palliative care specialists are skilled in EoL issues and questions, while ethicists are skilled in the methodology of facilitated conversations. Ethicists also ensure that different value systems are respected and integrated into the conversation. Thus, integrating these specialists into GOC conversations, along with the primary inpatient teams and oncologists, provides greater value for patients/caregivers, whose decision-making is optimized when they are presented with a global view of their treatment options and overall prognosis. Our model suggests more compassionate outcomes when utilizing a goal-concordant approach to those patients with cancer plus multiple comorbidities including Covid-19, so that they are educated early in the disease process on the option of a palliative approach and thus, may receive timely and high-quality palliative care when appropriate. Accordingly, we conclude that there is notable utility in implementing a multidisciplinary approach to goal concordant care in the hospitalized cancer population with Covid-19 illness. This concept likely has broader benefit in fundamental application to all hospitalized cancer patients. Covid-19 will likely continue its significant impact on our vulnerable immunocompromised community of patients, thus as clinicians, it is our ethical responsibility to provide patients and caregivers with the tools and education to make informed decisions regarding end-of-life care. Acknowledgements: We would like to acknowledge Anastasia Turin and Chingyi Young for their contributions on collection and assembly of data. We would also like to acknowledge the D3CODE Team (Data-Driven Determinants for COVID-19 Oncology Discovery Effort) at the University of Texas MD Anderson Cancer Center, an inter-departmental institutional collaborative of data scientists, data quality, and analytics teams for their contributions in research administration and data assembly. For a complete list of the D3CODE Team members, please refer to supplementary information labeled “Online Resource 1 – D3CODE Team Consortium Members List”. Funding: This study was supported in part by the National Institutes of Health through MD Anderson’s Cancer Support Grant PC30CA016672. Figure 1 Timeline of GOC Implementation Figure 2 Code Status in the Deceased Cohort of Hospitalized Covid Patients in a Cancer Institution Figure 3 Percentage of Covid-19 Patients Referred to Supportive Services in a Cancer Institution Figure 4 Provider ACP Note Documentation Rate on the COVID Unit in a Cancer Institution Figure 5 Provider ACP Note Documentation Rate on the General Inpatient Units in a Cancer Institution Figure 6 Provider ACP Note Documentation Rate (by Age Group) on the Covid Unit in a Cancer Institution Figure 7 Illness severity (by O2 status) on the Covid Unit in a Cancer Institution Figure 8 Illness Severity by Deterioration Index on the Covid Unit in a Cancer Institution *NULL = if length of stay is less than 72 hours or if O2 status was not consistently recorded within 72 hours Table I: 3-Tiered GOC Model MODELS OF GOC conversations Primary Oncologist led GOC (Same day) Co-managed GOC discussion (1–2 days) GQC-RRT (Rapid Response Team) (Urgent/Same day/24 hours) Why To establish clear GOC for patient Oncologist with peer support; complicated medical situation or family dynamics Clear GOC absent, patient is declining Patient population All pts admitted within past 24 hours with risk of escalation level 2 or 3 Any patient with complex clinical or psychosocial needs Any patient needing timely, integrated approach [Primary oncologist supported] who is present Primary oncologist/On-call oncologist Primary oncologist/On-call Oncologist Palliative care, Social work Primary Oncologist/On-call oncologist GOC RRT (Social work. Ethics, Palliative tare) Aim To give clear information and clarity the patient’s wishes to ensure goal-concordant care, Requires periodic re-evaluation, To give clear information and clarify the patient’s wishs in the context of the complex ongoing issues. Requires periodic ne-evaluation. To give rapid, coordinated, clear information and clarify the patient’s wishes, anticipating imminent define. Requires periodic re-evaluation. Conversation model Self-directed Faculty education conversation guide available Performed hy primary attending with Palliative care and/or social work using the briefing/debriefing model Performed by primary attending with Palliative care, ethics, and social work using the briefing/debriefing model How to obtain Self directed Consult to palliative care via EPIC Reach out to the on-call Case Management Documentation Primary oncologist documents available in the ALP tab Primary Oncologist or Palliative care documents in the ACP tab Primary Oncologist or Palliative care documents in the ACP tab Support Inpatient medical director Inpatient medical director Palliative care leader Case Management Palliative care leader Clinical Ethics Competing interests: All others declare that they have no competing interests regarding this work. Ethics approval: Approval was obtained from the institutional review board / ethics committee of the University of Texas MD Anderson Cancer Center, IRB-approved protocol 2020-0348. The procedures used in this study adhere to the tenets of the Declaration of Helsinki. Consent to participate: Not applicable. A waiver of informed consent has been granted by the institutional review board at The University of Texas MD Anderson Cancer Center. Consent for publication: Not applicable. There is no patient identifiable data in this publication. Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. OnlineResource1D3CODETeamConsortiumMembersList.pdf ==== Refs References 1. Shah P , Smith H , Olarewaju A , Jani Y , Cobb A , Owens J , Moore J , Chenna A , Hess D . Is Cardiopulmonary Resuscitation Futile in Coronavirus Disease 2019 Patients Experiencing In-Hospital Cardiac Arrest? Crit Care Med. 2021 Feb 1;49 (2 ):201–208.33093278 2. Fulmer T , Escobedo M , Berman A , Koren MJ , Hernandez S , Hult A . Physicians’ Views on Advance Care Planning and End-of-Life Care Conversations. J Am Geriatr Soc. 2018;66 (6 ):1201–1205.29797314 3. McDermott E , Bingley AF , Thomas C , Payne S , Seymour J , Clark D . Viewing patient need through professional writings: a systematic ‘ethnographic’ review of palliative care professionals’ experiences of caring for people with cancer at the end of life. Prog Palliat Care. 2006;14 (1 ):9–18. 4. Teno JM , Fisher ES , Hamel MB , Coppola K , Dawson NV . Medical care inconsistent with patients’ treatment goals: association with 1-year Medicare resource use and survival. J Am Geriatr Soc. 2002;50 (3 ):496–500.11943046 5. Khandelwal N , Curtis JR , Freedman VA , How Often Is End-of-Life Care in the United States Inconsistent with Patients’ Goals of Care? J Palliat Med. 2017;20 (12 ):1400–1404.28665781 6. Curtis JR , Downey L , Back AL , Effect of a Patient and Clinician Communication-Priming Intervention on Patient-Reported Goals-of-Care Discussions Between Patients With Serious Illness and Clinicians: A Randomized Clinical Trial. JAMA Intern Med. 2018;178 (7 ):930–940.29802770 7. Boissy A , Windover AK , Bokar D , Communication Skills Training for Physicians Improves Patient Satisfaction. J Gen Intern Med. 2016;31 (7 ):755–761.26921153 8. Temel JS , Greer JA , Muzikansky A , Early palliative care for patients with metastatic non-small-cell lung cancer. N Engl J Med. 2010;363 (8 ):733–742.20818875 9. Mack JW , Weeks JC , Wright AA , Block SD , Prigerson HG . End-of-life discussions, goal attainment, and distress at the end of life: predictors and outcomes of receipt of care consistent with preferences. J Clin Oncol. 2010;28 (7 ):1203–1208.20124172 10. Wright AA , Zhang B , Ray A , Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment. JAMA. 2008;300 (14 ):1665–1673.18840840 11. Bakitas M , Lyons KD , Hegel MT , The project ENABLE II randomized controlled trial to improve palliative care for rural patients with advanced cancer: baseline findings, methodological challenges, and solutions. Palliat Support Care. 2009;7 (1 ):75–86.19619377 12. Harris PA , Taylor R , Thielke R , Payne J , Gonzalez N , Conde JG. , Research electronic data capture (REDCap) – A metadata-driven methodology and workflow process for providing translational research informatics support, J Biomed Inform. 2009 Apr;42 (2 ):377–81.18929686 13. Harris PA , Taylor R , Minor BL , Elliott V , Fernandez M , O’Neal L , McLeod L , Delacqua G , Delacqua F , Kirby J , Duda SN , REDCap Consortium, The REDCap consortium: Building an international community of software partners, J Biomed Inform. 2019 May 9. 14. Fu C , Stoeckle JH , Masri L , Pandey A , Cao M , Littman D , Rybstein M , Saith SE , Yarta K , Rohatgi A , Makarov DV , Sherman SE , Morrissey C , Jordan AC , Razzo B , Theprungsirikul P , Tsai J , Becker DJ . COVID-19 outcomes in hospitalized patients with active cancer: Experiences from a major New York City health care system. Cancer. 2021 Sep 15;127 (18 ):3466–3475.34096048 15. Rüthrich MM , Giessen-Jung C , Borgmann S , Classen AY , Dolff S , Grüner B , Hanses F , Isberner N , Köhler P , Lanznaster J , Merle U , Nadalin S , Piepel C , Schneider J , Schons M , Strauss R , Tometten L , Vehreschild JJ , von Lilienfeld-Toal M , Beutel G , Wille K ; LEOSS Study Group. COVID-19 in cancer patients: clinical characteristics and outcome-an analysis of the LEOSS registry. Ann Hematol. 2021 Feb;100 (2 ):383–393.33159569 16. Murthy S , Archambault PM , Atique A , Carrier FM , Cheng MP , Codan C , Daneman N , Dechert W , Douglas S , Fiest KM , Fowler R , Goco G , Gu Y , Guerguerian AM , Hall R , Hsu JM , Joffe A , Jouvet P , Kelly L , Kho ME , Kruisselbrink RJ , Kumar D , Kutsogiannis DJ , Lamontagne F , Lee TC , Menon K , O’Grady H , O’Hearn K , Ovakim DH , Pharand SG , Pitre T , Reel R , Reeve B , Rewa O , Richardson D , Rishu A , Sandhu G , Sarfo-Mensah S , Shadowitz E , Sligl W , Solomon J , Stelfox HT , Swanson A , Tessier-Grenier H , Tsang JLY , Wood G ; SPRINT-SARI Canada Investigators and the Canadian Critical Care Trials Group. Characteristics and outcomes of patients with COVID-19 admitted to hospital and intensive care in the first phase of the pandemic in Canada: a national cohort study. CMAJ Open. 2021 Mar 8;9 (1 ):E181–E188. 17. Bowen A , Zucker J , Shen Y , Huang S , Yan Q , Annavajhala MK , Uhlemann AC , Kuhn L , Sobieszczyk M , Castor D . Reduction in Risk of Death Among Patients Admitted With COVID-19 Between the First and Second Epidemic Waves in New York City. Open Forum Infect Dis. 2022 Aug 25;9 (9 ):ofac436.36131846

PMC010xxxxxx/PMC10312959.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398057 10.21203/rs.3.rs-3021800/v1 10.21203/rs.3.rs-3021800 preprint 1 Article Hypothyroidism Among Children and Adolescents With Nephrotic Syndrome in Mulago National Referral Hospital, Kampala, Uganda; a Cross-sectional Study Tumwesige Maureen Rujumba Joseph Makerere University College of Health Sciences Piloya Thereza Makerere University College of Health Sciences Aujo Carol Mulago National Referral Hospital Authors’ contributions All the authors contributed in improving the research and the manuscript that has been written. ✉ tmauryn7@gmail.com 12 6 2023 rs.3.rs-3021800https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs3021800v1.pdf Background Nephrotic syndrome (NS) is the commonest glomerular disease among children. It is characterized by heavy proteinuria and is a risk factor for hypothyroidism in the affected children. Hypothyroidism is of concern because it affects the physical and intellectual development of children and adolescents. This study sought to establish the prevalence and factors associated with hypothyroidism among children and adolescents with NS. Methods A cross-sectional design was used to study 70 children and adolescents aged 1–19 years diagnosed with nephrotic syndrome and being followed up in the kidney clinic in Mulago National Referral Hospital. Questionnaires were used to collect patients’ socio-demographics and clinical information. A blood sample was taken for analysis for thyroid stimulating hormone (TSH) and free thyroxine (FT4), renal function tests and serum albumin. Hypothyroidism included both overt and subclinical forms. Overt hypothyroidism was defined as TSH level > 10 mU/L and FT4 < 10pmol/L, or FT4 < 10pmol/l with normal TSH, or TSH < 0.5mU/l. Sub-clinical hypothyroidism was defined as TSH ranging between 5 and10 mU/L with normal age appropriate FT4 levels. Urine samples were collected and taken for a dipstick examination. The data was analyzed using STATA version 14 and a p-value < 0.05 was considered as significant. Results The mean age (standard deviation) of participants was 9 years (3.8). There were more males; 36 of 70 (51.4%). The prevalence of hypothyroidism was 23% (16/70 participants). Of the 16 children with hypothyroidism, 3 (18.7%) had overt hypothyroidism while 13 had subclinical hypothyroidism. Only low serum albumin, aOR 35.80 (confidence interval 5.97–214.69 and a p value of < 0.001) was associated with hypothyroidism. Conclusion The prevalence of hypothyroidism among children and adolescent with nephrotic syndrome attending Mulago Hospital paediatric kidney clinic was 23%. Hypolbuminemia was found to be associated with hypothyroidism. Therefore, children and adolescents that have severely low levels of serum albumin should be screened for hypothyroidism and linked to endocrinologists for care. hypothyroidism nephrotic syndrome hypoalbuminemia children adolescents Fogarty International Center of the National Institutes of Health1R25TW011213 ==== Body pmcINTRODUCTION Nephrotic syndrome (NS) is one of the most common renal glomerular diseases among children (1). The incidence of idiopathic NS alone has been reported to be 1.15–16.9 per 100,000 children and this has also been found to vary by ethnicity and region (2). The disease is classically defined by persistent massive range of proteinuria (≥ 40 mg/m2/hour or urine protein/creatinine ratio ≥ 200 mg/mL or 3 + protein on urine dipstick) and hypoalbuminemia(3). NS is a known risk factor for development of hypothyroidism as a result of the disease process and its treatment as well. During the massive loss of protein in urine, proteins like thyroxine binding protein (TBG), albumin and thyroxine itself are also lost (4–6). These contribute to the development of primary hypothyroidism which is characterized by a low free thyroxine hormone (FT4) and a raised thyroid stimulating hormone (TSH) in an attempt to compensate for the loss. Furthermore, steroids which are the cornerstone of treatment of NS have also been reported to affect the release of TSH by the pituitary gland and can thus contribute to the development of central hypothyroidism (6). Overt hypothyroidism can impair physical growth and intellectual development of children and adolescents if not diagnosed and treated in time (7, 8). Mario et al also documented that normalizing the thyroid status of children with NS can improve their response to treatment of the glomerular disease(5). Despite all this, there is no well-established protocol for screening of children and adolescents with NS for hypothyroidism (6). In Uganda, proteinuric diseases in children have been reported to be eight times that which is seen in the United Kingdom with an incidence of 160 per a million population per year(9). There is however, paucity of data on the prevalence of hypothyroidism among children and adolescents with NS in Uganda where a higher incidence of proteinuric diseases are reported. The aim of this study was to therefore establish the prevalence and factors associated with hypothyroidism among children and adolescents with NS. METHODS STUDY DESIGN AND SETTING: This cross-sectional study was carried out between February 2022 and July 2022 at Mulago national referral hospital (MNRH) children’s kidney clinic, in Uganda, East Africa. MNRH is Uganda’s largest and oldest national referral hospital and the services in this facility are free except certain laboratory investigations and specialized procedures. It is located in Kampala the capital city of Uganda but the hospital receives patients that are referred from different parts of the country. The paediatrics department was a children’s kidney ward where most patients with NS are first admitted before they are later followed up in the outpatient paediatric kidney clinic which runs on Mondays every week, alongside other specialized chronic diseases clinics. The clinic has a total of 100 children and adolescents registered with NS. All their medical information is stored in medical charts kept at the clinic. While NS on the kidney ward and clinic is managed in line with the Kidney Disease Improving Global Outcomes (3) guidelines, there is neither a guideline nor routine screening for hypothyroidism. However, growth monitoring is carried out at the clinic. STUDY POPULATION All children and adolescents aged 1 to 19 years attending the Paediatric clinic at Mulago hospital with a documented diagnosis of NS in their files at time of enrollment into the clinic that was based on at least two of the following; of edema, hypoalbuminemia (≤ 2.5g/dl) and proteinuria ≥ + 3 or more on dipstick, irrespective of whether they were in remission or not. Those aged < 18 years whose parents and adolescents > 18 years who had given written informed consent and/children > 8yrs who assented were enrolled. Those that had a prior documented diagnosis of hypothyroidism before NS diagnosis were not eligible to take part in this study. STUDY PRODURE During the study period, 71 patients with a diagnosis of NS attended the paediatric clinic at MNRH and by consecutive sampling, these were screened for eligibility; 70 children and adolescents were enrolled and one was excluded because they declined to provide consent and assent. An interviewer administered questionnaire was used to collect data on the patient socio-demographics, clinical and NS disease characteristics, other risk factors for hypothyroidism (whether they used iodized salt, history of neck or brain surgeries/irradiations, other chronic medical conditions and medications used). Clinical history suggestive of hypothyroidism was also assessed; history of cold intolerance, constipation, fatigue, weight gain and menstrual irregularities, where applicable. Physical examinations were done and anthropometric measurements including weight in kilograms and length/height in centimeters that were taken using a weighing scale and stadiometer respectively. The patients’ files were reviewed for disease related factors that included the type of NS, cumulative dose of steroids, age at diagnosis of NS and the duration since the diagnosis of NS. LABORATORY INVESTIGATIONS Blood samples were collected in plain red top vacutainers and transported to the laboratory within one hour of sample collection. The laboratory at MNRH was used to carry out the thyroid function tests (FT4, TSH), serum albumin and serum creatinine (from which the estimated glomerular filtration rate, eGFR, was calculated using Schwartz formula)(10). Thyroid function tests were measured using a fully automated COBAS 6000 ROCHE HITACHI machine from Germany which uses electrogenerated chemiluminescence (ECL) technology in which luminescence is produced during electrochemical reactions in solution and has been reported to be highly specific and sensitive (functional sensitivity at 0.01 mIU/L TSH)(11, 12). Hypothyroidism in this study was defined as overt or subclinical forms. Overt hypothyroidism was defined as TSH level > 10 mU/L and FT4 < 10pmol/L or FT4 < 10pmol/l with normal TSH or TSH < 0.5mU/l. Sub-clinical hypothyroidism was defined as TSH ranging between 5 and10 mU/L with normal age appropriate FT4 levels that is: 1–5years [10–23.2pmol/L];6–10 years [10–28pmol/L] and 11–19 years [10–30pmol/L](13, 14). Urine samples were collected from each participant and urinalysis was done. Urine protein was categorized as nil, trace, +, 2+, 3+. DATA MANAGEMENT The sample size was calculated by Leslie Kish formula using a study that was done in Egypt by El-aal et al with estimated prevalence of hypothyroidism at 23.52% (15). This was adjusted to the available population according to the clinic records and a sample size of 70 patients was obtained. Consecutive sampling method was used. Data was entered into an electronic database using Epidata version 3.1 software package with built-in quality control checks. It was then exported to Stata version 14.1 (STATA CORP, TEXAS USA for analysis. The continuous variables were summarized using means and standard deviations for normally distributed data. Simple Logistic regression analysis was used to test the association between hypothyroidism and independent variables. Crude odds ratio and its 95% CI was reported as measure of association at 5% level of significance and any variable that achieved a P value of < 0.05 was considered for multivariable analysis. Variables that attained a P value < 0.05 after multivariate analysis were considered statistically significantly and independently associated with hypothyroidism. RESULTS Seventy-one children and adolescents aged 1–19 years were screened during the study period from February 2022 to July 2022 and only 70 of these were enrolled. The prevalence of hypothyroidism was 23% (16 of 70 participants) with only 3 of the 16 children having overt hypothyroidism. The study enrollment profile is shown in Fig. 1 below. Baseline characteristics of patients enrolled The mean age (standard deviation) of the children and adolescents that were enrolled was 9 years (3.8). The youngest was 2 years old while the oldest was 18 years old. The median age in years at the time of diagnosis of NS for the participants enrolled was 6 ± 3.3 IQR (3–9). The median (IQR) duration since the diagnosis of NS was made was 15.5 (8–48) months. Thirteen out of 70 patients (18.6%) resided in mountainous areas. There were 6 out of 70 patients with co-morbidities and these included sickle cell disease (in 3/6), HIV, hepatitis B and tetralogy of fallot. The cumulative dose of steroids was calculated from the patients’ records right from the time of enrollment into the renal clinic and the median (IQR) dose was 255.3 (110,490) mg/kg/day. It was found that few patients reported symptoms suggestive of hypothyroidism and none of the children had more than two symptoms. Cold intolerance was reported by 11/70 patients (15.7%), 24/70 (34.3%) reported weight gain, 25/70 (35.7%) reported fatigue and 10/70 (14.3%) patients reported constipation. None of the female adolescents reported menstrual irregularities. Of the patients that took part in the study, 43 out of 70 (61.4%) were in remission (had nil or trace proteins on urine dipstick). Their median serum creatinine levels in mg/dL (normal range 0.3–1 mg/dL) was 0.42(0.3–0.51) and the mean estimated GFR in mL/min/1.73 m2 was 143.4 ± 62.9. The mean serum albumin in g/dL (normal range 3.5–5.5 g/dL) was 3.4 ± 1.2. The median TSH in mU/L (normal range 0.7–4.5 mU/L) was 2.1 IQR (0.9–3.8) while the mean thyroxine level in pmol/L (normal range 10–28 pmol/L) was 15.6 ± 4.1. The rest of the characteristics are shown in Table 1. Description of the children with NS that had overt hypothyroidism Three out of 16 participants with hypothyroidism had the overt form. One of these was a 3 year old male who had been first diagnosed with NS at the age of 2 years. He had no symptoms suggestive of hypothyroidism and had a cumulative dose of steroids (prednisolone) of 202mg/kg. At the time of enrollment into the study he had proteinuria + 2 on urine dipstick, serum albumin of 2.0 g/dL and a low FT4 of 7.49 pmol/L. The second participant was a 4year old female who had been diagnosed with NS 2 months prior to enroll into the study. She was in remission at the time of enrollment and her prednisolone treatment was being tapered down gradually. Her cumulative dose of prednisolone was 78mg/kg and she reported to have cold intolerance. Her serum albumin was at 1.26g/dL and a low FT4 of 5.13 pmol/L. The third participant was an 11 year old female who had only been diagnosed with NS a month prior to enrollment into the study. She was at Tanner stage I and reported cold intolerance and fatigue. Her cumulative dose of prednisolone was 82.5mg/kg. She had a proteinuria of + 4 on urine dipstick at the time of enrollment into the study, with serum albumin of 1.28g/dL and a low FT4 of 8.45pmol/L. Factors independently associated with hypothyroidism. The factors that were found to be statistically significant at bivariate analysis (non-remission, presence of edema and reduced serum albumin) were then subjected to multivariate analysis and only reduced serum albumin < 2.5g/dl was found to be significantly associated with hypothyroidism, P value < 0.001 aOR 35.80(5.97,214.69). More information has been stated in Table 2. DISCUSSION This study found a significantly high prevalence of hypothyroidism (23%) among children and adolescents with NS. However, we observed that majority had subclinical hypothyroidism with many being above 10 years of age. This can be explained by the fact that there are various physiological and physical changes that occur in the body of an adolescent that may result in increased secretion of TSH hence predisposing this age group to subclinical hypothyroidism (16). It is important to note that while subclinical hypothyroidism has been reported to spontaneously resolve in some cases, it can also persist and progress to overt hypothyroidism and hence there is need for continuous monitoring of the affected patients(17). The prevalence found in this study was slightly lower than, the 33.3% prevalence of hypothyroidism reported by Marimuthu et al in a cross-sectional study done in India among children and adolescents aged 1–18 years with NS in the outpatient department(18). Although Marimuthu’s study had participants in a similar age group to this current study, with a mean age of 7.2 years (SD 3.9), the researchers probably reported a slightly higher prevalence because they only enrolled children and adolescents with steroid resistant NS(18). This type of NS is associated with longer duration of nephrosis therefore they may ultimately be losing more protein and thus loss of T4 and TBG causing hypothyroidism. (3). In this study however, we enrolled all patients with nephrotic syndrome of which the majority (67.6%, 55 out of 70) had steroid sensitive nephrotic syndrome and as such might be less likely to lose protein (including thyroglobulin, thyroxine and TSH) in urine. It is also important to note that autoimmune thyroid disorders have been reported to be higher in the Asian population and could have perhaps contributed to the high prevalence found by the researchers in this Indian study (19). Noteworthy, Marimuthu too found a similar proportion of children with subclinical hypothyroidism just like our study despite having a slightly younger population. (18). Therefore subclinical hypothyroidism may be the commonest form of hypothyroidism in NS population and whether this has long term clinical implications for the children and adolescents with NS may need further research through prospective studies. This study found that patients with a reduced serum albumin were more likely to have hypothyroidism compared to those that had normal levels of the same. Children with NS lose protein in urine and among these proteins are thyroglobulin and serum albumin which are important carriers of the thyroid hormones but also the latter which are themselves protein in nature, are lost as well(4). Serum albumin also acts as a buffer of serum levels of thyroxine before hypothyroidism eventually occurs and once it is lost in urine together with thyroglobulin, the serum concentration of the thyroid hormones also decreases (4). This explains why the present study found that the children and adolescents that had hypoalbuminemia were more likely to have hypothyroidism. Although, it was not significant in this study, we also noted that majority of the participants with hypothyroidism were not in remission for the Nephrotic syndrome. This would further emphasize that prolonged proteinuria may ultimately lead to hypoalbuminemia with subsequent hypothyroidism. These findings of hypoalbuminemia being associated with hypothyroidism are similar to the findings in a case-control study by Saffari et al which was conducted at a paediatric hospital in Qazvin, Iran. (20). The researchers reported a negative correlation between serum albumin and TSH levels in serum. TSH levels increase as a compensatory mechanism to the decrease in serum levels of thyroid hormones resulting from the loss of protein in urine among children and adolescents with NS. Similarly, El-aal et al conducted a prospective study at Sohag University Hospital, Egypt over a one-year period among 51 children aged between 1 and 12 years old with NS and found that low levels of thyroid hormones and high levels of TSH (hypothyroidism) was significantly associated with low levels of serum albumin(15). While universally most studies have reported a relationship between hypothyroidism and hypoalbuminemia, Jung et al reported contrary findings in a study conducted at Inje University Busan Paik Hospital, Korea. The researchers enrolled 31 children with NS between January 2001 and December 2017 where they compared their thyroid status during active nephrosis and in remission and they found no significant correlations between serum albumin and T4, TSH, or Free T4 levels(21). The researchers only found a negative correlation between T3 and serum albumin and they hypothesized that there were probably other mechanisms that could explain this other than the loss of protein in urine. However, the study by Jung et al had a very small sample size that may not have reached the power to detect the difference between those with low serum albumen and those with normal levels. STRENGTH AND LIMITATIONS The study was conducted in the paediatric renal clinic in Uganda’s biggest national referral hospital that serves patients from different regions of the country and therefore the findings can be generalized to the rest of the population since there was a fair representation of all the regions of the country. The study also brings new information that shades light on the prevalence and factors associated with hypothyroidism among children with NS in Uganda, East Africa. However, the study did not exclude autoimmune causes of hypothyroidism among the children with NS that we enrolled. The accessible population of children with NS in this study offered a limited sample size and hence the study may not have adequate power to detect other factors associated with hypothyroidism. CONCLUSION The prevalence of hypothyroidism among children and adolescents with NS in MNRH is quite high as it affects 1 in 4 children. The factor that is associated with hypothyroidism among these children is hypoalbuminemia. Therefore, children and adolescents with NS that have hypoalbuminemia should be screened for hypothyroidism and the treating clinicians should endeavor to achieve and maintain normal albumin levels in these patients. Further studies that are preferably multi-center, with a larger sample size, are recommended to assess more factors that could be associated with hypothyroidism in NS. Acknowledgement We acknowledge all the caretakers and the children who participated in this study as well as the study team members. We also acknowledge Health professional Education Partnership Initiative (HEPI-SHSSU) and Prof. Sarah Kiguli for all the support they offered in getting this work done. We acknowledge all the paediatricians in the department of paediatrics, Makerere University and the 2019 master of medicine paediatrics class for their tireless effort to improve this work at various stages. Funding This work was supported by the Fogarty International Center of the National Institutes of Health under Award Number 1R25TW011213. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Availability of data and materials The dataset that was generated during this study is not publicly available because we did not obtain consent from all the participants to publish raw data. It can however be availed by the corresponding author upon reasonable request. Abbreviations eGFR estimated glomerular filtration rate FT4 Free thyroxine KDIGO Kidney Disease Improving Global Outcomes MNRH Mulago National Referral Hospital NS Nephrotic syndrome TSH Thyroid Stimulating Hormone Figure 1 Study profile of children and adolescents with NS that were enrolled into this study. *those who were found to have hypothyroidism were linked to care in the paediatric endocrinology clinic Table 1 Baseline characteristics of children and adolescents with NS that were enrolled in this study Variable Frequency (N = 70) Percentage (%) Age category < 5years 11 15.7 5–10 years 31 44.3 > 10 years 28 40 Sex Male 36 51.4 Address Urban 42 60.0 Type of salt used Iodized salt 58 82.9 Local and iodized salt 7 10.0 Local salt 4 5.7 No salt 1 1.4 Weight for Age Z score (≤ 10yrs, n = 41) WAZ − 2 to + 2 Z score-Normal 32 78.1 WAZ > + 2 Z score Obese/Overweight* 9 21.9 Height for Age Z score HAZ ≥ −2 to ≤ + 2-Normal 60 85.7 HAZ <-3 to <-2 –Moderate-severely stunted 10 14.3 Type of nephrotic syndrome Steroid sensitive NS 55 67.6. Steroid resistant NS 11 15.7 Newly diagnosed** 4 5.7 Degree of proteinuria on urine dipstick at enrollment Nil and Trace 43 61.4 + 1 and + 2*** 6 8.6 + 3 and + 4 21 30 Presence of edema 9 12.9 Medication taken at enrollment Prednisolone alone 30 41.1 Prednisolone and other drugs 13 18.8 Other drugs**** 9 13.0. None 18 26.1 Laboratory results Estimated GFR <90mL/min/1.73 m2 7 10.0 Reduced serum albumin (2.5g/dL) 17 24.3 * none of the children that were overweight/obese had edema, ** newly diagnosed were children and adolescents diagnosed in a period of less than 6 weeks and as such could not be classified as SSNS or SRNS, *** did not meet nephrotic range proteinuria since they were on treatment,, **** Other drugs included Tenofovir, abacavir, lamuvidine, efervarenz, captopril, folic acid, mycophenolate mofetil, tacrolimu Table 2 Factors associated with hypothyroidism among children and adolescents with NS at Bivariate and multivariate analysis Variable Hypothyroidism cOR, 95% CI P value aOR 95% CI P value Yes (n = 16) (f,%) No (n = 54) (f,%) Age category < 5years 5(31.3) 6(11.1) 5.63(1.15,2744) 0.053 5–10 years 4(25.0) 27(50.0) 1.00 > 10 years 7(43.8) 21(38.9) 2.25(0.58,8.72) 0.241 Sex Male 10(62.5) 26(48.2) 1.79(0.57,5.64) 0.316 Female 6(37.5) 28(51.8) 1.00 Address Urban 12(75.0) 30(55.6) 2.40(0.69,8.39) 0.171 Rural 4(25.0) 24(44.4) 1.00 Comorbidity known Yes 14(87.5) 50(92.6) 1.79(0.29,10.78) 0.527 No 2(12.5) 4(7.4) 1.00 Age of the child at time of diagnosis < 5 years 7(43.7) 18(33.3) 0.58(0.13,2.71) 0.492 5–10 years 5(31.3) 30(55.6) 1.00 > 10 years 4(25.0) 6(11.1) 0.58(0.13,2.71) 0.086 Type of nephrotic syndrome Steroid sensitive NS 11(68.8) 44(81.5) 1.00 1.00 Steroid resistant NS 2(12.5) 9(16.7) 0.89(0.17,4.72) 0.89 0.71(0.03,14.29) 0.822 Newly diagnosed 3(18.8) 1(1.8) 12.0(1.13,126.79) 0.039 2.63(0.04,158) 0.664 Remission status at enrollment Yes 5(31.2) 38(70.4) 1.00 1.00 No 11(68.8) 16(29.6) 5.33 (1.56, 17.48) 0.007 3.56(0.52,24.45) 0.196 Medication taken at enrollment Prednisolone alone 7(43.8) 22(41.5) 5.41(0.61,48.27) 0.131 Pred and other drugsa 4(25.0) 9(16.9) 7.55(0.73,78.08) 0.091 None 1 (6.2) 18(32.1) 1.00 Edema Yes 6(37.5) 3(5.6) 10.2(2.18,47.71) 0.003 1.29(0.11,15.11) 0.836 No 10(62.5) 51 (94.4) 1.00 1.00 Laboratory findings Estimated GFR Normal eGFR 14(87.5) 49(90.7) 1.00 Reduced eGFR 2(12.5) 5(9.3) 1.40(0.24,8.01) 0.705 Serum albumin Normal (NR3.5–5.5g/dL) 3(18.7) 50(92.6) 1.00 1.00 Reduced(≤ 2.5g/dL) 13(81.3) 5(7.4) 54.17(10.75,272.76) 0.000 35.80(5.97,214.69) 0.000 a Prednisolone and other drugs which included captopril, tacrolimus, mycophenolate mofetil Competing interest The authors declare they have no competing interests. Declarations Ethics approval and consent to participate The School of Medicine Research Ethics Committee of Makerere University (Mak-SOMREC) gave permission to carry out this study; approval reference #Mak-SOMEREC-2021-172. Administrative clearance was obtained from Mulago National referral Hospital ethics committee and the paediatric renal clinic. The caregivers of participants below 18years of age provided written informed consent and all children aged 8–17 years provided assent and those 18 and 19 years gave informed consent to participate in the study. All the information collected in this study was treated with uttermost confidentiality and results were only disclosed to the participants, their guardians and clinicians. The participants that were found to have hypothyroidism were linked to the paediatric endocrinology clinic for further medical attention. This research was conducted in line with the declaration of Helsinki guidelines on human subjects. Consent for publication Not applicable. ==== Refs References 1. Eddy AA , Symons JM . Nephrotic syndrome in childhood. The Lancet Child. 2003;362 :629. 2. Noone DG , Iijima K , Parekh R . Idiopathic nephrotic syndrome in children. Lancet (London England). 2018;392 (10141 ):61–74.29910038 3. KDIGO. Clinical practice guideline for glomerulonephritis. Official J Int Soc Nephrol. 2012;3 (1 ). 4. Ito S , Kano K , Ando T , Ichimura T . Thyroid function in children with nephrotic syndrome. Pediatr Nephrol. 1994;8 (4 ):412–5.7947028 5. Mario FD , Pofi R , Gigante A , Rivoli L , Rosato E , Isidori AM , Hypothyroidism and Nephrotic Syndrome: Why, When and How to Treat. Curr Vasc Pharmacol. 2017;15 (5 ):398–403.28176633 6. Radhakrishnan J. Endocrine dysfunction in the nephrotic syndrome. UpToDate; 2020. 7. Counts D , Varma SK . Hypothyroidism in children. Pediatr Rev. 2009;30 (7 ):251–8.19570923 8. Biondi B , Wartofsky L . Treatment with thyroid hormone. Endocr Rev. 2014;35 (3 ):433–512.24433025 9. Odiit A. Common childhood kidney diseases in Uganda and their prevention. Afr Health Sci. 2017;17 (4 ):1126–9.29937884 10. Schwartz GJ , Haycock GB , Edelmann CM Jr , Spitzer A . A simple estimate of glomerular filtration rate in children derived from body length and plasma creatinine. Pediatrics. 1976;58 (2 ):259–63.951142 11. Kazerouni F , Amirrasouli H . Performance characteristics of three automated immunoassays for thyroid hormones. Casp J Intern Med. 2012;3 (2 ):400–104. 12. Esfandiari NH , Papaleontiou M . Biochemical Testing in Thyroid Disorders. Endocrinol Metab Clin North Am. 2017;46 (3 ):631–48.28760230 13. pediatric catalog, Mayo Foundation for Medical Education and Research., 2021 [Available from: https://pediatric.testcatalog.org/show/T4. 14. Wassner AJ , Smith JR . Nelson textbook of paediatrics. 21st ed. New Elsevier; 2019. 15. El-aal MA , Hegab AM , Elsayed E . Thyroid function in children with nephrotic syndrome: A prospective hospital-based study. Sohag Med J. 2020;24 (2 ). 16. Sahay R , Kalra S . Thyroid tantrums in teenagers. JPMA The Journal of the Pakistan Medical Association. 2021;71 (11 ):2672–3.34783759 17. Gallizzi R , Crisafulli C , Aversa T , Salzano G , Luca FD , Valenzise M Subclinical hypothyroidism in children: is it always subclinical? 2018:1–5. 18. Marimuthu V , Krishnamurthy S , Rajappa M . Non-Autoimmune Subclinical and Overt Hypothyroidism in Idiopathic Steroid-resistant Nephrotic Syndrome in Children. Indian Pediatr. 2017;54 (11 ):925–9.28849770 19. Li Y , Inomata K , Nishihara E , Kakudo K . IgG4 thyroiditis in the Asian population. Gland Surg. 2020;9 (5 ):1838–46.33224859 20. Saffari F , Ahadi S , Dalirani R , Esfandiar N , Yazdi Z , Arad B . Thyroid Dysfunction in Children with Idiopathic Nephrotic Syndrome Attending a Paediatric Hospital in Qazvin, Iran. Sultan Qaboos University medical journal. 2020;20 (4 ):e332–e8.33414938 21. Jung SH , Lee JE , Chung WY . Changes in the thyroid hormone profiles in children with nephrotic syndrome. Korean J Pediatr. 2019;62 (3 ):85–9.30304897

PMC010xxxxxx/PMC10312970.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398370 10.21203/rs.3.rs-3038923/v1 10.21203/rs.3.rs-3038923 preprint 1 Article Biomarkers for Duchenne muscular dystrophy progression: impact of age in the mdx tongue spared muscle Lorena Marcelo dos Santos Voltani Department of Structural and Functional Biology, Institute of Biosciences of Botucatu, São Paulo State University (UNESP) Santos Estela Kato Department of Structural and Functional Biology, Institute of Biosciences of Botucatu, São Paulo State University (UNESP) Ferretti Renato Department of Structural and Functional Biology, Institute of Biosciences of Botucatu, São Paulo State University (UNESP) Gowda G.A. Nagana Northwest Metabolomics Research Center; Mitochondria and Metabolism Center, Anesthesiology and Pain Medicine, University of Washington Odom Guy L. Department of Neurology, Wellstone Muscular Dystrophy Specialized Research Center, University of Washington School of Medicine Chamberlain Jeffrey S. Department of Neurology, Wellstone Muscular Dystrophy Specialized Research Center, University of Washington School of Medicine Matsumura Cintia Yuri Department of Structural and Functional Biology, Institute of Biosciences of Botucatu, São Paulo State University (UNESP) Authors’ contributions MSVL JSC and CYM conceived and designed experiments; MSVL, EKS, NGGA and CYM performed experiments; MSVL, RF, NGGA, GLO, JSC and CYM analyzed the data; MSVL, EKS, GLO, JSC and CYM interpreted the results of the experiments and wrote the manuscript with editing from MSVL. The authors read and approved the final manuscript. ✉ cintia.matsumura@unesp.br 13 6 2023 rs.3.rs-3038923https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs3038923v1.pdf Background: Duchenne muscular dystrophy (DMD) is a severe form of muscular dystrophy without an effective treatment, caused by mutations in the DMD gene, leading to the absence of dystrophin. DMD results in muscle weakness, loss of ambulation and death at an early age. Metabolomics studies in mdx mice, the most used model for DMD, reveal changes in metabolites associated with muscle degeneration and aging. In DMD, the tongue muscles exhibit unique behavior, initially showing partial protection against inflammation but later experiencing fibrosis and loss of muscle fibers. Certain metabolites and proteins, like TNF-α and TGF-β, are potential biomarkers for dystrophic muscle characterization. Methods: To investigate disease progression and aging, we utilized young (1-month old) and old (21–25 months old) mdx and wild-type mice. Metabolite changes were analyzed using 1-H Nuclear Magnetic Resonance, while TNF-α and TGF-β were assessed using Western blotting to examine inflammation, and fibrosis. Morphometric analysis was conducted to assess the extent of myofiber damage between groups. Results: The histological analysis of the tongue showed no differences between groups. No differences were found between the concentrations of metabolites from wild type or mdx animals of the same age. The metabolites alanine, methionine, 3-methylhistidine were higher, and taurine and glycerol were lower in young animals in both wild type and mdx (p < 0.001). The metabolites glycine (p < 0.001) and glutamic acid (p = 0.0018) were different only in the mdx groups, being higher in young mdx mice. Acetic acid, phosphocreatine, isoleucine, succinic acid, creatine and the proteins TNF-α and TGF-β had no difference in the analysis between groups (p > 0.05). Conclusions: Surprisingly, histological and protein analysis reveals that the tongue of young and old mdx animals is protected from severe myonecrosis observed in other muscles. The metabolites alanine, methionine, 3-methylhistidine, taurine, and glycerol may be effective for specific assessments, although their use for disease progression monitoring should be cautious due to age-related changes. Acetic acid, phosphocreatine, isoleucine, succinate, creatine, TNF-α, and TGF-β do not vary with aging and remain constant in spared muscles, suggesting their potential as specific biomarkers for DMD progression independent of aging. São Paulo Research Foundation (FAPESP)2018/25598-0 2019/20162-1 2013/00312-2 CNPq456357/2014-9 NIHR01AR040864 DODW81XWH-18-1-0624 ==== Body pmcIntroduction Duchenne muscular dystrophy (DMD) affects approximately 1 in 3600–6000 male live births [1] and is a severe form of muscular dystrophy without an effective treatment [2]. It is an X-linked recessive disease [2] caused by mutations in the DMD gene [3]) which leads to the absence of dystrophin [4]. The lack of dystrophin causes continuous loss of muscle strength, myofiber damage, chronic inflammation, progressive fibrosis, and muscle stem cell dysfunction. This dystrophic scenario leads to a loss of ambulation in the early teens to twenties. Although patients’ life expectancy has improved with current standards of cardioprotective care and respiratory support, they often die around the third or fourth decade of life, mainly due to cardiac and respiratory complications [2][5][6]. The mdx mice is the most used animal model for research on DMD [7][8]. Around 20 days old, these animals begin to show their first signs of muscle degeneration and regeneration [9]. The mdx mice have reduced life spans, by about 17 to 19% compared to wild-type. At 26 months, average lifespans, the muscles present typical dystrophic characteristics: loss of muscle fibers with increased fibrosis, fat infiltration, necrotic fibers and regenerated fibers [10]. Metabolomics is the quantitative analysis of metabolites produced by an organism under certain conditions. Metabolomics provides an integrated view of biochemical pathways in complex organisms, thus producing a more detailed and systematic overview of the cellular processes and its response to diseases. Therefore, this approach is essential for the definition of personalized medicine, through the establishment of metabolite profiles and biomarkers for certain pathological states [11]. Studies in mdx mice show changes in metabolites related to the progression of muscle degeneration and aging [12][13] [14] In DMD, the muscles of the oral cavity are also affected, causing dysphagia in late stages of the disease, which worsens with advancing age and disease progression [15] [16]. In mdx mice, the tongue muscles (TON) show an unusual behavior as the disease progresses compared to other muscles. In 3-month mdx mice, inflammatory cells were hardly found in the TON, unlike other masticatory muscles. In addition, the expression of collagen did not change in the TON of these mdx, while it was three times higher in masseter muscle (MAS), when compared to the control [17]. These results indicate partial protection of this muscle against myonecrosis and inflammation at 3 months of age. However, at 26 months of age, Chamberlain et al. (2007) [10] described the TON of the mdx as the second most affected muscle, due to fibrosis and the loss of fibers in the central portion of the muscle, just after the diaphragm. Some metabolites from dystrophic mice analyzed by Nuclear Magnetic Resonance (NMR) have been suggested to determine biomarkers for the state of muscle fibers [18], such as: increase of glutamate, glutamine, succinate, isoleucine, acetate, alanine and glycerol, in contrast to decreased of carnosine, taurine, glycine, methionine and creatine in the mdx’s diaphragm and quadriceps muscles, compared to the wild-type [13]. Besides metabolites, there are some proteins that are well established in the current literature related to inflammation and fibrosis, like Tumor Necrosis Factor alpha (TNF-α) and Transforming Growth Factor beta (TGF-β), respectively. These proteins are very promising biomarkers for the dystrophic muscle characterization, for their relation to the myonecrosis as seen in the mdx animal model and human patients [19]. The present study aimed to validate biomarkers for diagnosis and progression of DMD through the analysis of metabolites, proteins related to inflammation and fibrosis, and histology of the tongue of mdx mice over time. We evaluated metabolites and proteins in the tongue muscles, within young (30 days old) and old (21 to 25 months old) mdx and wild type mice. Surprisingly, our histological results found that the tongue remains protected in older mdx mice (21 to 25 months). Our results allowed us to identify biomarkers that change with aging, regardless of the absence of dystrophin. Other biomarkers seem to be excellent candidates to indicate the progression of dystrophinopathy over time, as they do not change with aging and remain constant in the spared muscles. Furthermore, analyzing these possible biomarkers in TON at different ages can help to understand the protection mechanisms involved in the pathological process and support the development of future approaches for the diagnosis and monitoring progression of DMD. Materials and Methods Animals Male and female, young and old mdx (C57BL/10-DMDmdx/PasUnib) and age-match wild type mice (C57BL/10ScCr/PasUnib) were obtained and maintained by our institutional animal care facility of Institute of Bioscience (Botucatu) - UNESP. All mouse experimentation was approved by our Institution committee and done in accordance with the guidelines of the Brazilian College for Animal Experimentation (protocol n° 1095-CEUA). The animals were divided into four groups: young mdx (1-month old), old mdx (21–25 months old), young wild type (1 month-old) and old wild type (21–25 months old). Tissue harvesting The animals were euthanized with an overdose of intraperitoneal anesthesia of xylazine hydrochloride (30 mg / kg) and ketamine hydrochloride (300 mg / kg). The tongue (TON) was dissected and fixed for histological techniques or frozen in liquid nitrogen for western blotting and metabolome assays. Histology The TON muscles of the young wild type (n = 5), young mdx (n = 5), old wild type (n = 4) and old mdx (n = 6) groups were sectioned and stained with Masson’s Trichrome, to distinguish and quantify the areas of fibrosis (FIB), areas of muscle fibers with peripheral nuclei (PN) and central nuclei (CN). The analyzes were performed blindly, and the areas were expressed in relation to the area of total transverse fibers of the section. Proteins analysis Proteins related to the mechanisms of inflammation (TNF-α) and fibrosis (TGF-β) were quantified with the Western Blotting assay, as described previously [20]. The values were normalized with the glyceraldehyde 3-phosphate dehydrogenase protein (GAPDH), incubated on the same membrane after routine stripping methods. 3.8 Metabolomics by nuclear magnetic resonance (NMR) spectroscopy In order to verify possible changes in the metabolic profile, metabolomics analysis was performed by NMR spectroscopy of the tongue muscles of young (n = 10) and old (n = 5) wild type mice and young (n = 11) and old (n = 5) mdx mice. Data acquisition was performed on a Varian INOVA spectrometer operating at a resonance frequency of 1 H of 600 MHz. The samples were homogenized in a methanol / chloroform solution (2:1). After fifteen minutes, a solution of chloroform / milli Q water (1:1) was added to the pellet. Then, the samples were centrifuged at 4000 RPM for 20 minutes. The supernatant was collected and lyophilized. The obtained powder was resuspended in deuterated water (D2O) with trimethylsilyl tetradeuteropropionic acid (TSP, 10 mM). Subsequently, the samples were transferred to a standard NMR tube for spectral analysis. The D2O allowed the device to be monitored and blocked by the device’s resonant frequency. The TSP reference signals were used to assess the quality of the spectra as well as to quantify the identified substances. Proton spectra in one dimension, using pulse sequences optimized for suppression of the water signal and were collected at 25 ° C [21]. Spectrum treatment and identification and quantification of metabolites were performed using the Bruker Topspin 3.1 and Chenomx NMR Suite (Version 7.1; Chenomx Inc., Edmonton, Canada) application packages, in conjunction with the Human Metabolome Database [22] and literature already published. All the chemical shift intervals (ppm) are listed in Table 1. In parallel with the chemometric analysis using the separation of the spectrum in small intervals, the statistical analysis was performed directly with the concentrations of the identified metabolites, through the targeted profiling methodology, developed by the creators of the Chenomx NMR Suite application. In this methodology the groups of peaks corresponding to each metabolite are identified and quantified using a database of pure substance spectra. We also used the Spectral binning methodology, in which the spectra are divided into predefined frequency intervals, while the integrals of the signals within each interval are used in the statistical analysis [23]. Statistical analysis Statistical analysis was performed through analysis of variance (ANOVA, p ≤ 0.05) with Tukey’s post hoc test for histological and western blotting data. For the metabolomics analysis, Principal Component Analysis (PCA) unsupervised was used in order to visualize the complex sample space and multivariate after identification and quantification of metabolites. These discriminant analyses were performed through standard procedures implemented in Pirouette 4.0 (Infometrix, Washington, USA) [24]. The input variables will consist of the integral of the area intervals (spectral binning) and / or the concentration values obtained with Targeted Profiling. Analysis of Variance (ANOVA, p ≤ 0.05) with Tukey’s post hoc test was performed directly with the concentrations of the identified metabolites, using the MetaboAnalyst platform. Results Tongue muscles remain spared in old mdx In the qualitative histopathological analysis of the sections, we observed different histological aspects in the muscle fibers. Fibers with peripheral nuclei (PN) were observed, indicating normal muscle tissue status; fibers with central nuclei (CN), indicating regenerated muscle fibers; and areas of fibrosis (FIB), shown in blue by Masson’s Trichrome stain (Fig. 1). In the quantitative histological analysis of the TON of the groups described (Table 2), there was no difference (p > 0.05) in the areas of PN, CN, or FIB, both between ages, young and old, and between lineages, wild type and mdx. Changes in proteins related to inflammation and fibrosis The quantification of TNF-α and TGF-β in the TON muscle was performed by Western blotting to verify the presence of inflammation and fibrosis, respectively. There was no significant difference in the concentrations of both proteins in the TON muscle between the groups (two-way ANOVA, TGF-β p = 9.968; TNF-α p = 7.558), as shown in Fig. 2. Changes in the metabolomic profile The tongues of the mice are grouped according to their metabolic profile through the PCA. It was possible to distinguish animals of the same lineage at different ages (Fig. 3, C and D), however it was not possible to distinguish mdx mice from wild type mice, in both young or old for the tongue muscle (Fig. 3, A and B). No differences were found in the comparison between the concentrations of metabolites from wild type or mdx animals of the same age, whether young or old (Fig. 4), suggesting that the protection of the dystrophic tongue muscle is observed in both ages. Table 3 summarizes the metabolites responsible for the differentiation of the tongue muscle of same strain mice at different ages. The metabolites alanine, methionine, 3-methylhistidine were higher in young animals in both wild type and mdx (p < 0.001), and the metabolites taurine and glycerol were lower in both the young wild type and mdx groups (p < 0.001). The metabolites glycine (p < 0.001) and glutamic acid (p = 0.0018) were different only in the mdx groups, being more concentrated in young mdx mice. The metabolites acetic acid, phosphocreatine, isoleucine, succinic acid, creatine had no difference in the analysis between groups (p > 0.05). Discussion This work aimed to validate biomarkers for diagnosis and progression of DMD through the analysis of metabolites, protein and histology of the tongue muscles of dystrophic and non-dystrophic animals, young and old. Surprisingly, our histological results demonstrated protection of TON against fibrosis and myonecrosis in both young (30 days) and old mdx animals (21 to 25 months). Chamberlain et al. (2007) studied the 26-month mdx TON and described those muscles as displaying significant histological abnormalities such as fibrosis and focal loss of muscle fibers. However, the muscle was analyzed qualitatively and comparisons were not made to TON from younger animals [10]. In contrast, here the qualitative and quantitative analysis of mid-belly TON did not show such severity. This protection was evidenced by the biomarker proteins for inflammation and fibrosis, and by the concentrations of the metabolites 3-methylhistidine, acetic acid, glutamic acid, alanine, creatine, phosphocreatine, glycerol, glycine, isoleucine, methionine, succinic acid and taurine already used for dystrophic muscle differentiation and that in this work presented similar results between wild type and mdx. This similarity between the chosen biomarkers was observed in young and old animals, corroborating the histological result. Kunert-Keil and colleagues [25] believe that dystrophic MAS and temporalis (TEM) are histologically similar to other skeletal muscles involved in the degeneration process, whereas the tongue remains with a milder phenotype. However, from the findings stated in the article, MAS and TEM are more resistant to the calcium (Ca2+) overload when compared to TON muscles of 100 days old mdx. Hence, histological findings assert that inflammatory foci is hardly detectable and dystrophic TON contains only 11.2% of regenerated muscle in the calcium-regulating genes, when compared to MAS, TEM and even soleus [25]. In DMD rats, the TON showed hypertrophy of myofibers with less advanced dystrophic changes until 8 months old compared to MAS. This resistance against degeneration might be related to a higher level of utrophin transcription in TON of wild-type and DMD rats compared to MAS of wild-type rats (Yamanouchi 2022). The study of biomarkers in a muscle protected from the absence of dystrophin allows us to identify dystrophinopathy markers that change with age, regardless of muscle degeneration. Among the possible biomarkers analyzed, the metabolites alanine, methionine, 3-methylhistidine, taurine and glycerol change with aging, but not between control and dystrophic tongue muscles. Thus, it is suggested that these biomarkers may be efficient for specific assessments, but care should be taken when using them to monitor the progression of the disease, since they change throughout life. The analysis of the metabolic profile of TON suggests that the muscle aging process has a high impact on its metabolism, regardless of the lineage, since it was possible to distinguish the different ages of animals of the same lineage (Fig. 3, C and D). High-resolution 1H NMR spectroscopy has been shown to differentiate skeletal muscle from adult and old mice. In addition, there is a general difference in composition between younger and older muscles in mice [30], and these results are confirmed in the scenario of dystrophinopathy. It was not possible to distinguish the metabolic profile of the tongue between the wild type and mdx strains. This result corroborates our histological analysis, which showed protection against fibrosis and myonecrosis in the tongue of young and old dystrophic animals. The chosen metabolites (3-methylhistidine, acetate, alanine, creatine, glutamic acid, glycerol, glycine, isoleucine, methionine, phosphocreatine, succinic acid and taurine) were altered between the wild type and mdx animals in muscles that suffer degeneration, such as the diaphragm, quadriceps and soleus[13][31][32]. Since these metabolites were not different between the strains in this study, it suggests that the protection previously observed in the TON of 3-month-old adult animals [17] is also observed in young (1 month-old) and old (21–25 months-old) animals. Taurine is considered a biomarker for the aging of skeletal muscle in mice [30], corroborating the data from this project, since its concentration was higher in old dystrophic and non-dystrophic mice. Its decrease in young mdx mice, compared to wild type, has already been seen in other muscles, indicating the possibility of a deficient taurine synthesis by the mdx muscles [31]. In the tongue, this decrease occurs both in the mdx and in the wild type compared to older age, corroborating the protection of the tongue muscle in dystrophic pathology. Taurine has an osmoregulatory function that helps to balance intracellular Ca2+ levels, helping with cell integrity and membrane stability [13]). Kunert-Keil et al (2014) [25] studied the differential expressions of genes involved in Ca2+ homeostasis in dystrophic masticatory muscles and found uneven expressions in the studied muscles. Despite being a priori protected muscle, in its study the tongue of 100 days-old mice presented levels of expression of Ca2+ regulatory proteins typical of dystrophic muscles. Analyzing Ca2+ regulatory proteins at the age of 21 to 25 months old, would provide a better view on the relationship of increased taurine, found in this study, with Ca2+ homeostasis in old mice. In addition, in muscle regeneration there is an increase in taurine, regardless of the type of muscle or genetic etiology of the damage to the fiber, suggesting that metabolic changes are significant indicators of muscle status [33]. Its greater concentration in the old group TON, in this project, suggests the activation of fiber metabolism to prevent muscle degeneration in the advanced stage of dystrophinopathy. Glycerol is a component of triglycerides (fats and oils) and phospholipids [22]. In muscles affected by DMD, the glycerol concentration was increased in relation to the wild type, at 6 months of age, being one of the metabolites responsible for the differentiation between the dystrophic and non-dystrophic muscles [13]. The results of this work showed that glycerol is more concentrated in the TON of old animals, both in wild type and mdx mice, with no difference between strains, suggesting its relationship with the aging process in the tongue mice muscle. Alanine is a non-essential amino acid resulting from the conversion of pyruvate or the breakdown of DNA and carnosine and anserine dipeptides [22]. It can be used as a fuel for gluconeogenesis directly from muscle tissue and therefore plays an important role in glucose homeostasis [34][35]. With the progression of dystrophinopathy in affected muscles, there is an increase in energy expenditure [10][36], for the incorporation of amino acids into new proteins [37]. Therefore, the lower concentration of alanine in the tongue muscle of old animals may indicate incorporation of alanine into proteins for processes of muscle maintenance and regeneration related to age, since the decrease is seen in old dystrophic and non-dystrophic mice. Methionine is an essential amino acid, substrate for protein synthesis, necessary for normal mammalian growth and development [22]. Martins-Bach et al. (2012) [13] showed that methionine increased with aging in quadriceps muscle samples from control mice from 3 months to 6 months of age. Besides, it showed that methionine concentrations did not change for mdx samples. However, in the tongue muscles, with a greater interval between the ages studied, methionine changed for both strains, being more concentrated in younger animals. The growth of the animals can elucidate this difference between the concentrations of methionine throughout the aging of the mice. From the 26th day of life until the 150th – 200th, approximately 7 months of age, the mice’s growth is observed, then remaining on a plateau until the end of life [38]. The high concentration of methionine at a young age in this study suggests growth in young animals. In older animals, which are not in the growth phase, the metabolite is decreased. Furthermore, this similarity of methionine concentrations between strains suggests that the TON muscle in mdx is metabolically closer to wild type, since this metabolite participates in the regulation of the immune system, lipid metabolism, oxidative stress and other metabolic regulation processes (for review see Martínez et al 2017). New studies should be carried out to characterize the involvement of this metabolite in the metabolism of DMD. 3-Methylhistidine (3-MeH) is an amino acid present in actin and myosin. It has been determined that more than 90% of body 3-MeH is located in the skeletal muscle. When skeletal muscle is degraded, 3-MeH is released, but is not reused for protein synthesis [39], therefore the urinary excretion of 3-MeH can be used to indicate muscle protein degradation [40]. Furthermore, there is a progressive loss of skeletal muscle mass and muscle strength with aging [41][42]. The higher concentration of 3-MeH observed in the tongue of the young groups may reflect the decrease in muscle mass and the consequent reduction in 3-MeH concentrations seen in older muscles. Muscle degeneration biomarkers, which in other studies proved to be efficient (acetic acid, phosphocreatine, isoleucine, succinic acid, creatine, TNF-α and TGF-β protein) [13][18][27] did not show any difference with aging and in the analysis between strains. The TGF-β protein is actively involved in the proliferation of fibrous connective tissue in the skeletal muscles of patients and mdx mice [43][44]. In previous studies, the levels of TGF-β have already been related to the presence of fibrous tissue in mdx’s respiratory muscles [27]. Spassov et al. [17] studied changes in the expression of collagen in the masticatory muscles in 100 days old animals and concluded that there was no difference between the tongue of mdx and control groups. Our results corroborate yours, due to the equality in the quantification of TGF-β in young animals indicating protection against tissue fibrosis of this muscle in the initial phase of the disease. In old animals, the levels of TGF-β were also equal between strains and suggest that this muscle protection against fibrosis continues until advanced stages of dystrophy in the mice, corroborating the histological results for fibrosis found in our study, for both ages. TNF-α is a pro-inflammatory protein produced by activated macrophages, mast cells, endothelial cells and some other cell types. It stimulates the expression of adhesion molecules in endothelial cells, increasing leukocyte recruitment and adhesion. TNF-α is secreted at the site of inflammation and can enter the bloodstream [26]. The TNF-α is related to the progression of DMD and its absence is related to muscle protection for the disease. Maranhão et al. [27] evidenced the progressive increase in the concentration of TNF-α in mdx diaphragm muscles, in comparison with the control and over age, with 1, 4 and 9 months. The intrinsic muscles of the larynx (ILM) were identified as muscles protected from dystrophinopathy in the mdx mice, as they did not show signs of muscle damage, degeneration or regeneration, during the course of the disease [28]. The concentrations of TNF-α in ILM were comparable to those of the control animal, during the progression of the disease, even in later stages, at 20 months of age [27]. These results corroborate with this work, since the concentrations of TNF-α in the TON of the mdx animals did not differ from the wild type animals, in young and old. Messina et al. (2011) [29] analyzed vastus lateralis muscle samples from DMD patients and showed an increase in the concentration of TNF-α in relation to the control group, besides pointing out the increase in its expression with age and disease progression, from 2 to 9 years of age. In this work, the concentration of TNF-α did not increase with aging in TON in both strains, corroborating with the protection previously mentioned, including with disease progression and aging. The metabolites glycine and glutamic acid were different only in dystrophic animals, being more concentrated in young mdx mice. Glycine is a non-essential amino acid involved in the production of DNA, phospholipids and collagen, in addition to being involved in the release of energy [22]. D.J. Ham et al. [45] demonstrated that glycine supplementation in mdx mice and dystrophin/utrophin double knockout mice can attenuate the progression of dystrophic pathology, as well as improve the effectiveness of prednisolone, the current gold standard treatment for DMD. Anderson and Skrabek [46] studied the heart of mdx mice treated with deflazacort, from a metabolomic perspective. They showed that the concentration of glycine in the mdx heart decreased with disease progression, but with a 2-week high-dose treatment with deflazacort, glycine levels increased above the levels of the mdx placebo group. Glycine supplementation has already been shown to protect against loss of myotubes in nutrient restricted/growth factor restriction models with C2C12 muscle cells in vitro. This in vitro protection has been shown to be dependent on mammalian target of rapamycin complex 1 (mTORC1) signaling [47] and also a specific glycine activation of mTORC1 involved in muscle regeneration in dystrophic mice [48]. Thus, the protection observed in the 1 month old tongue of the mdx muscle, close to the 20th day degeneration peak, may be related to the high levels of glycine presented in our study. Glutamic acid, also known as glutamate (the anion), is a non-essential amino acid, one of the 20 proteinogenic amino acids and the most abundant rapid excitatory neurotransmitter in the nervous system [22]. Glutathione is the main cellular antioxidant and regulates free radical homeostasis. Glutamate plays an essential regulatory role in the synthesis of glutathione, and is also related to its functions (for review see [49][50]. It is proven the relationship between decreased intracellular glutathione and stress states, such as chronic diseases [50], besides demonstrating benefits in the use of different antioxidant drugs for preclinical studies in mdx mice, such as improvement in dystrophinopathy with decreased necrosis [51]. In a metabolomic study of the biceps femoris muscle of the canine model for DMD (Golden Retriever Muscular Dystrophy), there was an increase in the concentration of glutamic acid when compared to the control [52]. Laferte, Rosenkrantz and Berlinguet [53] observed an increase in weight loss and acceleration of the beginning of the terminal phase of the disease in mice that received exogenous glutamic acid. The increased glutamate concentration in the dystrophic muscles of the quadriceps and diaphragm suggests its connection with muscle regeneration [13] and its relationship with the pathology of DMD. In this study, the amounts of glutamic acid in the mdx were equivalent to those of control animals of the same age, indicating protection against myonecrosis on the tongue. These results corroborate the protection observed in histology and the use of these as muscle biomarkers of degeneration. Other studies on affected muscles can elucidate their specificities as biomarkers for the progression of DMD, without the interference of aging. Conclusions The tongue of young (1 month) and old (21 to 25 months) mdx animals remains protected from the intense and progressive myonecrosis described in other muscles, as evidenced by histological analysis. This protection was verified by the biomarker proteins for inflammation and fibrosis, and by the concentrations of the metabolites 3-methylhistidine, acetic acid, glutamic acid, alanine, creatine, phosphocreatine, glycerol, glycine, isoleucine, methionine, succinic acid and taurine already used for dystrophic muscle differentiation and in this work presented similar results between wild type and mdx. Among the possible biomarkers analyzed, the metabolites alanine, methionine, 3-methylhistidine, taurine and glycerol may be efficient for specific assessments, but care should be taken when using them to monitor the progression of the disease, since they change throughout life. Muscle degeneration biomarkers, which in other studies proved to be efficient (acetic acid, phosphocreatine, isoleucine, succinate, creatine, TNF-α and TGF-β) did not show any difference with aging and in the spared muscles. Other studies on affected muscles can elucidate their specificities as biomarkers for the progression of DMD, without the interference of aging. Funding This work was supported by São Paulo Research Foundation (FAPESP) grant#2018/25598-0, grant#2019/20162-1 and grant#2013/00312-2 (to MSVL and CYM). CYM was supported by CNPq grant 456357/2014-9. JSC is supported by NIH grant R01AR040864. GLO is supported by DOD grant award#W81XWH-18-1-0624. Availability of data and materials NOT APPLICABLE Abbreviations 3-MeH 3-Methylhistidine ANOVA analysis of variance CEUA Brazilian College for Animal Experimentation CN central nuclei D2O deuterated water DMD Duchenne muscular dystrophy FIB areas of fibrosis GAPDH glyceraldehyde 3-phosphate dehydrogenase protein ILM intrinsic muscles of the larynx MAS masseter muscle mdx X chromosome-linked muscular dystrophy mTORC1 mammalian target of rapamycin complex 1 NMR Nuclear Magnetic Resonance PN areas of muscle fibers with peripheral nucle PCA Principal Component Analysis TEM temporalis TON tongue muscles TNF-α Tumor Necrosis Factor alpha TGF-β Transforming Growth Factor beta TSP trimethylsilyl tetradeuteropropionic acid UNESP São Paulo State University Figure 1 Cross-sections of mid-belly TON muscle. Young wild type (A), young mdx (B), old wild type (C) and old mdx (D) groups. In B, areas of fibrosis highlighted in blue by Masson’s Trichrome (*). In C, muscle fibers with a central nucleus (arrow) and muscle fibers with a peripheral nucleus (arrowhead). Scale: 50 μm Figure 2 Quantification of TNF-α and TGF-β. The quantification of TNF-α and TGF-β was performed by Western blotting analysis in crude extracts of TON muscles from young wild type, young mdx, old wild type and old mdx groups. Same blot reprobed for GAPDH as a loading control. Graphs represent the level of proteins expressed in arbitrary units and normalized to GAPDH levels. Bars represent standard deviation. No significant difference was observed between the groups (p>0.05, ANOVA). Figure 3 Principal component analysis (PCA) of TON with the scores plot between groups. mdx animals (mdx) vs. wild type (wt) young (A) and old (B); tongue of young vs. old mdx animals (C); tongue of young vs. old wild type animals (D). Each point on the graph represents the spectrum of an animal. Figure 4 Differences in the concentrations of metabolites between groups. Difference between metabolite concentrations between young and old mdxanimals (a). Difference between metabolite concentrations between young and old wild type animals (b), ANOVA, p< 0.001. The metabolites acetic acid, creatine, phosphocreatine, isoleucine and succinic acid showed no differences between groups (p> 0.05). Table 1 Assignments of resonance peaks from 1H Nuclear Magnetic Resonance (NMR) data. Areas were calculated considering the indicated chemical shift intervals. Chemical shifts interval (ppm) Compound 1.0056–1.0283 Isoleucine 1.4639–1.5053 Alanine 1.9155–1.9262 Acetic acid 2.4002–2.4134 Succinic acid 3.0253–3.0459 Creatine 3.0458–3.0533 Phosphocreatine 3.3964–3.4469 Taurine 3.5587–3.5648 Glycine 2.6372–2.6426 Methionine 2.3322–2.3754 Glutamic acid 7.0091–7.0516 3-Methylhistidine 3.639–3.6507 Glycerol −0.1362–0.1198 TSP Table 2 Quantitative histological analysis of TON. %PN %CN %FIB young wild type 91.33 ± 0.98 1.24 ± 0.63 7.42 ± 1.37 youngmdx 80.94 ± 13.62 2.40 ± 1.17 16.67 ± 13.30 old wild type 83.43 ± 10.68 4.99 ± 2.77 11.58 ± 9.67 oldmdx 75.67 ± 5.49 6.18 ± 3.87 18.16 ± 7.29 Mean ± SD of fiber percentage areas with peripheral nucleus (%PN), central nucleus (%CN) and areas of fibrosis (%FIB) in the TON of the young and old mdx and wild type groups. There were no differences found for areas of PN, CN and FIB between the four groups (ANOVA, p > 0,05). Table 3 Different metabolites between animals of the same lineage at different ages in tongue samples. Higher in young Lower in young wild type (young vs. old) alanine, methionine, 3-methylhistidine taurine, glycerol mdx (young vs. old) alanine, methionine, 3-methylhistidine, glutamic acid, glycine taurine, glycerol Competing interests The authors declare that they have no competing interests Declarations Ethics approval and consent to participate All experiments were carried out in accordance with the guidelines for animal experimentation of the Ethics Committee on Animal Use - CEUA-IBB/UNESP, protocol#1095. Consent for publication NOT APPLICABLE ==== Refs References 1. Mendell JR , Shilling C , Leslie ND , Flanigan KM , al-Dahhak R , Gastier-Foster J , Evidence-based path to newborn screening for duchenne muscular dystrophy. Annals of Neurology. 2012 Mar;71 (3 ):304–13. DOI: 10.1002/ana.23528 22451200 2. Hoffman EP , Brown RH , Kunkel LM . Dystrophin: The protein product of the duchenne muscular dystrophy locus. Cell. 1987 Dec; 51 (6 ):919–28. DOI: 10.1016/0092-8674(87)90579-4 3319190 3. Tuffery-Giraud S , Béroud C , Leturcq F , Yaou RB , Hamroun D , Michel-Calemard L , Genotype-phenotype analysis in 2,405 patients with a dystrophinopathy using the UMD-DMD database: a model of nationwide knowledgebase. Human Mutation. 2009 Jun;30 (6 ):934–45. DOI: 10.1002/humu.20976 19367636 4. Tinsley JM , Blake DJ , Zuellig RA , Davies KE . Increasing complexity of the dystrophin-associated protein complex. Proceedings of the National Academy of Sciences. 1994 Aug 30];91 (18 ):8307–13. DOI: 10.1073/pnas.91.18.8307 5. Bogdanovich S , Perkins KJ , Krag TOB , Khurana TS . Therapeutics for Duchenne muscular dystrophy: current approaches and future directions. Journal of Molecular Medicine. 2004 Feb 1;82 (2 ):102–15. DOI: 10.1007/s00109-003-0484-1 14673527 6. Landfeldt E , Thompson R , Sejersen T , McMillan HJ , Kirschner J , Lochmüller H . Life expectancy at birth in Duchenne muscular dystrophy: a systematic review and meta-analysis. European Journal of Epidemiology. 2020 Feb 27;35 (7 ):643–53. DOI: 10.1007/s00109-003-0484-1 32107739 7. Bulfield G , Siller WG , Wight PA , Moore KJ . X chromosome-linked muscular dystrophy (mdx) in the mouse. Proceedings of the National Academy of Sciences. 1984 Feb];81 (4 ):1189–92. DOI: 10.1073/pnas.81.4.1189 8. McGreevy JW , Hakim CH , McIntosh MA , Duan D . Animal models of Duchenne muscular dystrophy: from basic mechanisms to gene therapy. Disease Models & Mechanisms. 2015 Mar 1;8 (3 ):195–213. DOI:10.1242/dmm.018424 25740330 9. Coulton GR , Morgan JE , Partridge TA , Sloper JC . The mdx mouse skeletal muscle myopathy: I. A histological, morphometric and biochemical investigation. Neuropathology and Applied Neurobiology. 1988 Feb;14 (1 ):53–70. DOI: 10.1111/j.1365-2990.1988.tb00866.x 2967442 10. Chamberlain JS , Metzger J , Reyes M , Townsend D , Faulkner JA . Dystrophin-deficient mdx mice display a reduced life span and are susceptible to spontaneous rhabdomyosarcoma. The FASEB Journal. 2007 Mar 14;21 (9 ):2195–204. DOI: 10.1096/fj.06-7353com 17360850 11. Donald S , Chen R , Snyder M . Systems biology approaches to disease marker discovery. Disease markers. 2013;28 (4 ). DOI: 10.3233/DMA-2010-0707 12. Gulston MK , Rubtsov DV , Atherton HJ , Clarke K , Davies KE , Lilley KS , A Combined Metabolomic and Proteomic Investigation of the Effects of a Failure to Express Dystrophin in the Mouse Heart. Journal of Proteome Research [Internet]. 2008 Apr 3;7 (5 ):2069–77. DOI: 10.1021/pr800070p 18386883 13. Martins-Bach AB , Bloise AC , Vainzof M , Rahnamaye Rabbani S . Metabolic profile of dystrophic mdx mouse muscles analyzed with in vitro magnetic resonance spectroscopy (MRS). Magnetic Resonance Imaging. 2012 Oct;30 (8 ):1167–76. DOI: 10.1016/j.mri.2012.04.003 22673895 14. Maurício AF , de Carvalho SC , Santo Neto H , Marques MJ . Effects of dietary omega-3 on dystrophic cardiac and diaphragm muscles as evaluated by 1H magnetic resonance spectroscopy: Metabolic profile and calcium-related proteins. Clinical Nutrition ESPEN. 2017 Aug;20 :60–7. DOI: 10.1016/j.clnesp.2017.03.005 29072171 15. Archer SK , Garrod R , Hart N , Miller S . Dysphagia in Duchenne muscular dystrophy assessed by validated questionnaire. International Journal of Language & Communication Disorders. 2013 Mar;48 (2 ):240–6. DOI: 10.1111/j.1460-6984.2012.00197.x 23472962 16. van den Engel-Hoek L , de Groot IJM , Sie LT , van Bruggen HW , de Groot SAF , Erasmus CE , Dystrophic changes in masticatory muscles related chewing problems and malocclusions in Duchenne muscular dystrophy. Neuromuscular Disorders. 2016 Jun;26 (6 ):354–60. DOI: 10.1016/j.nmd.2016.03.008 27132120 17. Spassov A , Gredes T , Gedrange T , Lucke S , Pavlovic D , Kunert-Keil C . Histological changes in masticatory muscles of mdx mice. Archives of Oral Biology. 2010 Apr];55 (4 ):318–24. DOI: 10.1016/j.archoralbio.2010.02.005 20207343 18. Al-Khalili Szigyarto C , Spitali P . Biomarkers of Duchenne muscular dystrophy: current findings. Degenerative Neurological and Neuromuscular Disease. 2018 Jan;Volume 8 :1–13. DOI: 10.2147/DNND.S121099 30050384 19. Grounds MD , Terrill JR , Al-Mshhdani BA , Duong MN , Radley-Crabb HG , Arthur PG . Biomarkers for Duchenne muscular dystrophy: myonecrosis, inflammation and oxidative stress. Disease Models & Mechanisms. 2020 Mar;Volume 13 (2 ). DOI:10.1242/dmm.043638 20. Ferretti R. , Marques M. J. , Pertille A. , & Santo Neto H. (2009). Sarcoplasmic-endoplasmic-reticulum Ca2+-ATPase and calsequestrin are overexpressed in spared intrinsic laryngeal muscles of dystrophin-deficient mdx mice. Muscle & Nerve. 2019 May;39 (5 ), 609–615. DOI: 10.1002/mus.21154 21. Sukumaran DK , Garcia E , Hua J , Tabaczynski W , Odunsi K , Andrews C , Standard operating procedure for metabonomics studies of blood serum and plasma samples using a 1H-NMR micro-flow probe. Magnetic Resonance in Chemistry. 2009 Aug 18;47 (S1 ):S81–5. DOI: 10.1002/mrc.2469 19688872 22. Wishart D. S. , Guo A. , Oler E. , Wang F. , Anjum A. , Peters H. , Dizon R. , Sayeeda Z. , Tian S. , Lee B. L. , Berjanskii M. , Mah R. , Yamamoto M. , Jovel J. , Torres-Calzada C. , Hiebert-Giesbrecht M. , Lui V. W. , Varshavi D. , Varshavi D. , Allen D. , … Gautam V. (2022). HMDB 5.0: the Human Metabolome Database for 2022. Nucleic acids research, 50 (D1 ), D622–D631. 10.1093/nar/gkab1062 34986597 23. Weljie AM , Newton J , Mercier P , Carlson E , Slupsky CM . Targeted Profiling: Quantitative Analysis of 1H NMR Metabolomics Data. Analytical Chemistry [Internet]. 2006 May 17;78 (13 ):4430–42. DOI: 10.1021/ac060209g 16808451 24. Beckonert O , Keun HC , Ebbels TMD , Bundy J , Holmes E , Lindon JC , Metabolic profiling, metabolomic and metabonomic procedures for NMR spectroscopy of urine, plasma, serum and tissue extracts. Nature Protocol. 2007 Oct 25.];2 (11 ):2692–703. DOI: 10.1038/nprot.2007.376 25. Kunert-Keil CH ; Gredes T ; Lucke S ; Botzenhart U ; Dominiak M ; Gedrange T . Differential expression of genes involved in the calcium homeostasis in masticatory muscles of MDX mice. Journal of physiology and pharmacology: an official journal of the Polish Physiological Society. 2014;65 (2 ). 26. Kumar V. ; Abbas A. ; Fausto N. Robbins e Cotran Pathologic Basis of Disease. 8th. ed. Rio de Janeiro: Elsevier, 2010; 27. Maranhão J , de Oliveira Moreira D , Maurício AF , de Carvalho SC , Ferretti R , Pereira JA , Changes in calsequestrin, TNF-α, TGF-β and MyoD levels during the progression of skeletal muscle dystrophy in mdx mice: a comparative analysis of the quadriceps, diaphragm and intrinsic laryngeal muscles. International Journal of Experimental Pathology. 2015 Oct;96 (5 ):285–93. DOI: 10.1111/iep.12142 26515458 28. Marques MJ , Ferretti R , Vomero VU , Minatel E , Neto HS . Intrinsic laryngeal muscles are spared from myonecrosis in the mdx mouse model of Duchenne muscular dystrophy. Muscle & Nerve. 2007;35 (3 ):349–53. DOI: 10.1002/mus.20697 17143878 29. Messina S , Vita GL , Aguennouz M , Sframeli M , Romeo S , Rodolico C , Activation of NF-kappaB pathway in Duchenne muscular dystrophy: relation to age. Acta myologica: myopathies and cardiomyopathies : official journal of the Mediterranean Society of Myology. 2011;30 (1 ):16–23.21842588 30. Sobolev A. P. , Mannina L. , Costanzo M. , Cisterna B. , Malatesta M. , & Zancanaro C. (2017). Age-related changes in skeletal muscle composition: A pilot nuclear magnetic resonance spectroscopy study in mice. Experimental Gerontology. 2017 Jun;92 , 23–27. DOI: 10.1016/j.exger.2017.03.005 28286172 31. Terrill JR , Grounds MD , Arthur PG . Taurine deficiency, synthesis and transport in the mdx mouse model for Duchenne Muscular Dystrophy. The International Journal of Biochemistry & Cell Biology 2015 Sep;66 :141–8. DOI: 10.1016/j.biocel.2015.07.016 26239309 32. Griffin JL , Williams HJ , Sang E , Clarke K , Rae C , Nicholson JK . Metabolic Profiling of Genetic Disorders: A Multitissue 1H Nuclear Magnetic Resonance Spectroscopic and Pattern Recognition Study into Dystrophic Tissue. Analytical Biochemistry. 2001 Jun;293 (1 ):16–21. DOI: 10.1006/abio.2001.5096 11373073 33. McIntosh LM , Garrett KL , Megeney L , Rudnicki MA , Anderson JE . Regeneration and myogenic cell proliferation correlate with taurine levels in dystrophin- and MyoD-Deficient muscles. The Anatomical Record. 1998 Oct];252 (2 ):311–24. DOI: 10.1002/(SICI)1097-0185(199810)252:2<311::AID-AR17>3.0.CO;2-Q 9776086 34. Nishio H , Wada H , Matsuo T , Horikawa H , Takahashi K , Nakajima T , Glucose, free fatty acid and ketone body metabolism in duchenne muscular dystrophy. Brain and Development. 1990 Jan;12 (4 ):390–402. DOI: 10.1016/s0387-7604(12)80071-4 2240459 35. Srivastava NK , Yadav R , Mukherjee S , Sinha N . Perturbation of muscle metabolism in patients with muscular dystrophy in early or acute phase of disease: In vitro, high resolution NMR spectroscopy based analysis. Clinica Chimica Acta. 2018 Mar;478 :171–81. DOI: 10.1016/j.cca.2017.12.036 36. Radley-Crabb HG , Marini JC , Sosa HA , Castillo LI , Grounds MD , Fiorotto ML . Dystropathology Increases Energy Expenditure and Protein Turnover in the Mdx Mouse Model of Duchenne Muscular Dystrophy. Fraidenraich D, editor. PLoS ONE. 2014 Feb 19;9 (2 ):e89277. DOI:10.1371/journal.pone.0089277 24586653 37. Strakova J , Kamdar F , Kulhanek D , Razzoli M , Garry DJ , Ervasti JM , Integrative effects of dystrophin loss on metabolic function of the mdx mouse. Scientific Reports. 2018 Sep 11.;8 (1 ). DOI: 10.1038/s41598-018-31753-3 38. Schäfer R , Zweyer M , Knauf U , Mundegar RR , Wernig A . The ontogeny of soleus muscles in mdx and wild type mice. Neuromuscular Disorders. 2005 Jan;15 (1 ):57–64. DOI: 10.1016/j.nmd.2004.09.011 15639122 39. Williams A.P. AMINO ACIDS | Determination. Encyclopedia Of Food Sciences And Nutrition, [s.l.], p.192–197, 2003. Elsevier. DOI: 40. Keller U. Nutritional Laboratory Markers in Malnutrition. Journal of Clinical Medicine. 2019 May 31;8 (6 ):775. DOI: 10.3390/jcm8060775 31159248 41. Janssen I , Heymsfield SB , Wang Z , Ross R . Skeletal muscle mass and distribution in 468 men and women aged 18–88 year. Journal of Applied Physiology. 2000 Jul 1;89 (1 ):81–8. DOI: 10.1152/jappl.2000.89.1.81 10904038 42. Nilwik R , Snijders T , Leenders M , Groen BBL , van Kranenburg J , Verdijk LB , The decline in skeletal muscle mass with aging is mainly attributed to a reduction in type II muscle fiber size. Experimental Gerontology. 2013 May;48 (5 ):492–8. DOI: 10.1016/j.exger.2013.02.012 23425621 43. Bernasconi P , Torchiana E , Confalonieri P , Brugnoni R , Barresi R , Mora M , Expression of transforming growth factor-beta 1 in dystrophic patient muscles correlates with fibrosis. Pathogenetic role of a fibrogenic cytokine. Journal of Clinical Investigation. 1995 Aug 1;96 (2 ):1137–44. DOI: 10.1172/JCI118101 7635950 44. Carvalho SC , Matsumura CY , Santo Neto H , Marques MJ . Identification of plasma interleukins as biomarkers for deflazacort and omega-3 based Duchenne muscular dystrophy therapy. Cytokine. 2018 Feb;102 :55–61. DOI: 10.1016/j.cyto.2017.12.006 29276972 45. Ham DJ , Gardner A , Kennedy TL , Trieu J , Naim T , Chee A , Glycine administration attenuates progression of dystrophic pathology in prednisolone-treated dystrophin/utrophin null mice. Scientific Reports [Internet]. 2019 Sep 10;9 (1 ). DOI: 10.1038/s41598-019-49140-x 46. Skrabek RQ ; Anderson JE . Metabolic shifts and myocyte hypertrophy in deflazacort treatment of mdx mouse cardiomyopathy. Muscle & Nerve. 2016;24 (2 ). DOI: 10.1002/1097-4598(200102)24:2<192::aid-mus40>3.0.co;2-s 47. Caldow MK , Ham DJ , Trieu J , Chung JD , Lynch GS , Koopman R . Glycine Protects Muscle Cells From Wasting in vitro via mTORC1 Signaling. Frontiers in Nutrition. 2019 Nov 13. DOI: 10.3389/fnut.2019.00172 48. Lin C , Han G , Ning H , Song J , Ran N , Yi X , Glycine Enhances Satellite Cell Proliferation, Cell Transplantation, and Oligonucleotide Efficacy in Dystrophic Muscle. Molecular Therapy. 2020 May;28 (5 ):1339–58. DOI:32209436 49. Wu G , Fang YZ ., Yang S , Lupton JR , Turner ND . Glutathione Metabolism and Its Implications for Health. The Journal of Nutrition. 2004 Mar 1.;134 (3 ):489–92. DOI: 10.1093/jn/134.3.489 14988435 50. Rutten EP , Engelen MP , Schols AM , Deutz NE . Skeletal muscle glutamate metabolism in health and disease: state of the art. Current Opinion in Clinical Nutrition and Metabolic Care. 2005 Jan;8 (1 ):41–51. DOI: 10.1097/00075197-200501000-00007 15585999 51. Petrillo S , Pelosi L , Piemonte F , Travaglini L , Forcina L , Catteruccia M , Oxidative stress in Duchenne muscular dystrophy: focus on the NRF2 redox pathway. Human Molecular Genetics. 2017 May 3;26 (14 ):2781–90. DOI:10.1093/hmg/ddx173 28472288 52. Abdullah M , Kornegay JN , Honcoop A , Parry TL , Balog-Alvarez CJ , O’Neal SK , Non-Targeted Metabolomics Analysis of Golden Retriever Muscular Dystrophy-Affected Muscles Reveals Alterations in Arginine and Proline Metabolism, and Elevations in Glutamic and Oleic Acid In Vivo. Metabolites. 2017 Jul 29;7 (3 ):E38. DOI: 10.3390/metabo7030038 53. Laferté RO , Rosenkrantz H , Berlinguet L . TRANSAMINATION IN MUSCULAR DYSTROPHY AND THE EFFECT OF EXOGENOUS GLUTAMATE: A STUDY ON VITAMIN E DEFICIENT RABBITS, AND MICE WITH HEREDITARY DYSTROPHY. Canadian Journal of Biochemistry and Physiology. 1963 Jan 1;41 (1 ):1423–32. DOI: 10.1139/y63-162 13928008

PMC010xxxxxx/PMC10312974.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398318 10.21203/rs.3.rs-2684226/v1 10.21203/rs.3.rs-2684226 preprint 1 Article Gender as a determinant of health in under-five children in Ethiopia; a secondary data analysis from EDHS 2016 Ketema Elbet Addis Ababa University Hassen Saria Emory University Author’s contribution Dr. Elbet ketema was involved in developing the research concept as well as preparing the manuscript. Dr. Elbet was also involved in analyzing the secondary data and also manuscript writing. Dr. Saria Hassan was involved in guiding and editing the research work all through proposal development, analysis and write up process. All authors read and approved the final manuscript. ✉ Elbetkete@gmail.com 31 5 2023 rs.3.rs-2684226https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs2684226v1.pdf Background: Under-five mortality is one of the key sustainable development goal targets. Despite the great strides made globally, Under-five mortality remains high in many developing countries like Ethiopia. Child health status is determined by a myriad of factors at the individual, family and community level, furthermore, a child’s gender has been shown to affect the probability of infant and child mortality. Methods: A secondary data analysis was conducted using Ethiopian demographic health survey 2016 to assess association between gender and under-five child health. A representative sample of 18,008 households was selected. After data cleaning and entry, analysis was done using the Statistical Package for Social Sciences (SPSS) version 23. Uni-variable and multivariable logistic regression model were employed to determine the association between under-five child health and gender. In the final multivariable logistic regression model, the association of gender with childhood mortality was declared statistically significant at P value < 0.05. Result: A total of 2,075 under five children from EDHS 2016 were included in the analysis. Majority (92%) were rural dwellers. More male children were found to be underweight (53% Vs 47%) and wasted (56.2% Vs 43.8%) compared to female children. A higher proportions of females were vaccinated (52.2%) compared to 47.8% in males. Health seeking behavior for fever (54.4%) and diarrheal diseases (51.6%) were also found to be higher for females. However, in a multivariable logistic regression model, there was no statistically significant association found between gender and under-five child children health measures. Conclusion: Although it was not statistically significant association, females were found to have a better health and nutritional outcomes compared to boys in our study. Gender Under-five mortality Ethiopian demographic health survey 2016 Fogarty international center of the national institute of healthD43TW011404 ==== Body pmcIntroduction Improving child survival remains a matter of urgent concern. In 2019 alone, roughly 14,000 under-five deaths occurred worldwide every day which is an intolerably high number of largely preventable child deaths (1). This burden of child mortality is unevenly distributed throughout the world. Over 70% of all under-five death occurs in Africa and south East Asia. One in twelve children in sub Saharan Africa dies before their fifth birthday which is 15 times higher than the risk for children born in high-income countries and 20 years behind the world average (1). Ethiopia is one of the hardest hit countries with an alarming number of child deaths. In the country 1 in 15 children die before reaching age 5, and 7 in 10 of the deaths occur during infancy (2). To achieve the sustainable development goal (SDG) target of under-five mortality rate of 25 or fewer deaths per 1000 live births by 2030, a total of 47 countries need to increase their pace of progress including Ethiopia (3). Child health status is determined by a myriad of factors at the individual, family and community levels. A child’s gender has been shown to affect the probability of infant and child mortality. Owing to biological factors, male infants have a higher risk of mortality during the first year of life. In addition differential treatment of boys and girls, owing to cultural and socioeconomic factors also affect the chances of survival during childhood (4). Gender is a major determinant of health and as such should be properly understood and thoroughly investigated. Although gender inequality in child health has been consistently reported throughout the world, there is still substantial knowledge gap on how gender mediates child health in Ethiopia. In addition implementation of interventions to mitigate gender inequalities that hinder child health requires additional perspective and research. This study seeks to evaluate gender as a determinant of under-five health in Ethiopia and explore factors associated with it. This work will inform the disease prevention and control strategy and serve as a base line assessment laying the foundation to devise policies and programs to address gender inequality in health during childhood in Ethiopia. Method and Materials Administratively, regions in Ethiopia are divided into zones, and subsequently into administrative units called weredas. Each wereda is further subdivided into the lowest administrative unit, called kebele. During the 2007 census each kebele was subdivided into census enumeration areas (EAs), which were convenient for the implementation of the census. The data source for this analysis is the 2016 Ethiopian DHS which was undertaken over a 5-months period from 18 January, 2016 to 27 June, 2016 which was designed to provide population and health indicators at the national (urban and rural) and regional levels. The data is publicly available at this site (http://www.measuredhs.com/data/available-datasets.cfm). The 2007 Population and Housing Census, conducted by the central statistical agency (CSA), provided the sampling frame from which the 2016 EDHS sample was drawn. The sample was selected using a stratified, two stage cluster design and EAs were the sampling units for the first stage. In the second stage, a fixed number of 28 household per cluster were selected with an equal probability systematic selection from the newly created household listing. The sample included 645 EAs, 202 in urban areas and 443 in rural areas. Households comprised the second stage of sampling. A complete listing of households was carried out in each of the 645 selected EAs from September to December 2015. A representative sample of 18,008 households was selected (EDHS 2016). Dependent variable (Outcome of interest): Under-five mortality, vaccinations, seeking care in past 2 weeks for ARI/fever/diarrhea, stunted growth, under weight, exclusive breastfeeding, and anemia. Independent variable: Gender of the child. Gender of the child is defined as the socially constructed characteristics of a boy and a girl. Covariates Based on literature review, the following covariates were selected: Parental wealth index, birth order, and number of children in household, age of the mother, place of residence, region, Parental educational level, mode of delivery, antenatal care check-up, place of delivery, toilet facility, source of drinking water. Data analysis and Interpretation After data cleaning and entry, analysis was done using the Statistical Package for Social Sciences (SPSS) version 23. Descriptive statistics and cross tabulation was performed to describe the study variables. Bivariate association between each child’s gender was first examined using chi square test. Univariable and multivariable logistic regression models were employed to determine the association between under-five child health and gender. Variance inflation factor was used to assess presence of co linearity. Crudes odds ratio (COR) and adjusted odds ratio (AOR) was presented with 95% confidence interval. Each covariate was included in the multivariable model regardless of their statistical significance in the uni-variable analysis. In the final multivariable logistic regression model, the association of gender with childhood mortality was declared statistically significant if p-value < 0.05. Sample weights that account for complex survey design and unequal probabilities of selection were incorporated in all the analysis. Results 1. Socio-demographic characteristics A total of 2,075 under five children from EDHS 2016 were included in the analysis. The majority (92%) were rural dwellers. More males (57%) were from the urban area when compared to females (43%) but parents of males and females had a similar wealth index. In both males and females there were more infants than any other age group. Boys had relatively younger (54.7% Vs 45.3%) and uneducated mothers (51% Vs 49%) compared with girls (Table 1). 2. Under five child health In this survey, 38% of under-five children were stunted, 24% underweight and 10% wasted. However, more male children were underweight (53% Vs 47%) and wasted (56.2% Vs 43.8%) compared to female children. The proportion of male and female children who were stunted was similar (50.8% Vs 49.2%). Around 52.2% of under-five children who were at least vaccinated once were females compared to a 47.8% of males. Female children were also found to have higher health seeking for fever (54.4%) and diarrheal diseases (51.6%) than males (45.6% and 48.4$) respectively. More than half (57%) of children aged 6–59 months in this survey were anemic. There were more females with severe anemia (55.3%) when compared to males (44.7%) (Table 2). 3. Gender and under five child health Among 2,075 under five children evaluated in this study, there was no statistically significant association found between gender of a child and under-five children health outcomes. The odds of being wasted was three times more in females compared to males, AOR = 3.09 (CI 0.50, 18.89). The odds of care seeking behavior for fever was also 2.53 AOR = (CI 0.61, 10.48) times more in females when compared to boys. Similarly the odds of being exclusively breast fed was AOR = 2.79(CI 0.70, 11.02) times more in females. Again, the odds of being vaccinated among female under fives was AOR = 1.27(CI 0.47, 3.39) times those of males. However, these associations were not found to be statistically significant. Females were also found to be less stunted and underweight when compared to males with odds ratio of AOR = 0.48(0.16, 1.41) and AOR = 0.36 (0.086, 1.54) respectively (Table 3). Discussion Child health remains a global health priority (3). In Ethiopia, under-five mortality is declining steadily but still remains high (5). There are a myriad of factors determining child health including gender (6). In this study the EDHS 2016 was used to assess the association of under five child health parameters with gender. Our study did not find a statistically significant association between under-five child health and gender. The result of this study is unexpected since the majority of Ethiopians live in rural area where1905 samples out of 2070 in this study were from. The rural community is thought to be a very traditional community and females are usually disadvantaged. Though there was no statistically significant association found in this study, females were found to be less stunted AOR of 0.48 (CI: 0.16, 1.41) and underweight AOR of 0.36(0.086, 1.54) compared to boys. This is contrary to one study in Bangladesh which showed a substantially higher prevalence of malnutrition among female children than male. In depth dietary surveys in the same study also found males to consistently consume more calories and proteins than females at all ages (9). A similar study from rural eastern Kenya were boys were found to have consistently higher energy intake than girls and more girls were found to be stunted, underweight and wasted (10). Contrary to the above studies, however, other studies have not found significant difference between girls and boys in terms of nutrition and health outcomes (12). Exclusive breast feeding was also one of the parameters where girls were found to benefit more than boys in our study, although, this was not statistically significant. The odds of being exclusively breast fed was 1.35 times more common for females with COR = 1.35(CI 0.86, 2.13) and AOR = 2.79(CI 0.70, 11.02). Similarly, for sub-Saharan countries, the male breast feeding advantage is much smaller. In contrary, Boys are breast fed for 0.657 months longer than girls in North African countries (11). In our study, statistically significant difference in care seeking for sick boys and girls during febrile and diarrheal diseases was not observed. However the odds of care seeking for sick boys and girls during diarrheal and febrile illness was AOR = 0.95(CI 0.26, 3.49) and AOR = 2.53(CI 0.61, 10.48) respectively. Our finding is similar to one meta-analysis which evaluated 57 countries and significant difference in care seeking for sick boys and girls were not observed in most countries (13). In contrary to this study, females were found to die more than boys at the age of 5, 10, 15 and 20 in Egypt and the major cause of excess female mortality was attributed to the favored treatment boys received for digestive and respiratory illness (8). Similarly, utilization of health care services showed marked male preferences in Bangladesh (9). A statistically significant difference in vaccination status was also not found between boys and girls in this study. The odds of being vaccinated among female was AOR = 1.27(CI 0.47, 3.39) when compared to boys. Similar to our study, in one systematic review, the pooled odds ratio for sex did not show significant differences between girls and boys in vaccination out-come (14). This is in contrary to one study in India which was conducted among 4000 children between the age of 1 and 2, the likeli-hood of female being fully vaccinated was 5% less than that for boys. In the same study, in certain subgroups of children, especially children from poorest households, boys were more likely to not being vaccinated than girls (7). Conclusion Ethiopia is considered a culturally conservative country rooted in traditional beliefs among them priority given to sons over daughters. Yet, there was no statistically significant association between gender of a child and under five child health based on our analysis. Females were found to have a better health and nutritional outcomes compared to boys in our study. This can be in part explained by the higher number of boys with young and uneducated mothers. Acknowledgment First of all, I would like to thank Emory University’s COALESCCE research scholars program for enrolling me and giving me chance to prepare this manuscript. My heartfelt appreciation also goes to my advisor Dr. Saria Hassan for her guidance and encouragement all through the process. I would also like to recognize Shivani A. Patel, and Solveig A. Cunningham who have been a big support. Funding The funding of this study was covered by the Fogarty international center of the national institute of health under award number D43TW011404. The content is solely the responsibility of the authors and doesn’t necessarily represent the official views of the national institutes of health. Availability of data The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. https://dhsprogram.com/data/dataset/Ethiopia_Standard-DHS_2016.cfm?flag=0 Abbreviations EDHS Ethiopian demographic health survey SDG Sustainable development goals WHO World health organization UNICEF United Nations children’s emergency fund Table 1 Frequency table of characteristics of the study sample by gender in Ethiopia, EDHS 2016 (n = 2075) Variables Gender p-value Male Female N (%) N (%) Maternal age 164(54.7%) 148(45.3%) 0.441 <20 717(49.5%) 683(50.5%) 20–34 134(45.6%) 147(54.4%) 35+ Wealth index 507(49.7%) 473(50.3%) 0.991 Poorest 196(50.6%) 196(49.4%) Poorer 149(49.3%) 147(49.3%) Middle 114(50.5%) 113(49.5%) Richer 93(47.9%) 87(52.1%) Richest Place of residence 87(57.0%) 78(43.0%) 0.241 Urban 972(49.5%) 938(50.5%) Rural Education level of mother 761(50.4%) 707(49.6%) 0.914 No education 238(48.5%) 253(51.5%) Primary 45(47.7%) 44(52.3%) Secondary 15(56.6%) 12(43.4%) Higher Age of the child in month 306(49.0%) 303(51.0%) 0.940 Under 6 months 256(50.5%) 224(49.5%) 6–11 months 497(50.0%) 489(50.0%) 12–23 months Number of under 5 in house hold 312(50.5%) 319(49.5%) 0.780 1 504(51.8%) 462(48.2%) 2 210(44.2%) 196(55.8%) 3 27(52.2%) 31 (47.8%) 4 3(3.7%) 5(96.3%) 5 3(50.1%) 3(49.9%) 6 Table 2 Characteristics of mother-infant pairs in Ethiopia, EDHS 2016 (n = 2075) Variable Gender p-value Male Female N (%) N (%) Stunting 245(50.8%) 211(49.2%) 0.571 Yes 711(48.7%) 724(51.3%) No Underweight 265(53.0%) 231(47.0%) 0.221 Yes 746(48.8%) 755(51.2%) No Wasting 171(56.2%) 138(43.8%) 0.101 Yes 843(48.9%) 851(51.1%) No Toilet facility 587(53.3%) 526(46.7%) 0.139 Open defecation 365(47.5%) 398(52.5%) Un improved facility 107(46.2%) 92(53.8%) Improved facility Water source 557(51.5%) 507(48.5%) 0.276 Improved 502(48.2%) 509(51.8%) Unimproved ANC visit 488(50.8%) 440(49.2%) 0.735 No ANC visits 319(50.0%) 304(50.0%) 1 to 3 visits 252(47.8%) 272(52.2%) More than 3 visits Place of delivery 771 (48.8%) 773(51.2%) 0.264 Home 288(53.0%) 243(47.0%) Health facility Delivery by caesarean section 18(60.1%) 12(39.9%) 0.465 Yes 1041(49.7%) 1004(50.3%) No Ever vacdnated 624(47.8%) 597(52.2%) 0.186 Yes 435(52.7%) 419(47.3%) No Diarrhea medical treatment 71(48.4%) 66(51.6%) 0.398 Yes 82(55.5%) 66(44.5%) No Fever/cough medical treatment 65(45.6%) 59(54.4%) 0.995 Yes 165(53.1%) 145(46.9%) No Stunting 245(50.8%) 211(49.2%) 0.571 Yes 711(48.7%) 724(51.3%) No Anemia 23(44.7%) 28(55.3%) 0.356 Severe 115(55.0%) 106(54.0%) Moderate 275(45.1%) 265(54.9%) Mild 646(51.1%) 617(48.9%) Not anemic Currently breast feeding 920(50.3%) 872(49.7%) 0.408 Yes 139(46.3%) 144(53.7%) No Given child anything other than breast milk 214(56.8%) 156(43.2%) 0.187 Yes 816(49.2%) 837(50.8%) No Table 3 Association between Gender and vaccination, care seeking in the past 2 weeks, stunting, underweight, wasting, anemia and exclusive breast feeding among children aged less than five years in Ethiopia, EDHS 2016. Stunting Wasting Underweight Gender Yes No COR(95% CI) AOR(95% CI) Yes No COR (95% CI) AOR(95% CI) Yes No COR (95% CI) AOR(95% CI) Male 245 711 1.00 1.00 171 843 1.00 1.00 265 746 1.00 1.00 Female 211 724 0.92(0.68,1.22) 0.48(0.16, 1.41) 138 851 0.74(0.52,1.06) 3.09(0.50, 18.89) 231 755 0.84(0.64, 1.1) 0.36(0.08, 1.54) Anemia Health seeking for diarrhea Health seeking for fever Gender Yes No COR(95% CI) AOR(95% CI) Yes No COR (95% CI) AOR(95% CI) Yes No COR (95% CI) AOR(95% CI) Male 413 646 1.00 1.00 71 82 1.00 1.00 65 165 1.00 1.00 Female 399 617 1.16(0.92, 1.47) 0.65(0.19,2.1) 66 66 1.32(0.67,2.59) 0.95(0.26, 3.49) 59 145 1.35(0.76, 2.39) 2.53(0.61, 10) Exclusive breast feeding Vaccination status Gender Yes No COR(95% CI) AOR(95% CI) Yes No COR (95% CI) AOR(95% CI) Male 816 214 1.00 1.00 624 435 1.00 1.00 Female 837 156 1.35(0.86, 2.13) 2.79(0.70, 11) 595 419 1.21(0.91, 1.62) 1.27(0.47, 3.3) Declarations Competing interest No competing interest. Ethical consideration The data was downloaded and used after permission was taken from measure DHS. The original DHS data was collected in confirmation with international and ethical guidelines. Consent for publication Not applicable ==== Refs References 1. Level and trend in child mortality.UNICEF 2019 report. 2. Central statistics agency of Ethiopia and ORC Macro,Ethiopia Demographic and Health survey 2016,Central statistics agency and ORC Macro,Calverton, UK,2016. 3. Sustainabledevelopment.un.org 4. Roland P Why is infant mortality higher in boys than in girls? A new hypothesis based on preconception environment and evidence from a large sample of twins. Demography. 2013;50 (2 ):421–44.23151996 5. Ethiopian demographic health survey.EDHS 2005,2011,2015,2019. 6. Jahidur RK ,Nabil AW .A comprehensive analysis on child mortality and its determinants in Bangladesh using frailty models.Archives of public health(2017)75 :58.28912949 7. Borooah VK .Gender bias amng children in India in their diet and immunization against disease.Soc Sci Med.2004;58 (9 ):1719–31.14990373 8. Ahmed FA .Gender difference in child mortality.Egypt popul Fam Plann Rev. 1990 Dec;24 (2 ):60–79. 9. Lincon C ,Emdadul H ,Stan D .SexbiasinfamilyallocationoffoodandhealthcareinruralBangladish.Populationanddevelopmentreview 1981(7 ). 10. Ndiku M ,Jaceldo S .Gender inequalities in food intake and nutritional status of children under five years old in rural eastern Kenya.European clinical journal of clinical nutrition.2011;(65 ):26–31. 11. UNICEF 2010 report 12. Janaina C ,Fernando C etal.Gender bias in care seeking practices in 57 low and middle in come countries.J Global Health 2017 Jun;7 (1 ). 13. WHO 2010.Gender and immunization summary report. 14. 2014 Ghana Demographic and health survey

PMC010xxxxxx/PMC10312977.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398019 10.21203/rs.3.rs-3005686/v1 10.21203/rs.3.rs-3005686 preprint 1 Article Male histone deacetylase 6 (HDAC6) knockout mice have enhanced ventilatory responses to hypoxic challenge Getsy Paulina M. Coffee Gregory A. Kelley Thomas J. Lewis Stephen J. Case Western Reserve University Author contributions P.M.G., G.A.C., T.J.K. and S.J.L. designed and conceived the experiments. P.MG., G.A.C., performed the studies. P.M.G., G.A.C., and S.J.L. collated and analyzed the data. P.M.G., G.A.C., T.J.K. and S.J.L. wrote the manuscript. All authors approved the final version of the manuscript and agree to account for all aspects of the work. All persons designated as authors qualify for authorship, and all of those qualify for authorship are listed. ✉ sjl78@case.edu 13 6 2023 rs.3.rs-3005686https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs3005686v1.pdf Histone deacetylase 6 (HDAC6) is a class II histone deacetylase that is predominantly localized in the cytoplasm of cells. HDAC6 associates with microtubules, regulating acetylation of tubulin and other proteins. The possibility that HDAC6 participates in hypoxic signaling is supported by evidence that (1) hypoxic gas challenges cause microtubule depolymerization, (2) expression of hypoxia inducible factor alpha (HIF)-1α is regulated by microtubule alterations in response to hypoxia, and (3) inhibition of HDAC6 prevents HIF-1α expression and protects tissue from hypoxic/ischemic insults. The aim of this study was to address whether the absence of HDAC6 alters ventilatory responses during and/or after hypoxic gas challenges (10% O2, 90% N2 for 15 min) in adult male wild-type (WT) C57BL/6 mice and HDAC6 knockout (KO) mice. Key findings were that (1) baseline values for frequency of breathing, tidal volume, inspiratory and expiratory times and end expiratory pause were different between KO mice and WT mice, (2) ventilatory responses during hypoxic challenge were more robust in KO mice than WT mice for parameters including frequency of breathing, minute ventilation, inspiratory and expiratory durations, peak inspiratory and expiratory flows, inspiratory and expiratory drives, and (3) responses upon return to room-air were markedly different in KO mice than WT mice for frequency of breathing, minute ventilation, inspiratory and expiratory durations, end expiratory (but not end inspiratory) pauses, peak inspiratory and expiratory flows, and inspiratory or expiratory drives. These data suggest that HDAC6 may have a fundamentally important role in regulating the neural responses to hypoxia. C57BL/6 mice breathing hypoxic gas challenge Histone deacetylase 6 (HDAC 6) knockout mice CF Mouse Model Resource Center at CWRUCFF HODGES19R1 NIH/NHLBIR01 HL156928-01A1 ==== Body pmcIntroduction Histone deacetylase 6 (HDAC6) is a class II histone deacetylase that exists predominantly within the cytosolic compartment of cells where it associates with microtubules to regulate the acetylation of tubulin and other cytosolic/intracellular protein targets.1–6 Numerous studies have demonstrated that pharmacological inhibition of HDAC6 improves neuronal function in multiple disease states.1–3,7–12 For example, the inhibition of HDAC6 improves microtubule-mediated transport in neurons in Huntington’s disease directly by increasing tubulin acetylation.13 The peripheral nerve disease, Charcot-Marie-Tooth, is characterized by reduced tubulin acetylation.14 HDAC6 inhibitors improve neuronal transmission and alleviate phenotypes in a mouse model of this disease.14,15 In addition, HDAC6 inhibitors have been assessed in vascular dementia (e.g., Alzheimer’s) diseases and Parkinson’s disease based on their ability to improve neuronal function via tubulin acetylation.16–19 Clear relationships between hypoxia and microtubule regulation have been demonstrated in cardiomyocyte preparations.20,21 For example, Dang et al20 demonstrated that hypoxic challenge leads to microtubule depolymerization. These findings are consistent with those from Teng et al21 who demonstrated that hypoxia inducible factor-1α (HIF-1α) expression is regulated by microtubule alterations in response to hypoxic challenge. Stable microtubules are preferentially acetylated, suggesting that HDAC6 inhibition should protect against damage induced by hypoxic challenge. Multiple studies have demonstrated that HDAC6 inhibition prevents HIF-1α expression and protects tissue from hypoxic/ischemia challenges.22–25 Taken together, it is evident that pharmacological inhibition of HDAC6 protects against hypoxic challenge-induced tissue damage and also improves central and peripheral neuronal function in a variety of disease states.1–25 Ventilatory responses to hypoxia are dependent on carotid body sensing and neuronal (chemoreceptor afferent) signal propagation to the commissural nucleus tractus solitarius in the brainstem.26–29 At present, there is no information as to (1) whether HDAC6 exists in primary glomus (hypoxia-sensing) cells within the carotid bodies or in key brain structures such as the commissural nucleus tractus solitarius that receive and process chemoreceptor afferent input. The evidence that HDAC6 inhibition prevents HIF-1α expression and protects tissue from hypoxic and/or ischemic damage30–33 is consistent with a role for HDAC6 in carotid body function since there is extensive evidence that HIF-1α plays many roles in hypoxic signaling in primary carotid body glomus cells.34–39 To our knowledge, there are no studies that have directly addressed whether or not HDAC6 has a role in signaling processes involved in expression of the ventilatory responses that occur upon exposure to hypoxic challenges. Accordingly, the primary objective of this study was to compare the ventilatory responses elicited by a hypoxic gas challenge (10% O2, 90% N2) in adult male wild-type (WT) C57BL/6 mice and HDAC6 knock-out (KO) mice using whole-body plethysmography.40–46 The data from these experiments clearly demonstrate that HDAC6 has a fundamentally important role in regulating the neural responses that drive the ventilatory responses to hypoxic challenge. Experimental Procedures Mice: Adult wild type C57BL/6 mice were obtained from Jackson Laboratories (Bar Harbor, Maine). Adult male and female HDAC6 KO mice were generously provided by Dr. Tso-Pang Yao (Duke University). Breeding pairs of these mice provided the adult male HDAC6 KO mice used in this study. All studies described were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No. 80 – 23) revised in 1996 and in strict compliance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines (https://arriveguidelines.org). The protocols were approved by the Animal Care and Use Committee of Case Western Reserve University. Whole-body plethysmography Ventilatory parameters in freely-moving mice were recorded by whole body plethysmography (PLY3223; Data Sciences International, St. Paul, MN) as described previously.40–46 The parameters (see Supplemental Table 1, Supplemental Fig. 1) were frequency of breathing (Freq); tidal volume (TV, volume of inspired air per breath), minute ventilation (Freq × TV, total volume of air inspired/min); inspiratory time (Ti, duration of inspiration); expiratory time (Te, duration of expiration); expiratory/inspiratory time (Te/Ti, expiratory quotient); end inspiratory pause (EIP, pause between end of inspiration and start of expiration); end expiratory pause (EEP, pause between end of expiration and start of inspiration); peak inspiratory flow (PIF); peak expiratory flow (PEF); airflow at 50% expired TV (EF50), relaxation time (RT, time to exhale 64% of TV), expiratory delay (Te-RT), inspiratory drive (TV/Ti) and expiratory drive (TV/Te), and non-eupneic breathing index (NEBI, % of breaths non-eupneic breaths including irregular events and apneas and type 1 and 2 sighs) and NEBI/Freq (NEBI corrected for Freq). The Fine Pointe (BUXCO) software constantly corrected digitized values for changes in chamber temperature and humidity. Pressure changes associated with the respiratory waveform were converted to volumes (e.g., TV, PIF, PEF) using the algorithm of Epstein and colleagues.47,48 Factoring in the chamber temperature and humidity, the cycle analyzers filtered the acquired signals, and algorithms (Fine Pointe, BUXCO) generated an array of box flow data that identified a waveform segment as an acceptable breath. From that data vector, the minimum and maximum values were determined. The flows at this point were “box flow” signals. From this array, the minimum and maximum box flow values were then determined and multiplied by a compensation factor provided by the selected algorithm,47,48 thus producing TV, PIF and PEF that are used to determine accepted and rejected waveforms. In all protocols described below, the conscious unrestrained mice were placed in the plethysmography chambers and allowed at least 60 min to acclimatize before exposure to the gas challenges. Protocols for gas challenges On the day of study, C57BL/6 mice and HDAC6 KO mice were placed in whole-body plethysmography chambers and allowed at least 60 min to acclimatize and settle so that baseline breathing values could be ascertained. The mice were then exposed to a hypoxic gas (10% O2, 90% N2) challenge for 5 min after which time they were re-exposed to room air. Statistics All data are shown as mean ± SEM. To determine total responses (cumulative %changes from pre-hypoxia values) during gas challenge and return to room air for each mouse, we summed the values recorded before and during the challenge and those upon return to room air. We then determined the cumulative response by the formula, total response (%change) = {[(sum of values during hypoxic challenge or return to room air) − (sum of values before hypoxic challenge)]/sum of values before hypoxic challenge} × 100. We then determined the mean and SEM of the group data. All data were analyzed by one-way or two-way ANOVA followed by Student’s modified t-test with Bonferroni corrections for multiple comparisons between means as described in detail previously.49 Results Baseline parameters The ages of the HDAC6 KO mice were slightly lower (−3.0%) than those of WT mice whereas the body weights of the HDAC6 KO mice were slightly higher (+ 13.2%) than the WT mice (see Table 1). As such the body weight/age ratio for the HDAC6 KO mice (0.34 ± 0.01) was higher than that of the WT mice (0.29 ± 0.1). The heavier body weights of the HDAC6 KO mice could certainly influence the findings related to flow parameters, namely, tidal volume, minute ventilation (tidal volume × frequency of breathing), peak inspiratory and expiratory flows, and EF50. Accordingly, (1) resting tidal volumes were lower in HDAC6 KO mice than in WT mice but were similar to one another when corrected for body weight, (2) corrections for body weights did not alter the lack of differences between the two groups with respect to minute ventilation, peak inspiratory and expiratory flows, or EF50. As also summarized in Table 1, resting frequency of breathing was lower in the HDAC6 KO mice than in the WT mice and inspiratory and expiratory times were longer in the HDAC6 KO mice. In addition, resting expiratory delay (Te-RT) was longer in HDAC6 KO mice than in WT mice. All other baseline parameters were similar between the two groups. It was also evident that the moment-to moment variability of many parameters was higher in the HDAC6 KO mice than in the WT mice. As shown in Table 2, values of standard deviation/corrected for mean (STDEV/mean) for frequency of breathing, minute ventilation, inspiratory time, expiratory time, end inspiratory pause and inspiratory drive were higher in the HDAC6 KO mice than in the WT mice. The finding that neither NEBI or NEBI/Freq were different between the groups suggests that the variability is due to simple changes in the breath-to-breath levels of frequency of breathing, for example, rather than enhanced expression of non-eupneic breathing including irregular breaths and apneas. Ventilatory responses to hypoxic gas challenges and upon return to room-air Frequency of breathing, tidal volume, minute ventilation The frequency of breathing, tidal volume and minute ventilation values recorded before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in WT mice and in HDAC6 KO mice are shown in the left-hand panels of Fig. 1. As seen in Panel A, resting frequency of breathing prior to the hypoxic gas challenge was similar in WT and HDAC6 KO. The hypoxic challenge in WT mice elicited a typical increase in frequency of breathing associated with expected roll-off. The increases in Freq were somewhat higher in HDAC6 KO mice. The return to room-air elicited a typical increase in frequency of breathing in WT mice that gradually declined over the first 5 min of the recording period. The corresponding frequency of breathing values were higher in HDAC6 KO mice especially over the 5–15 min time-period when values had returned to baseline values in WT mice. The total responses summarized in Panel B show that the total frequency of breathing responses elicited by the hypoxic challenge (HX) and upon return to room-air (RA5 and RA15) were higher in HDAC6 KO mice than in WT mice. As seen in Panel C, resting tidal volume prior to the hypoxic challenge was consistently higher in the HDAC6 KO mice than WT mice, perhaps due to the slightly larger body weights of the KO mice (see Table 1). The hypoxic challenge in the WT mice elicited a sustained increase in tidal volume that did not display roll-off. The increases in tidal volume were similarly robust in HDAC6 KO mice and reached higher values, mostly because of the higher resting values. Upon return to room-air, tidal volume values returned to pre-hypoxic levels within 5 min in WT mice but remained elevated in the HDAC6 KO mice over the 15 min recording period. The total responses summarized in Panel D shows that the total tidal volume responses elicited by hypoxia (HX) were similar in HDAC6 KO mice than in WT mice. The total responses upon return to room-air over the first 5 min (RA5) were similar in the HDAC6 KO and WT mice. The total tidal volume response over the entire 15 min recording period was significant in the HDAC6 KO mice but not the WT mice. As seen in Panel E, resting minute ventilation prior to the hypoxic challenge was similar in WT and HDAC6 KO mice. The hypoxic challenge in the WT mice elicited a typical increase in minute ventilation that was associated with the expected roll-off. The increases in minute ventilation were consistently higher in HDAC6 KO mice. The return to room-air elicited a typical increase in minute ventilation in WT mice that subsided within 5 min. The corresponding MV values were higher in HDAC6 KO mice and remained elevated over the 15 min recording period. As seen in Panel F, the total minute ventilation responses elicited by hypoxia (HX) and return to room-air (RA 5 and RA15) were higher in the HDAC6 KO mice than in WT mice. Inspiratory time, expiratory time, expiratory time/inspiratory time The Ti, Te and Te/Ti values recorded before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in WT mice and in HDAC6 KO mice are shown in the left-hand panels of Fig. 2. As seen in Panel A, resting Ti prior to the hypoxic gas challenge tended to be higher in HDAC6 KO mice than WT mice during the 5 min period immediately before the hypoxic challenge. The hypoxic challenge in WT mice elicited a decrease in Ti that was associated with the expected roll-off. The decreases in Ti were somewhat greater in HDAC6 KO mice. The return to room-air elicited an initial further decrease in Ti in the WT mice that was followed by a gradual return toward pre-hypoxia values. Corresponding Ti values in HDAC6 KO mice showed the same pattern of changes as in WT mice but reached lower values most probably in part perhaps because of the lower values reached at the end of the hypoxic challenge. The total responses summarized in Panel B shows that the total Ti responses elicited by hypoxia (HX) and upon return to room-air (RA5 and RA15) were greater in HDAC6 KO mice than in WT mice. As seen in Panel C, resting expiratory time prior to the hypoxic challenge tended to be higher in HDAC6 KO mice than WT mice during the 5 min period before the hypoxic challenge. The hypoxic challenge in WT mice elicited a decrease in Te that was associated with the expected roll-off. The decreases in Te were somewhat greater in HDAC6 KO mice. The return to room-air elicited an initial further decrease in Te in the WT mice that was followed by a rapid return to pre-hypoxia values. The corresponding Te values in HDAC6 KO mice followed the same pattern of changes as in WT mice but stayed at lower values for longer before returning to pre-hypoxia values. The total responses summarized in the Panel D shows that the total Ti responses elicited by hypoxia (HX) and upon return to room-air (RA5 and RA15) were greater in HDAC6 KO mice than in WT mice. As seen in Panel E, prior to the hypoxic challenge, resting Te/Ti was similar in HDAC6 KO and WT mice. The hypoxic challenge elicited minimal increases in Te/Ti in both groups. Upon return to room-air, Te/Ti rose further (with a spike evident in WT mice) before gradually declining to pre-hypoxia levels. As seen in Panel F, the the total Te/Ti responses that occurred during the hypoxic challenge (HX) and upon return to room-air (RA5 and RA15) were similar in HDAC6 KO and WT mice. End Inspiratory Pause, End Expiratory Pause EIP and EEP values recorded before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in WT mice and in HDAC6 KO mice are shown in the left-hand panels of Fig. 3. As seen in Panel A, resting EIP and EEP values prior to the hypoxic gas challenge were similar between HDAC6 KO mice and WT mice. The hypoxic challenge elicited similar sustained decreases in EIP in WT and HDAC6 KO mice. Upon return to room-air, the EIP values gradually returned to pre-HX levels in both groups. As seen in Panel B, the total EIP responses elicited by hypoxia and upon return to room-air (RA5 and RA15) were similar in HDAC6 KO and WT mice. As seen in Panel C, resting EEP values tended to be higher in the HDAC6 KO mice. The hypoxic challenge elicited an initial decrease in EEP in both groups of mice that tended to recover toward baseline toward the end of the challenge. Upon return to room-air, EEP rose well above baseline levels in WT mice but stayed at baseline values in the HDAC6 KO mice. As can be seen in Panel D, the total decreases in EEP elicited by the hypoxic challenge (HX) were similar in WT and HDAC6 KO mice. In contrast, the substantial total increases in EEP observed upon return to room-air in the WT mice (RA5 and RA15) were not seen in the HDAC6 KO mice. Peak Inspiratory and Expiratory Flows The PIF and PEF values recorded before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in WT mice and in HDAC6 KO mice are shown in the left-hand panels of Fig. 4. As seen in Panels A and C, resting PIF and PEF values were similar between the HDAC6 KO mice and WT mice. The HX challenge elicited sustained increases in PIF and PEF in WT mice. These responses were markedly augmented in HDAC6 KO mice. Upon return to room-air, PIF and PEF spiked upward before gradually returning to baseline levels in WT and HDAC6 KO mice with PIF and PEF values remaining substantially higher in HDAC6 KO mice. As seen in Panels B and D, total PIF and PEF responses elicited by hypoxia (HX) and upon return to room-air (RA5 and RA15) were greater in the HDAC6 KO than in the WT mice. EF50, relaxation time, expiratory delay (Te-RT) EF50, relaxation time, expiratory delay (Te-RT) values recorded before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in WT mice and in HDAC6 KO mice are shown in the left-hand panels of Fig. 5. As seen in Panel A, resting EF50 values were similar between the HDAC6 KO mice and WT mice. The HX challenge elicited greater increases in EF50 in the HDAC6 KO mice than in the WT mice. These responses were markedly augmented in HDAC6 KO mice. The responses that occurred upon return to room-air were also markedly greater in the HDAC6 KO mice. As seen in Panel B, the total increases EF50 elicited by the hypoxic challenge and upon return to room-air (RA5 and RA15) were greater in the HDAC6 KO than in the WT mice. As seen in Panel C, resting relaxation time values were similar in HDAC6 KO mice and WT mice. The hypoxic challenge elicited minimal changes in relaxation time in the WT mice but substantial initial falls in the HDAC6 KO mice that recovered within 3 min of the hypoxic challenge. Relaxation time dropped substantially upon return to room-air in the WT and HDAC6 KO mice. Relaxation time then rose above baseline values in the WT mice but returned to baseline values in HDAC6 KO mice. As seen in Panel D, total decreases in relaxation elicited by hypoxia and upon return to room-air (RA5 and RA15) were greater in HDAC6 KO than in WT mice. As seen in Panel E, resting expiratory delay (Te-RT) values tended to be higher in HDAC6 KO mice than in the WT mice immediately before the hypoxic challenge. The hypoxic challenge elicited slightly greater decreases in expiratory delay in HDAC6 KO mice. Expiratory delay returned rapidly to baseline levels in the WT mice whereas these values remained below baseline values in HDAC6 KO mice for 5–6 min. As seen in Panel F, the total decreases in expiratory delay elicited during hypoxic challenge (HX) were greater in HDAC6 KO than in WT mice. Decreases in expiratory delay upon return to room air (RA5 and RA15) occurred in the HDAC6 KO mice only. Inspiratory and Expiratory Drives Inspiratory and expiratory drive values recorded before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in WT mice and in HDAC6 KO mice are shown in the left-hand panels of Fig. 6. As seen in Panels A and C, resting drives prior to the hypoxic challenge were similar in HDAC6 KO mice and WT mice. The hypoxic challenge elicited sustained increases in inspiratory drive and expiratory drive in WT mice. These responses were markedly augmented in HDAC6 KO mice. Upon return to room-air, the inspiratory drive and expiratory drive values spiked upward in WT mice before gradually returning to baseline levels. These values were considerable higher in the HDAC 6 KO mice. As seen in Panels B and D, the responses elicited during the hypoxic challenge (HX) and upon return to room-air (RA5 and RA15) were greater in HDAC6 KO mice than WT mice. NEBI, NEBI/Freq Non-eupneic breathing index (NEBI) and NEBI/Freq values recorded before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in WT mice and in HDAC6 KO mice are shown in the left-hand panels of Fig. 7. As seen in Panel A, resting NEBI values prior to the hypoxic challenge were similar in HDAC6 KO mice and WT mice. The hypoxic challenge elicited similar increases in NEBI in WT and HDAC6 KO mice. The return to room-air caused remarkable increases in NEBI in both groups with NEBI subsiding more rapidly in the WT mice. These values were considerable higher in the HDAC 6 KO mice. As seen in Panel B, the increases in NEBI elicited during the hypoxic challenge (HX) and during the first 5 min upon return to room-air (RA5) were similar in the WT and HDAC6 KO mice whereas the overall increase in NEBY (RA15) was much greater in the HDAC6 mice. and RA15) were greater in HDAC6 KO mice than WT mice. As seen in Panels C and D, normalizing the changes in NEBI for the changes in frequency of breathing (NEBI/Freq) resulted in the changes during hypoxia (HX) and upon return to room-air (RA5 and RA15) that were similar in both groups of mice. Body weight considerations The heavier body weights of HDAC 6 KO mice may influence findings related to the effects of hypoxic challenge and return to room-air on flow parameters; tidal volume, minute ventilation, peak inspiratory and expiratory flows, and EF50. Table 3 summarizes the total arithmetic changes in ventilatory parameters during hypoxic gas challenge with flow parameters also shown corrected for body weight. The total arithmetic changes for frequency of breathing, inspiratory and expiratory times, expiratory time/expiratory time, end inspiratory and expiratory pauses, relaxation time, expiratory time-relaxation time, NEBI and NEBI/Freq provide the exact same conclusions given by the %change data provided in Figs. 1–7 in that hypoxia-mediated changes in these parameters were greater in the HDAC6 KO mice. Second, the delta changes in flow variables corrected for body weights (delta/body weight) for tidal volume, minute ventilation, peak inspiratory and expiratory flows, EF50 and inspiratory and expiratory drives were also consistent with %change data provided in Figs. 1–7 in that except for tidal volume, the hypoxia-mediated changes in these parameters were greater in the HDAC6 KO mice than in the WT mice. Discussion The C57BL/6 mouse is a common inbred strain that is widely used in ventilatory and pulmonary function studies50,51 and to produce mice lacking genes for numerous functional proteins.41,52–54 The C57BL/6 mouse is used as a mouse model with “normal” physiology and indeed they display many “normal” traits.41–44,55–58 For example, resting systemic and pulmonary arterial blood pressures at rest and cardiovascular responses upon challenges with hypoxic, hypercapnic and hypoxic-hypercapnic gas mixtures are representative of other healthy mouse and rat strains.56–58 Moreover, the ability of hypercapnia to modulate the effects of hypoxia on arterial blood pressure (hypoxia elicits pronounced depressor response, hypercapnia elicits a minor pressor response, hypoxia-hypercapnia elicits a minimal response) is to be intact in C57BL/6 mice.56,57 As such, C57BL/6 mice have been used extensively to study the effects of hypoxic, hypercapnic and hypoxic-hypercapnic gas challenges on ventilatory function41–44 and disordered breathing during both wakefulness and sleep.59–67 Despite being slightly younger, the HDAC6 KO mice were slightly heavier than their WT (C57BL/6) littermate controls. Whether this means that deletion of HDAC6 affects body metabolism or other factors regulating general health/body weight in C56BL/6 mice are yet to be established. In addition, the loss of HDAC6 could directly/indirectly impact ventilatory parameters in C57BL/6 mice by numerous mechanisms. For example, HDAC6 exists in smooth muscle and vascular endothelium of pulmonary arteries and Inhibition of HDAC6 improves the functions of both cell types.68,69 Resting ventilatory parameters A key finding of this study was that baseline (pre-HX gas challenge) frequency of breathing values were lower in the HDAC6 KO mice than in the WT mice. The reduced frequency of breathing in HDAC6 KO mice was accompanied by longer inspiratory and expiratory times. These findings certainly suggest that the possible presence of HDAC6 in key brainstem sites controlling respiratory frequency such as the NTS has a vital role in setting resting inspiratory and expiratory times. Moreover, the increased baseline variability in breathing parameters in HDAC6 KO mice (in the absence of enhanced non-eupneic breathing) certainly suggests that the presence of HDAC6 is essential for normal patterning of breathing. Similarly, the findings that end expiratory pause and expiratory delay (Te-RT) were greater in the HDAC6 KO mice suggests that HDAC6 is important for regulation of expiratory dynamics. Finally, the findings that the majority of the resting ventilatory parameters were similar in HDAC6 KO and WT mice (e.g., tidal volume, peak inspiratory and expiratory flows) does not negate a role for HDAC6 in the control of these parameters but rather that C57BL/6 mice are able to compensate for the loss of this important signaling element. Ventilatory Responses to Hypoxic Gas Challenge The hypoxic gas challenge elicited substantially greater increases in frequency of breathing (but not tidal volume) and therefore minute ventilation in HDAC6 KO mice than in WT mice. These novel findings suggest that stabilizing microtubules has a very important positive effect on respiratory timing but perhaps not ventilatory mechanics. The carotid body and chemoafferents in the carotid sinus nerve play an essential role in detecting and transmitting hypoxic signals to the commissural nuclei tractus solitarii in the brainstem26–29 and we have provided evidence that hypoxic ventilatory responses are markedly reduced in freely-moving male C57BL/6 (WT) mice with bilateral carotid sinus nerve transection.43 Although currently lacking, evidence that HDAC6 exists within the carotid bodies and key brain structures such as the nucleus tractus solitarii (see below for further discussion) would supports our evidence that HDAC6 is vital to the robustness of hypoxic ventilatory signaling. The hypoxia-induced increases in frequency of breathing were, as expected, associated with temporally consistent decreases in inspiratory and expiratory times in WT and HDAC6 KO mice. The decreases in inspiratory and expiratory times and were greater than the HDAC6 KO mice than the WT mice, consistent with the more pronounced increases in Freq in HDAC6 KO mice. The decrease in inspiratory time was somewhat greater than the decrease in Te in the WT and HDAC6 KO mice such that there was a slight increase in expiratory quotient (Te/Ti ratio) in both groups of mice with the increase in this ratio being larger in the HDAC6 KO mice. The combinations of increased tidal volume coupled to decreases in inspiratory and expiratory times resulted in marked increases in inspiratory drive (TV/Ti) and expiratory drive (TV/Te) in both groups but which were substantially larger in the HDAC6 KO mice. Again, although data is not available as to the precise brain sites that HDAC6 may participate in hypoxic signaling, it is known that HDAC6 exists widely throughout the brain although it is particularly associated with serotonergic neurons such as the dorsal and median raphe nuclei70–72 that are known to have important roles in the control of ventilatory processes73–75 although it appears that whereas these neurons play a key role in the expression of the ventilatory responses to hypercapnic challenges, they do not play a major role in expression of ventilatory responses to hypoxic gas challenges.76–82 As would be expected, end inspiratory and expiratory pauses decreased during exposure to the hypoxic challenge in the WT and HDAC6 KO mice. The decreases in end inspiratory pause were identical in the WT and HDAC6 KO mice whereas the decreases in end inspiratory pause were substantially greater in the HDAC6 KO mice. In addition, although relaxation times and expiratory delay (expiratory time − relaxation time) shortened remarkably during the hypoxic challenge in both groups, the decreases were substantially greater in the HDAC6 KO mice. Moreover, the increases in PIF, PEF and EF50 during the hypoxic challenge were dramatically augmented in the HDAC6 KO mice compared to the WT mice. Again, although it is not known if HDAC6 exists in the diaphragm and/or chest wall, histone deacetylases do exist in skeletal muscle,83–85 and as such reduced expression and/or pharmacological blockade of HDAC6 may increase the force of contraction generated by ventilatory muscles thereby enhancing PIF, PEF and EF50 responses during hypoxic gas challenges. Finally, the hypoxic challenge caused a substantially greater increase in the non-eupneic index (NEBI; e.g., disordered breathing, apneas, type 1 and 2 sighs) of HDAC6 KO mice than in WT mice although when corrected for the values for frequency of breathing (also more greatly elevated in HDAC6 KO mice) NEBI/Freq was similar in both groups. Although, we have argued that NEBI may reach higher values with higher values of frequency of breathing,44–46 this may not always be the case and so it remains plausible that the lack of HDAC6 destabilizes breathing patterns during hypoxic challenges. Ventilatory Responses upon Return to Room-Air The return to room-air in mice having undergone hypoxic gas challenges often results in an abrupt dramatic increase in Freq, TV and therefore MV in mice40–46, 86 that can result in unstable breathing.86–89 The mechanisms responsible for post-hypoxia alterations in breathing have received considerable investigation and at present, evidence favors disturbances in central signaling87,90 including the pons91,92 rather than processes within the carotid bodies.93,94 The present study demonstrated that C57BL/6 WT mice displayed the expected abrupt increase frequency of breathing, tidal volume and minute ventilation upon return to room-air which returned to baseline within 5 min. The increases in frequency of breathing and minute ventilation (but not tidal volume) upon return to room-air were greater in the HDAC6 KO mice than WT mice over the first 5 min of return to room-air and took substantially longer to return to baseline values. These results clearly suggest that HDAC6 within peripheral and central neural structures normally plays a vital role in the ventilatory adaptations that occur upon the return to room-air. As would be expected, the decreases in inspiratory and expiratory times were greater in the HDAC6 KO mice than the WT mice over the first 5 min following return to room-air and were sustained for a longer period of time. A careful view of the data shows that expiratory time returned to pre-hypoxia (baseline) values relatively abruptly in the WT mice whereas it remained decreased for 5–10 min in HDAC6 KO mice. This evidence is strongly supported by the finding that the abrupt and sustained increases in end expiratory pause that occurred upon return to room-air in WT mice was virtually absent in the HDAC6 KO mice. This contrasts with the gradual return of end inspiratory pause to baseline levels upon return to room-air in the WT and HDAC6 KO mice. Taken together, this data provides evidence that HDAC6 has a major role in brain neural circuitry regulating expiratory timing. The increases in peak inspiratory and expiratory flows and EF50 upon return to room-air were greater over the first 5 min in HDAC6 KO mice than in WT mice and remained greater for longer periods. Again, enhanced activity of skeletal muscle in chest wall and diaphragm may be directly responsible for the enhanced responses in the HDAC6 KO mice although augmented central output to these muscles cannot be discounted. The findings that the decreases in relaxation time and expiratory delay (Te-RT) were remarkably greater in the HDAC6 KO mice also points to an important role for HDAC6 in expiratory control processes. Taking the changes in tidal volume and inspiratory and expiratory times into account, it was evident that the increases in inspiratory drive (TV/Ti) and expiratory drive (TV/Te) upon return to room-air were substantially greater in HDAC6 KO mice. Taken together, the data reinforce the overall impression that HDAC6 has a major role in regulating inspiratory and expiratory timing in C57BL/6 mice. The finding that the increase in NEBI upon return to room-air was greater in the HDAC6 KO mice tentatively suggests that HDAC6 plays a vital role in ventilatory stability during this phase and that the loss of HDAC6 may contribute to ventilatory instability (increased expression of abnormal breaths and apneas) during this phase and perhaps in general. The finding that post-hypoxia (post-apnea) breathing in humans is associated with severe glottal closures95,96 raises the possibility that decreased expression of HDAC6 may contribute to obstruction of the upper airway in patients with obstructive apneas and perhaps the expression of central apneas. Summary The genetic bases for different breathing patterns in mice at rest and in response to hypoxic and hypercapnic challenges in mouse strains have received extensive investigation96–105 as have neurochemical processes,66,67,88,89,106–108 and structural features of respiratory structures such as the carotid bodies.109–111 The possibility that HDAC6 is a key player in the genetic factors that regulate breathing opens up intriguing avenues of research and especially testing whether selective HDAC6 inhibitors such as CAY10603, Tubacin and Nexturastat112–116 augment and/or stabilize ventilatory responses to hypoxic and/or hypercapnic gas challenges in mouse models such as C57BL/6 mice.63–65, 88,89 The results of the present study in male mice raises the question of whether female HDAC6 KO mice will also display many of the ventilatory features displayed by males and especially the ventilatory responses to hypoxia and those that occur upon return to room-air. The question of how male and female HDAC6 KO mice respond to a hypercapnic gas challenge is also of great interest with respect to understanding the physiological role of HDAC6. Acknowledgement We gratefully acknowledge the Case Western Reserve University Cystic Fibrosis Mouse Models Core (www.cfmice.org), specifically Alma Wilson, Amanda Barabas, and Molly Schneider, for their work maintaining the mouse colony. We thank Dr. Tso-Pang Yao for generously sharing the HDAC6-null mouse line. Funding This work was supported by CF Mouse Model Resource Center at CWRU (CFF HODGES19R1) and NIH/NHLBI R01 HL156928-01A1 (Dr. Tom Kelly and Dr. Becky Darrah). Data availability The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request. Figure 1 Left-hand panels: Frequency of breathing (Freq), tidal volume (TV) and minute ventilation (MV) before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in wild-type (WT) mice (n = 7) and in HDAC6 knockout (HDAC6 KO) mice (n=14). Right-hand panels: Total responses recorded during the hypoxic (HX) gas challenge, during the first 5 min (RA5) or 15 min (RA15) upon return to room air. All data are presented as mean ± SEM. *P < 0.05, significant response. †P < 0.05, HDAC6 KO versus WT. Figure 2 Left-hand panels: Inspiratory time (Ti), expiratory time (Te) and Te/Ti before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in wild-type (WT) mice (n = 7) and in HDAC6 knockout (HDAC6 KO) mice (n=14). Right-hand panels: Total responses recorded during the hypoxic (HX) gas challenge, during the first 5 min (RA5) or 15 min (RA15) upon return to room air. All data are presented as mean ± SEM. *P < 0.05, significant response. †P < 0.05, HDAC6 KO versus WT. Figure 3 Left-hand panels: End inspiratory pause (EIP) and end expiratory pause (EEP) before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in wild-type (WT) mice (n = 7) and in HDAC6 knockout (HDAC6 KO) mice (n=14). Right-hand panels: Total responses recorded during the hypoxic (HX) gas challenge, during the first 5 min (RA5) or 15 min (RA15) upon return to room air. All data are presented as mean ± SEM. *P < 0.05, significant response. †P < 0.05, HDAC6 KO versus WT. Figure 4 Left-hand panels: Peak inspiratory flow (PIF), peak expiratory flow (PEF) and PEF/PIF before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in wild-type (WT) mice (n = 7) and in HDAC6 knockout (HDAC6 KO) mice (n=14). Right-hand panels: Total responses recorded during the hypoxic (HX) gas challenge, during the first 5 min (RA5) or 15 min (RA15) upon return to room air. All data are presented as mean ± SEM. *P < 0.05, significant response. †P < 0.05, HDAC6 KO versus WT. Figure 5 Left-hand panels: Expiratory flow at 50% tidal volume (EF50), relaxation time (RT) and expiratory delay (Te-RT) before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in wild-type (WT) mice (n = 7) and in HDAC6 knockout (HDAC6 KO) mice (n=14). Right-hand panels: Total responses recorded during the hypoxic (HX) gas challenge, during the first 5 min (RA5) or 15 min (RA15) upon return to room air. All data are presented as mean ± SEM. *P < 0.05, significant response. †P < 0.05, HDAC6 KO versus WT. Figure 6 Left-hand panels: Inspiratory drive (tidal volume/inspiratory time, TV/Ti) and expiratory time (tidal volume/expiratory time, TV/Te) before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in wild-type (WT) mice (n = 7) and in HDAC6 knockout (HDAC6 KO) mice (n=14). Right-hand panels: Total responses recorded during the hypoxic (HX) gas challenge, during the first 5 min (RA5) or 15 min (RA15) upon return to room air. All data are presented as mean ± SEM. *P < 0.05, significant response. †P < 0.05, HDAC6 KO versus WT. Figure 7 Left-hand panels: Non-eupneic breathing index (NEBI) and NEBI/frequency of breathing (NEBI/Freq) before (Pre-HX), during a 5 min hypoxic gas (HX; 10% O2, 90% N2) challenge and upon return to room-air (Post-HX) in wild-type (WT) mice (n = 7) and in HDAC6 knockout (HDAC6 KO) mice (n=14). Right-hand panels: Total responses recorded during the hypoxic (HX) gas challenge, during the first 5 min (RA5) or 15 min (RA15) upon return to room air. All data are presented as mean ± SEM. *P < 0.05, significant response. †P < 0.05, HDAC6 KO versus WT. Table 1 Baseline parameters in wild-type (WT) and HDAC6 knockout (HDAC6 KO) mice with the shown delta/body weight values represent the actual values multiplied by 1,000. Parameter Abbreviation WT mice HDAC6 KO mice Number of mice in each group 7 14 Age, (days) 94.7 ± 0.9 91.9 ± 0.6* Body Weight, (g) 27.3 ± 0.6 30.9 ± 0.9* Body Weight/Age, (g/days) 0.29 ± 0.1 0.34 ± 0.01* Frequency, (breaths/min) Freq 200 ± 6 170 ± 5* Inspiratory Time, (sec) Ti 0.111 ± 0.003 0.133 ± 0.005* Expiratory Time, (sec) Te 0.210 ± 0.008 0.253 ± 0.009* Expiratory/Inspiratory Time Te/Ti 2.01 ± 0.08 1.97 ± 0.10 End Inspiratory Pause, (msec) EIP 2.59 ± 0.09 2.68 ± 0.07 End Expiratory Pause, (msec) EEP 30.6 ± 6.1 61.0 ± 8.5* Tidal Volume, (ml) TV 0.166 ± 0.010 0.200 ± 0.008* *TV/body weight, (ml/g) × 1000 TV/BW 6.10 ± 0.35 6.51 ± 0.26 Minute Ventilation, (ml/min) MV 32.9 ± 3.0 34.9 ± 2.5 *MV/body weight, (ml/g) × 1000 MV/BW 1215 ± 133 1134 ± 79 Peak Inspiratory Flow, (ml/sec) PIF 2.73 ± 0.38 2.72 ± 0.26 *PIF/body weight, (ml/g) × 1000 PIF/BW 101.4 ± 16.4 88.5 ± 8.5 Peak Expiratory Flow (ml/sec) PEF 1.63 ± 0.15 1.77 ± 0.10 *PEF/body weight, (ml/g) × 1000 PEF/BW 60.3 ± 6.2 57.6 ± 3.3 PEF/PIF 0.62 ± 0.04 0.71 ± 0.03 *(PEF/PIF)/body weight (ratio/g) × 1000 22.7 ± 1.1 23. 2 ± 1.1 Air-flow at 50% expired TV (ml/sec) EF50 0.079 ± 0.007 0.077 ± 0.006 *EF50/body weight (ratio/g) × 1000 EF50/BW 2.93 ± 0.30 2.51 ± 0.21 Relaxation Time RT 0.104 ± 0.006 0.122 ± 0.005 Expiratory Delay Te-RT 0.106 ± 0.005 0.128 ± 0.005* Inspiratory Drive (ml/sec) TV/Ti (InspD) 1.62 ± 0.15 1.46 ± 0.10 *(TV/Ti)/body weight (ratio/g) × 1000 InspD/BW 59.7 ± 6.4 47.5 ± 2.9 Expiratory Drive (ml/sec) TV/Te (ExpD) 0.80 ± 0.05 0.75 ± 0.03 *(TV/Te)/body weight (ratio/g) × 1000 ExD/BW 29.5 ± 2.3 24.5 ± 1.1 Non-eupneic breathing Index (%) NEBI 10.3 ± 1.3 13.8 ± 1.8 NEBI/Freq (%/(breaths/min)) NEBI/Freq 0.057 ± 0.012 0.088 ± 0.013 The data are presented as mean ± SEM. * P < 0.05, histone deacetylase 6 knockout (HDAC6 KO) mice versus wild-type (WT). Table 2 Variability in resting variables in wild-type (WT) and HDAC6 knockout (HDAC6 KO) mice Parameter Parameter WT mice HDAC6 KO mice Frequency (breaths/min) STDEV 45.5 ± 7.4 74.7 ± 7.8* Mean 228 ± 20 225 ± 16 STDEV/Mean 0.195 ± 0.019 0.325 ± 0.023* Tidal Volume (ml) STDEV 0.031 ± 0.003 0.038 ± 0.003 Mean 0.170 ± 0.010 0.200 ± 0.008 STDEV/Mean 0.180 ± 0.014 0.188 ± 0.009 Minute Ventilation (ml/min) STDEV 12.5 ± 2.6 22.8 ± 3.0* Mean 39.4 ± 5.7 47.1 ± 5.1 STDEV/Mean 0.304 ± 0.028 0.474 ± 0.036* Inspiratory Time (sec) STDEV 0.017 ± 0.002 0.032 ± 0.003* Mean 0.100 ± 0.009 0.113 ± 0.007 STDEV/Mean 0.179 ± 0.030 0.297 ± 0.032* Expiratory Time (sec) STDEV 0.037 ± 0.004 0.059 ± 0.003* Mean 0.200 ± 0.010 0.217 ± 0.010 STDEV/Mean 0.184 ± 0.020 0.277 ± 0.019* Expiratory Time/Inspiratory Time STDEV 0.457 ± 0.044 0.580 ± 0.028* Mean 2.122 ± 0.135 2.076 ± 0.082 STDEV/Mean 0.218 ± 0.024 0.286 ± 0.019 End Inspiratory Pause (msec) STDEV 0.440 ± 0.16 1.56 ± 0.68* Mean 2.58 ± 0.10 2.84 ± 0.14 STDEV/Mean 0.165 ± 0.054 0.460 ± 0.149* End Expiratory Pause (msec) STDEV 29.7 ± 2.3 44.4 ± 4.4* Mean 31.0 ± 6.0 47.9 ± 4.9* STDEV/Mean 1.156 ± 0.209 0.954 ± 0.058 Peak Inspiratory Flow (PIF, ml/sec) STDEV 1.215 ± 0.227 1.895 ± 0.168 Mean 3.36 ± 0.59 3.91 ± 0.46 STDEV/Mean 0.363 ± 0.035 0.508 ± 0.032* Peak Expiratory Flow (PEF, ml/sec) STDEV 0.575 ± 0.099 1.030 ± 0.130* Mean 1.85 ± 0.22 2.35 ± 0.19* STDEV/Mean 0.301 ± 0.025 0.426 ± 0.032* PEF/PIF STDEV 0.109 ± 0.016 0.133 ± 0.008 Mean 0.529 ± 0.027 0.670 ± 0.032* STDEV/Mean 0.180 ± 0.020 0.200 ± 0.010 EF50(ml/sec) STDEV 0.027 ± 006 0.054 ± 0.008* Mean 0.089 ± 0.010 0.106 ± 0.010 STDEV/Mean 0.298 ± 0.0.051 0.489 ± 0.047* Relaxation Time (sec) STDEV 0.019 ± 0.002 0.027 ± 0.001 Mean 0.100 ± 0.004 0.106 ± 0.005 STDEV/Mean 0.191 ± 0.024 0.259 ± 0.016 Expiratory Time − Relaxation Time STDEV 0.024 ± 0.002 0.037 ± 0.002* Mean 0.100 ± 0.007 0.111 ± 0.005 STDEV/Mean 0.246 ± 0.019 0.342 ± 0.021* Inspiratory Drive (ml/sec) STDEV 0.728 ± 0.145 1.140 ± 0.118* Mean 1.98 ± 0.34 2.28 ± 0.27 STDEV/Mean 0.363 ± 0.036 0.517 ± 0.036* Expiratory Drive (ml/sec) STDEV 0.251 ± 0.051 0.486 ± 0.076* Mean 0.90 ± 0.01 1.08 ± 0.10 STDEV/Mean 0.264 ± 0.028 0.428 ± 0.041 NEBI (%) STDEV 18.9 ± 2.5 20.9 ± 1.0 Mean 24.0 ± 8.0 30.3 ± 5.5 STDEV/Mean 1.099 ± 0.160 1.031 ± 0.157 NEBI/Freq (%/(breaths/min)) STDEV 0.068 ± 0.004 0.088 ± 0.007 Mean 0.089 ± 0.022 0.122 ± 0.020 STDEV/Mean 0.999 ± 0.163 0.953 ± 0.131 The data are presented as mean ± SEM. * P < 0.05, histone deacetylase 6 knockout (HDAC6 KO) mice versus wild-type (WT). Table 3 Total arithmetic responses during hypoxic gas challenge Parameter Parameter WT mice HDAC6 KO mice Flow-independent parameters Frequency (breaths/min) Delta response +1570 ± 157* + 3076 ± 212*,† Inspiratory Time (sec) Delta response −0.63 ± 0.07* −1.33 ± 0.09*,† Expiratory Time (sec) Delta response −0.85 ± 0.16* −2.12 ± 0.19*,† Expiratory Time/Inspiratory Time Delta response + 3.72 ± 2.27 + 5.66 ± 1.82* End Inspiratory Pause (msec) Delta response −8.84 ± 1.44* −10.28 ± 1.41* End Expiratory Pause (msec) Delta response −227 ± 54* −781 ± 153*,† Relaxation Time (sec) Delta response −0.22 ± 0.13 −0.82 ± 0.11*,† Expiratory Time − Relaxation Time Delta response −0.63 ± 0.07* −1.26 ± 0.08*,† NEBI (%) Delta response + 292 ± 55* + 420 ± 40*,† NEBI/Freq (%/(breaths/min)) Delta response + 0.46 ± 0.17* + 0.40 ± 0.24* Flow-dependentparameters Tidal Volume (ml) Delta response + 0.91 ± 0.14* + 1.34 ± 0.15* Delta/Body Weight + 0.033 ± 0.005* + 0.045 ± 0.006 Minute Ventilation (ml/min) Delta response + 531 ± 74* + 1024 ± 54*,† Delta/Body Weight + 19.3 ± 2.5* + 33.6 ± 2.2*,† PIF (ml/sec) Delta response + 46.1 ± 6.6* + 88.5 ± 4.3*,† Delta/Body Weight + 1.67 ± 0.22* + 2.90 ± 0.17*,† PEF (ml/sec) Delta response + 23.6 ± 3.4* + 47.7 ± 2.7*,† Delta/Body Weight + 0.863 ± 0.12* + 1.55 ± 0.09*,† PEF/PIF Delta response − 0.93 ± 0.79 − 2.39 ± 0.51 Delta/Body Weight − 0.032 ± 0.029 − 0.081 ± 0.016 Flow-independent parameters EF50 (ml/sec) Delta response + 0.98 ± 0.18* + 2.34 ± 0.16*,† Delta/Body Weight 0.035 ± 0.006* + 0 076 ± 0 006*,† Inspiratory Drive (ml/sec Delta response + 26.9 ± 2.4* + 57.5 ± 3.0*,† Delta/Body Weight + 0.99 ± 0.08* + 1.88 ± 0.11 *,† Expiratory Drive (ml/sec) Delta response + 11.1 ± 1.6* 24.3 ± 17*,† Delta/Body Weight + 0.40 ± 0.05* + 0.80 ± 0.06*,† WT, wild-type; HDAC6 KO histone deacetylase 6 knockout mice; NEBI, non-eupneic breathing index; Freq, frequency of breathing; PIF, peak inspiratory flow; PEF, peak expiratory flow; EF50, airflow at 50% expired tidal volume. The data are presented as mean ± SEM. * P < 0.05, histone deacetylase 6 knockout (HDAC6 KO) mice versus wild-type (WT). Declarations Competing interests None Supplementary Files This is a list of supplementary files associated with this preprint. Click to download. 2.HDAC6MS1SJLSciRep3supplementalfile.docx ==== Refs References 1. Li G , Jiang H , Chang M , Xie H , & Hu L . HDAC6 α-tubulin deacetylase: a potential therapeutic target in neurodegenerative diseases. J Neurol Sci. 304 , 1–8 (2011).21377170 2. Seidel C , Schnekenburger M , Dicato M , & Diederich M . Histone deacetylase 6 in health and disease. Epigenomics 7 , 103–118 (2015).25687470 3. Rodrigues DA , Thota S , & Fraga CA . Beyond the Selective Inhibition of Histone Deacetylase 6. Mini Rev Med Chem. 16 , 1175–1184 (2016).27121714 4. Balmik AA , & Chinnathambi S . Inter-relationship of Histone Deacetylase-6 with Tau-cytoskeletal organization and remodeling. Eur J Cell Biol. 101 , 151202 (2022).35092942 5. Bonanni D , Citarella A , Moi D , Pinzi L , Bergamini E , & Rastelli G . Dual Targeting Strategies on Histone Deacetylase 6 (HDAC6) and Heat Shock Protein 90 (Hsp90). Curr Med Chem. 29 , 1474–1502 (2022).34477503 6. Xu Y , & Wan W . Acetylation in the regulation of autophagy. Autophagy 19 , 379–387 (2023).35435793 7. Liang T , Fang H . Structure, Functions and Selective Inhibitors of HDAC6. Curr Top Med Chem. 18 , 2429–2447 (2018).30499393 8. Ma J , Kavelaars A , Dougherty PM , & Heijnen CJ . Beyond symptomatic relief for chemotherapy-induced peripheral neuropathy: Targeting the source. Cancer 124 , 2289–2298 (2018).29461625 9. Prior R , Van Helleputte L , Klingl YE , & Van Den Bosch L . HDAC6 as a potential therapeutic target for peripheral nerve disorders. Expert Opin Ther Targets 22 , 993–1007 (2018).30360671 10. LoPresti P. HDAC6 in Diseases of Cognition and of Neurons. Cells 10 , 12 (2020).33374719 11. Kumar V , Kundu S , Singh A , & Singh S . Understanding the Role of Histone Deacetylase and their Inhibitors in Neurodegenerative Disorders: Current Targets and Future Perspective. Curr Neuropharmacol. 20 , 158–178 (2022).34151764 12. Qureshi T , & Chinnathambi S . Histone deacetylase-6 modulates Tau function in Alzheimer’s disease. Biochim Biophys Acta Mol Cell Res. 1869 , 119275 (2022).35452751 13. Dompierre JP , Godin JD , Charrin BC , Cordelières FP , King SJ , Humbert S , & Saudou F . Histone deacetylase 6 inhibition compensates for the transport deficit in Huntington’s disease by increasing tubulin acetylation. J Neurosci. 27 , 3571–3583 (2007).17392473 14. d’Ydewalle C , Krishnan J , Chiheb DM , Van Damme P , Irobi J , Kozikowski AP , Vanden Berghe P , Timmerman V , Robberecht W , & Van Den Bosch L . HDAC6 inhibitors reverse axonal loss in a mouse model of mutant HSPB1-induced Charcot-Marie-Tooth disease. Nat Med. 17 , 968–974 (2011).21785432 15. Benoy V , Van Helleputte L , Prior R , d’Ydewalle C , Haeck W , Geens N , Scheveneels W , Schevenels B , Cader MZ , Talbot K , Kozikowski AP , Vanden Berghe P , Van Damme P , Robberecht W , & Van Den Bosch L (2018) HDAC6 is a therapeutic target in mutant GARS-induced Charcot-Marie-Tooth disease. Brain 141 , 673–687 (2018).29415205 16. Godena VK , Brookes-Hocking N , Moller A , Shaw G , Oswald M , Sancho RM , Miller CC , Whitworth AJ , & De Vos KJ . Increasing microtubule acetylation rescues axonal transport and locomotor deficits caused by LRRK2 Roc-COR domain mutations. Nat Commun. 5 , 5245 (2014).25316291 17. Majid T , Griffin D , Criss Z 2nd , Jarpe M , & Pautler RG Pharmocologic treatment with histone deacetylase 6 inhibitor (ACY-738) recovers Alzheimer’s disease phenotype in amyloid precursor protein/presenilin 1 (APP/PS1) mice. Alzheimers Dement. (NY) 1 , 170–181 (2015). 18. Cuadrado-Tejedor M , Garcia-Barroso C , Sánchez-Arias JA , Rabal O , Pérez-González M , Mederos S , Ugarte A , Franco R , Segura V , Perea G , Oyarzabal J , & Garcia-Osta A . A First-in-Class Small-Molecule that Acts as a Dual Inhibitor of HDAC and PDE5 and that Rescues Hippocampal Synaptic Impairment in Alzheimer’s Disease Mice. Neuropsychopharmacology 42 , 524–539 (2017).27550730 19. Esteves AR , Palma AM , Gomes R , Santos D , Silva DF , & Cardoso SM . Acetylation as a major determinant to microtubule-dependent autophagy: Relevance to Alzheimer’s and Parkinson disease pathology. Biochim Biophys Acta Mol Basis Dis. pii: S0925-4439(18)30475-7 (2018). 20. Dang Y , Lan X , Zhang Q , Li L , & Huang Y . Analysis of grayscale characteristics in images of labeled microtubules from cultured cardiac myocytes. Microsc Microanal. 21 , 334–342 (2015).25772206 21. Teng M , Dang YM , Zhang JP , Zhang Q , Fang YD , Ren J , & Huang YS . Microtubular stability affects cardiomyocyte glycolysis by HIF-1alpha expression and endonuclear aggregation during early stages of hypoxia. Am J Physiol Heart Circ Physiol. 298 , H1919–H1931 (2010).20228255 22. Kong X , Lin Z , Liang D , Fath D , Sang N , & Caro J . Histone deacetylase inhibitors induce VHL and ubiquitin-independent proteasomal degradation of hypoxia-inducible factor 1alpha. Mol Cell Biol. 26 , 2019–2028 (2006).16507982 23. Qian DZ , Kachhap SK , Collis SJ , Verheul HM , Carducci MA , Atadja P , & Pili R . Class II histone deacetylases are associated with VHL-independent regulation of hypoxia-inducible factor 1 alpha. Cancer Res. 66 , 8814–8821 (2006).16951198 24. Su M , Guan H , Zhang F , Gao Y , Teng X , & Yang W . HDAC6 Regulates the Chaperone-Mediated Autophagy to Prevent Oxidative Damage in Injured Neurons after Experimental Spinal Cord Injury. Oxid Med Cell Longev. 2016 , 7263736 (2016).26649145 25. Leng Y , Wu Y , Lei S , Zhou B , Qiu Z , Wang K , Xia Z (2018) Inhibition of HDAC6 Activity Alleviates Myocardial Ischemia/Reperfusion Injury in Diabetic Rats: Potential Role of Peroxiredoxin 1 Acetylation and Redox Regulation. Oxid Med Cell Longev 2018 : 9494052.30046381 26. Lahiri S , Roy A , Baby SM , Hoshi T , Semenza GL , & Prabhakar NR . Oxygen sensing in the body. Prog Biophys Mol Biol. 91 , 249–286 (2006).16137743 27. Prabhakar NR , & Peers C . Gasotransmitter regulation of ion channels: a key step in O2 sensing by the carotid body. Physiology (Bethesda) 29 , 49–57 (2014).24382871 28. López-Barneo J , González-Rodríguez P , Gao L , Fernández-Agüera MC , Pardal R , Ortega-Sáenz P (2016) Oxygen sensing by the carotid body: mechanisms and role in adaptation to hypoxia. Am J Physiol Cell Physiol. 310 , C629–C642 (2016).26764048 29. Baby SM , Gruber RB , Young AP , MacFarlane PM , Teppema LJ , & Lewis SJ . Bilateral carotid sinus nerve transection exacerbates morphine-induced respiratory depression. Eur J Pharmacol. 834 , 17–29 (2018).30012498 30. Qian DZ , Kachhap SK , Collis SJ , Verheul HM , Carducci MA , Atadja P , & Pili R . Class II histone deacetylases are associated with VHL-independent regulation of hypoxia-inducible factor 1 alpha. Cancer Res. 66 , 8814–8821 (2006).16951198 31. Su M , Guan H , Zhang F , Gao Y , Teng X , Yang W . HDAC6 Regulates the Chaperone-Mediated Autophagy to Prevent Oxidative Damage in Injured Neurons after Experimental Spinal Cord Injury. Oxid Med Cell Longev. 2016 , 7263736 (2016).26649145 32. Tsai HD , Wu JS , Kao MH , Chen JJ , Sun GY , Ong WY , & Lin TN . Clinacanthus nutans Protects Cortical Neurons Against Hypoxia-Induced Toxicity by Downregulating HDAC1/6. Neuromolecular Med. 18 , 274–282 (2016).27165113 33. Yi EH , Xu F , Li P , & Guo JQ . Transactive response DNA binding protein of 43/histone deacetylase 6 axis alleviates H 2 O 2 -induced retinal ganglion cells injury through inhibiting apoptosis and autophagy. J Cell Biochem. 120 , 4312–4320 (2019).30320895 34. Baby SM , Roy A , Mokashi AM , & Lahiri S Effects of hypoxia and intracellular iron chelation on hypoxia-inducible factor-1alpha and – 1beta in the rat carotid body and glomus cells. Histochem Cell Biol. 120 , 343–352 (2003).14600837 35. Baby SM , Roy A , & Lahiri S . Role of mitochondria in the regulation of hypoxia-inducible factor-1alpha in the rat carotid body glomus cells. Histochem Cell Biol. 124 , 69–76 (2005).16034640 36. Roy A , Li J , Baby SM , Mokashi A , Buerk DG , Lahiri S . Effects of iron-chelators on ion-channels and HIF-1alpha in the carotid body. Respir Physiol Neurobiol. 141 , 115–123 (2004).15239962 37. Roy A , Volgin DV , Baby SM , Mokashi A , Kubin L , & Lahiri S . Activation of HIF-1alpha mRNA by hypoxia and iron chelator in isolated rat carotid body. Neurosci Lett. 363 , 229–232 (2004).15182949 38. Roy A , Baby SM , Wilson DF , & Lahiri S . Rat carotid body chemosensory discharge and glomus cell HIF-1 alpha expression in vitro: regulation by a common oxygen sensor. Am J Physiol Regul Integr Comp Physiol. 293 , R829–R836 (2007).17475674 39. Lahiri S , Roy A , Baby SM , Di Giulio C , & Wilson DF . Carotid body sensory discharge and glomus cell HIF-1 alpha are regulated by a common oxygen sensor. Adv Exp Med Biol. 645 , 87–94 (2009).19227455 40. Palmer LA , May WJ , deRonde K , Brown-Steinke K , Lewis SJ . Hypoxia-induced ventilatory responses in conscious mice: Gender differences in ventilatory roll-off and facilitation. Respir Physiol Neurobiol. 185 , 497–505 (2013).23183420 41. Palmer LA , May WJ , deRonde K , Brown-Steinke K , Gaston B , Bates JN , & Lewis SJ . Ventilatory responses during and following exposure to a hypoxic challenge in conscious mice deficient or null in S-nitrosoglutathione reductase. Respir Physiol Neurobiol. 185 , 571–581 (2013).23183419 42. Palmer LA , deRonde K , Brown-Steinke K , Gunter S , Jyothikumar V , Forbes M , & Lewis SJ . (2015) Hypoxia-induced changes in protein S-nitrosylation in female mouse brainstem. Am J Respir Cell Mol Biol. 52 , 37–45.24922346 43. Gaston B , May WJ , Sullivan S , Yemen S , Marozkina NV , Palmer LA , Bates JN , & Lewis SJ . Essential role of hemoglobin beta-93-cysteine in post-hypoxia facilitation of breathing in conscious mice. J Appl Physiol. 116 , 1290–1299 (2014).24610531 44. Getsy PM , Davis J , Coffee GA , May WJ , Palmer LA , Strohl KP , & Lewis SJ . Enhanced non-eupneic breathing following hypoxic, hypercapnic or hypoxic-hypercapnic gas challenges in conscious mice. Respir Physiol Neurobiol. 204 , 147–159 (2014).25242462 45. Getsy PM , Coffee A , & Lewis SJ . Loss of ganglioglomerular nerve input to the carotid body impacts the hypoxic ventilatory response in freely-moving rats. Front Physiol. 14 , 1007043 (2023)37008015 46. Getsy PM , Davis J , Coffee A , Lewis THJ , & Lewis SJ . Hypercapnic signaling influences hypoxic signaling in the control of breathing in C57BL6 mice. J Appl Physiol. doi: 10.1152/japplphysiol.00548.2022 (2023). 47. Epstein MA , & Epstein RA . A theoretical analysis of the barometric method for measurement of tidal volume. Respir Physiol. 32 , 105–210 (1978).625610 48. Epstein RA , Epstein MA , Haddad GG , & Mellins RB Practical implementation of the barometric method for measurement of tidal volume. J Appl Physiol. 49 , 1107–1115 (1980).7440298 49. Getsy PM , Baby SM , May WJ , Lewis THJ , Bates JN , Hsieh YH , Gaston B , & Lewis SJ . L-NAC reverses of the adverse effects of fentanyl infusion on ventilation and blood-gas chemistry. Biomed Pharmacother. 153 , 113277 (2022).35724513 50. Soliz J , Soulage C , Borter E , van Patot MT , & Gassmann M . Ventilatory responses to acute and chronic hypoxia are altered in female but not male Paskin-deficient mice. Am J Physiol. 295 , R649–R658 (2008). 51. Soliz J , Thomsen JJ , Soulage C , Lundby C , & Gassmann M . Sex-dependent regulation of hypoxic ventilation in mice and humans is mediated by erythropoietin. Am J Physiol. 296 , R1837–R1846 (2009). 52. Kline DD , Overholt JL , & Prabhakar NR . Mutant mice deficient in NOS-1 exhibit attenuated long-term facilitation and short-term potentiation in breathing. J Physiol. 539 , 309–315 (2002).11850522 53. Liu L , Yan Y , Zeng M , Ahang J , Hanes MA , Ahearn G , McMahon TJ , Dickfeld T , Marshall HE , Que LG , & Stamler JS . Essential roles of S-nitrosothiols in vascular homeostasis and endotoxic shock. Cell 116 , 617–628 (2004).14980227 54. Duling LC , Cherng TW , Griego JR , Perrine MF , & Kanagy NL . (2006) Loss of α2B-adrenoceptors increases magnitude of hypertension following nitric oxide synthase inhibition. Am J Physiol. 291 , H2403–H2408 (2006). 55. Tankersley CG , Irizarry R , Flanders S , & Rabold R . Circadian rhythm variation in activity, body temperature, and heart rate between C3H/HeJ and C57BL/6J inbred strains. J Appl Physiol. 92 , 870–877 (2002).11796704 56. Campen MJ , Tagaito Y , Li J , Balbir A , Tankersley CG , Smith P , Schwartz A , & O’Donnell CP . Phenotypic variation in cardiovascular responses to acute hypoxic and hypercapnic exposure in mice. Physiological Genomics 20 , 15–20 (2004).15494473 57. Campen MJ , Tagaito Y , Jenkins TP , Balbir A , & O’Donnell CP . Heart rate variability responses to hypoxic and hypercapnic exposures in different mouse strains. J Appl Physiol. 99 , 807–813 (2005).15890760 58. Tewari SG , Bugenhagen SM , Wang Z , Schreier DA , Carlson BE , Chesler NC , & Beard DA . Analysis of cardiovascular dynamics in pulmonary hypertensive C57BL6/J mice. Front Physiol. 4 , 355 (2013).24376421 59. Tagaito Y , Polotsky VY , Campen MJ , Wilson JA , Balbir A , Smith PL , Schwartz AR , O’Donnell CP . A model of sleep-disordered breathing in the C57BL/6J mouse. J Appl Physiol. 91 , 2758–2766 (2001).11717244 60. Han F , Subramanian S , Dick TE , Dreshaj IA , & Strohl KP . Ventilatory behavior after hypoxia in C57BL/6J and A/J mice. J Appl Physiol. 91 , 1962–1970 (2001).11641331 61. Han F , Subramanian S , Price ER , Nadeau J , & Strohl KP . Periodic breathing in the mouse. J Appl Physiol. 92 , 1133–1140 (2002).11842050 62. Schneider H , Patil SP , Canisius S , Gladmon EA , Schwartz AR , O’Donnell CP , Smith, PL, & Tankersley CG. Hypercapnic duty cycle is an intermediate physiological phenotype linked to mouse chromosome 5. J Appl Physiol. 95 , 11–19 (2003).12794091 63. Yamauchi M , Dostal J , & Strohl KP . Acetazolamide protects against posthypoxic unstable breathing in the C57BL/6J mouse. J Appl Physiol. 103 , 1263–1268 (2007).17673555 64. Yamauchi M , Dostal J , Kimura H , & Strohl KP . Effects of buspirone on posthypoxic ventilatory behavior in the C57BL/6J and A/J mouse strains. J Appl Physiol. 105 , 518–526 (2008).18511527 65. Yamauchi M , Kimura H , & Strohl KP . Mouse models of apnea: strain differences in apnea expression and its pharmacologic and genetic modification. Adv Exper Med Biol. 669 , 303–307 (2010).20217371 66. Moore MW , Akladious A , Hu Y , Azzam S , Feng P , & Strohl KP Effects of orexin 2 receptor activation on apnea in the C57BL/6J mouse. Respir Physiol Neurobiol. 200 , 118–125 (2014).24929062 67. Moore MW , Chai S , Gillombardo CB , Carlo A , Donovan LM , Netzer N , & Strohl KP . Two weeks of buspirone protects against posthypoxic ventilatory pauses in the C57BL/6J mouse strain. Respir Physiol and Neurobiol. 183 , 35–40 (2012).22595368 68. Joshi AD , Barabutis N , Birmpas C , Dimitropoulou C , Thangjam G , Cherian-Shaw M , Dennison J , & Catravas JD . Histone deacetylase inhibitors prevent pulmonary endothelial hyperpermeability and acute lung injury by regulating heat shock protein 90 function. Am J Physiol Lung Cell Mol Physiol. 309 , L1410–L1419 (2015).26498249 69. Boucherat O , Chabot S , Paulin R , Trinh I , Bourgeois A , Potus F , Lampron MC , Lambert C , Breuils-Bonnet S , Nadeau V , Paradis R , Goncharova EA , Provencher S , & Bonnet S . HDAC6: A Novel Histone Deacetylase Implicated in Pulmonary Arterial Hypertension. Sci Rep. 7 , 4546 (2017).28674407 70. Espallergues J , Teegarden SL , Veerakumar A , Boulden J , Challis C , Jochems J , Chan M , Petersen T , Deneris E , Matthias P , Hahn CG , Lucki I , Beck SG , & Berton O . HDAC6 regulates glucocorticoid receptor signaling in serotonin pathways with critical impact on stress resilience. J Neurosci. 32 , 4400–4416 (2012).22457490 71. Fukada M , Hanai A , Nakayama A , Suzuki T , Miyata N , Rodriguiz RM , Wetsel WC , Yao TP , & Kawaguchi Y . Loss of deacetylation activity of Hdac6 affects emotional behavior in mice. PLoS One 7 , e30924 (2012).22328923 72. Tang Y , Lin YH , Ni HY , Dong J , Yuan HJ , Zhang Y , Liang HY , Yao MC , Zhou QG , Wu HY , Chang L , Luo CX , & Zhu DY . Inhibiting Histone Deacetylase 2 (HDAC2) Promotes Functional Recovery From Stroke. J Am Heart Assoc. 6 , pii: e007236 (2017).28982677 73. Hilaire G , Voituron N , Menuet C , & Ichiyama RM , Subramanian HH, & Dutschmann M The role of serotonin in respiratory function and dysfunction. Respir Physiol Neurobiol. 174 , 76–88 (2010).20801236 74. Pilowsky PM . Peptides, serotonin, and breathing: the role of the raphe in the control of respiration. Prog Brain Res. 209 , 169–189 (2014).24746048 75. Teran FA , & Massey CA , & Richerson GB Serotonin neurons and central respiratory chemoreception: where are we now? Prog Brain Res. 209 , 207–233 (2014).24746050 76. Taylor NC , Li A , & Nattie EE . Medullary serotonergic neurones modulate the ventilatory response to hypercapnia, but not hypoxia in conscious rats. J Physiol. 566 , 543–557 (2005).15878953 77. Hodges MR , & Richerson GB . Interaction between defects in ventilatory and thermoregulatory control in mice lacking 5-HT neurons. Respir Physiol Neurobiol. 164 , 350–357 (2008).18775520 78. Hodges MR , Tattersall GJ , Harris MB , McEvoy SD , Richerson DN , Deneris ES , Johnson RL , Chen ZF , & Richerson GB . Defects in breathing and thermoregulation in mice with near-complete absence of central serotonin neurons. J Neurosci. 28 , 2495–2505 (2008).18322094 79. Li A , & Nattie E . Serotonin transporter knockout mice have a reduced ventilatory response to hypercapnia (predominantly in males) but not to hypoxia. J Physiol. 586 , 2321–2329 (2008).18356199 80. Nattie E. Do medullary serotonin neurones modulate the ventilatory response to hypoxia? Acta Physiol (Oxf). 193 , 309 (2008). 81. Nucci TB , Branco LG , & Gargaglioni LH . 5-HT1A, but not 5-HT2 and 5-HT7, receptors in the nucleus raphe magnus modulate hypoxia-induced hyperpnoea. Acta Physiol (Oxf). 193 , 403–414 (2008).18363900 82. Richerson GB , Wang W , Tiwari J , & Bradley SR . Chemosensitivity of serotonergic neurons in the rostral ventral medulla. Respir Physiol. 129 , 175–89 (2001).11738653 83. Poddar S , Kesharwani D , & Datta M (2016) Histone deacetylase inhibition regulates miR-449a levels in skeletal muscle cells. Epigenetics 11 , 579–587 (2016).27184529 84. Luo L , Martin SC , Parkington J , Cadena SM , Zhu J , Ibebunjo C , Summermatter S , Londraville N , Patora-Komisarska K , Widler L , Zhai H , Trendelenburg AU , Glass DJ , & Shi J . HDAC4 Controls Muscle Homeostasis through Deacetylation of Myosin Heavy Chain, PGC-1α, and Hsc70. Cell Rep. 29 , 749–763.e12 (2019).31618641 85. Song S , Wen Y , Tong H , Loro E , Gong Y , Liu J , Hong S , Li L , Khurana TS , Chu M , & Sun Z . The HDAC3 enzymatic activity regulates skeletal muscle fuel metabolism. J Mol Cell Biol. 11 , 133–143 (2019).30428023 86. Powell FL , Milsom WK , & Mitchell GS . Time domains of the hypoxic ventilatory response. Respir Physiol. 112 , 123–134 (1998).9716296 87. Strohl KP . Periodic breathing and genetics. Respir Physiol Neurobiol. 135 , 179–185 (2003).12809618 88. Yamauchi M , Ocak H , Dostal J , Jacono FJ , Loparo KA , & Strohl KP . Post-sigh breathing behavior and spontaneous pauses in the C57BL/6J (B6) mouse. Resp Physiol Neurobiol. 162 , 117–125 (2008). 89. Yamauchi M , Dostal J , & Strohl KP . Post-hypoxic unstable breathing in the C57BL/6J mouse: effects of acetazolamide. Adv Exper Med Biol. 605 , 75–79 (2008).18085250 90. Wilkinson MH , Berger PJ , Blanch N , Brodecky V , & Jones CA . Paradoxical effect of oxygen administration on breathing stability following post-hyperventilation apnoea in lambs. J Physiol. 504 , 199–209 (1997).9350630 91. Coles SK , & Dick TE . Neurones in the ventrolateral pons are required for post-hypoxic frequency decline in rats. J Physiol. 497 , 79–94 (1996).8951713 92. Dick TE , & Coles SK . Ventrolateral pons mediates short-term depression of respiratory frequency after brief hypoxia. Respir Physiol. 121 , 87–100 (2000).10963767 93. Vizek M , Pickett CK , & Weil JV . Biphasic ventilatory response of adult cats to sustained hypoxia has central origin. J Appl Physiol. 63 , 1658–1664 (1987).3693202 94. Brown DR , Forster HV , Greene AS , & Lowry TF . Breathing periodicity in intact and carotid body-denervated ponies during normoxia and chronic hypoxia. J Appl Physiol. 74 , 1073–1082 (1993).8482645 95. Dempsey JA , Veasey SC , Morgan BJ , & O’Donnell C . Pathophysiology of sleep apnea. Physiol Rev. 90 , 47–112 (2010).20086074 96. Strohl KP , Butler JP , & Malhotra A . Mechanical properties of the upper airway. Compar Physiol. 2 , 1853–1872 (2012). 97. Tankersley CG . A genomic model for differential hypoxic ventilatory responses. Adv Exper Med Biol. 475 , 75–85 (2000).10849650 98. Tankersley CG . Selected contribution: variation in acute hypoxic ventilatory response is linked to mouse chromosome 9. J Appl Physiol. 90 , 1615–1622 (2001).11247969 99. Tankersley CG . Genetic aspects of breathing: on interactions between hypercapnia and hypoxia. Respir Physiol Neurobiol. 135 , 167–178 (2003).12809617 100. Tankersley CG , Fitzgerald RS , & Kleeberger SR . Differential control of ventilation among inbred strains of mice. Am J Physiol. 267 , R1371–R1377 (1994).7977867 101. Tankersley CG , DiSilvestre DA , Jedlicka AE , Wilkins HM , & Zhang L . Differential inspiratory timing is genetically linked to mouse chromosome 3. J Appl Physiol. 85 , 360–365 (1998).9655796 102. Tankersley CG , Elston RC , & Schnell AH . Genetic determinants of acute hypoxic ventilation: patterns of inheritance in mice. J Appl Physiol. 88 , 2310–2318 (2000).10846050 103. Han F , & Strohl KP Inheritance of ventilatory behavior in rodent models. Respir Physiol. 121 , 247–256 (2000).10963779 104. Tankersley CG , & Broman KW . Interactions in hypoxic and hypercapnic breathing are genetically linked to mouse chromosomes 1 and 5. J Appl Physiol. 97 , 77–84 (2004).14978007 105. Balbir A , Okumura M , Schofield B , Coram J , Tankersley CG , Fitzgerald RS , O’Donnell CP , & Shirahata M . Genetic regulation of chemoreceptor development in DBA/2J and A/J strains of mice. Adv Exper Med Biol. 580 , 99–104 (2006).16683704 106. Gillombardo CB , Yamauchi M , Adams MD , Dostal J , Chai S , Moore MW , Donovan LM , Han F , & Strohl KP . Identification of novel mouse genes conferring posthypoxic pauses. J Appl Physiol. 113 , 167–174 (2012).22539170 107. Tankersley CG , Haxhiu MA , & Gauda EB . Differential CO2-induced c-fos gene expression in the nucleus tractus solitarii of inbred mouse strains. J Appl Physiol. 92 , 1277–1284 (2002).11842068 108. Price ER , Han F , Dick TE , & Strohl KP . 7-nitroindazole and posthypoxic ventilatory behavior in the A/J and C57BL/6J mouse strains. J Appl Physiol. 95 , 1097–1104 (2003).12909599 109. Groeben H , Meier S , Tankersley CG , Mitzner W , & Brown RH . Heritable and pharmacological influences on pauses and apneas in inbred mice during anesthesia and emergence. Exper Lung Res. 31 , 839–853 (2005).16684716 110. Yamaguchi S , Balbir A , Schofield B , Coram J , Tankersley CG , Fitzgerald RS , O’Donnell CP , & Shirahata M . Structural and functional differences of the carotid body between DBA/2J and A/J strains of mice. J Appl Physiol. 9 , 1536–1542 (2003). 111. Yamaguchi S , Balbir A , Okumura M , Schofield B , Coram J , Tankersley CG , Fitzgerald RS , O’Donnell CP , & Shirahata M . Genetic influence on carotid body structure in DBA/2J and A/J strains of mice. Adv Exper Med Biol. 580 , 105–109 (2006).16683705 112. Chai S , Gillombardo CB , Donovan L , & Strohl KP . Morphological differences of the carotid body among C57/BL6 (B6), A/J, and CSS B6A1 mouse strains. Respir Physiol Neurobiol. 177 , 265–272 (2011).21555000 113. Govers AMAP Wiggers CRM , van Boxtel R , Mokry M , Nieuwenhuis EES , Creyghton MP , Bartels M , & Coffer PJ Transcriptomic and Epigenomic Profiling of Histone Deacetylase Inhibitor Treatment Reveals Distinct Gene Regulation Profiles Leading to Impaired Neutrophil Development. Hemasphere 3 , e270 (2019).31723844 114. Lu CY , Chang YC , Hua CH , Chuang C , Huang SH , Kung SH , Hour MJ , & Lin CW . Tubacin, an HDAC6 Selective Inhibitor, Reduces the Replication of the Japanese Encephalitis Virus via the Decrease of Viral RNA Synthesis. Int J Mol Sci. 18, pii: E954 (2017). 115. Ma XJ , Xu G , Li ZJ , Chen F , Wu D , Miao JN , Zhan Y , & Fan Y . HDAC-selective Inhibitor Cay10603 Has Single Anti-tumour Effect in Burkitt’s Lymphoma Cells by Impeding the Cell Cycle. Curr Med Sci. 39 , 228–236 (2019).31016515 116. Sanaei M , & Kavoosi F . Histone Deacetylases and Histone Deacetylase Inhibitors: Molecular Mechanisms of Action in Various Cancers. Adv Biomed Res. 8 , 63 (2019).31737580 117. Sun X , Xie Y , Sun X , Yao Y , Li H , Li Z , Yao R , & Xu K . The selective HDAC6 inhibitor Nexturastat A induces apoptosis, overcomes drug resistance and inhibits tumor growth in multiple myeloma. Biosci Rep. 39 , pii: BSR20181916 (2019).30782785

PMC010xxxxxx/PMC10312980.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398122 10.21203/rs.3.rs-2973069/v1 10.21203/rs.3.rs-2973069 preprint 1 Article Racial Discrimination is Associated with Binge-Eating Disorder in Early Adolescents: A Cross-Sectional Analysis Raney Julia H University of California, San Francisco Al-shoaibi Abubakr A University of California, San Francisco Shao Iris Y. University of California, San Francisco Ganson Kyle T University of Toronto Testa Alexander The University of Texas Health Science Center at Houston Jackson Dylan B. Johns Hopkins University He Jinbo Chinese University of Hong Kong, Shenzhen Glidden David V. University of California, San Francisco Nagata Jason M. University of California, San Francisco Author’s contributions J.H.R. was responsible for the co-development of the research study design and methods; J.H.R. also drafted the initial manuscript. A.A., I.S., K.T.G, A.T., D.B.J., and J.H. co-developed the study design, methods, and formal analysis; they also provided oversight and participated in the revision of the manuscript. J.M.N provided supervision; he also co-developed the conceptualization of the study, methods, and supported the analysis and manuscript revision. All authors approved the final submitted version. ✉ julia.raney@ucsf.edu 31 5 2023 rs.3.rs-2973069https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs2973069v1.pdf Background Racial and ethnic discrimination are known stressors and are associated with negative psychological and physical health outcomes. Previous studies have found relationships between racial/ethnic discrimination and binge-eating disorder (BED), though they have mainly focused on adult populations. The aim of this study was to determine associations between racial/ethnic discrimination and BED in a large, national cohort study of early adolescents. We further sought to explore associations between the racial/ethnic discrimination perpetrator (students, teachers, or other adults) and BED. Methods We analyzed cross-sectional data from the Adolescent Brain Cognitive Development Study (ABCD) (N = 11,075, 2018–2020). Logistic regression analyses examined associations between self-reported racial or ethnic discrimination and binge-eating behaviors and diagnosis. Racial/ethnic discrimination measures were assessed based on the Perceived Discrimination Scale, which measures experiences of discrimination based on race/ethnicity and frequency of ethnic discrimination by teachers, adults outside of school, and students. Binge-eating behaviors and diagnosis were based on the Kiddie Schedule for Affective Disorders and Schizophrenia (KSAD-5), adjusting for age, sex, race/ethnicity, household income, parental education, and site. Results In this racially diverse sample of adolescents (N = 11,075, mean age: 11 years), 4.7% of adolescents reported racial or ethnic discrimination and 1.1% met the criteria for BED at the one-year follow-up. In the adjusted models, racial/ethnic discrimination was associated with 3 times higher odds of having BED (OR 3.31, CI 1.66–7.74); when investigating associations between the racial/ethnic discrimination perpetrator (students, teachers, or other adults) and BED, experiencing ethnic discrimination by students and adults outside school were associated with significantly increased odds of BED diagnosis (OR 1.36, CI 1.10–1.68 & OR 1.42 CI 1.06–1.90, respectively); further, increased odds of binge eating behaviors was only significantly associated with ethnic discrimination perpetuated by students (OR 1.12, CI 1.02–1.23). Conclusions Children and adolescents who have experienced racial/ethnic discrimination, particularly when discrimination was perpetuated by other students, have higher odds of having binge-eating behaviors and diagnoses. Clinicians may consider screening for racial discrimination and providing anti-racist, trauma-informed care when evaluating and treating patients for BED. Plain English Summary Binge-eating disorder is associated with significant psychological and physical consequences including depression, anxiety, impaired relationships, and obesity. Recent research has demonstrated that many of these behaviors develop in early adolescence, a time of immense psychosocial development. Racial and ethnic discrimination are known stressors, and previous studies have found relationships between racial and ethnic discrimination and binge-eating disorder, though they have mainly focused on adult populations. This study helps fill that gap by using data from the Adolescent Brain Cognitive Development Study (ABCD) Study, the largest prospective study of adolescent brain development in the US. In this large, racially diverse, national study of 11,075 adolescents aged primary 11 – 12 years old, we find that early adolescents who have experienced racial/ethnic discrimination, particularly when discrimination was perpetuated by other students, have higher odds of having binge-eating behaviors and diagnoses. These findings have important school and clinical implications. For example, schools may consider implementing curricula focused on anti-racist practices that foster environments where all youth to thrive. In addition, we recommend that clinicians screen for racial discrimination and provide culturally sensitive, equity-focused care when evaluating and treating patients with binge-eating disorder. Racial discrimination binge-eating disorder adolescent health National Institutes of HealthK08HL159350 American Heart Association Career DevelopmentCDA34760281 Doris Duke Charitable Foundation2022056 ==== Body pmcBackground Binge-eating disorder (BED), characterized by consuming a large amount of food while feeling a loss of control and negative emotions,(1) is estimated to affect 4.5% of the population across the lifetime,(2) with those from minoritized backgrounds experience even higher rates of BED(3) and binge-eating behaviors.(4) BED is associated with significant psychological and physical consequences including depression, anxiety, impaired relationships, and obesity.(2, 5, 6) Given BED’s prevalence, consequences, and challenges with accessing treatment, identifying risk factors for the development of BED is critical to promote health and health equity. Racial discrimination, or personally mediated racism, has been recognized as a core driver of health inequities in adolescents and children.(7–9) Personally mediated racism includes experiences of stereotypes and prejudices about a person’s ability, intent, or motives on the basis of race.(10) Personally mediated racism can be expressed implicitly or explicitly and can be experienced directly or indirectly. Minoritized children and adolescents face personally mediated racism in their interactions with teachers and students at school, during extracurricular activities, and increasingly in online, digital environments. (11, 12) Experiencing racial harassment and taunts can over-activate the stress response, and have cascading effects including increased and prolonged levels of exposure to stress hormones and oxidative stress.(13) A growing body of literature has found associations between racial discrimination and BED. Several studies of Latino and African American adults have demonstrated significant associations between racial discrimination and binge eating.(14–16) For example, a US nationally representative sample of Latino individuals found a significant association with discrimination and binge eating, with the average age in the sample being 40 years old (range 18–97).(15) In addition, Assari et al.’s paper demonstrated significant associations between perceived discrimination and BED in a nationally representative example of African American adults (average age 42).(16) Studies have also significant associations between maladaptive eating behaviors in young adult Black women (ages 18–25)(17) and Latino young adults (ages 18–25).(18) However, little is known about the association between racial discrimination and the development of BED in early adolescents. A better understanding of the development of BED in this age-group is especially critical for several reasons. First, recent research has demonstrated that many of these behaviors develop in early adolescents, a time of immense psychosocial development.(19) A study of 10- to 11-year old children from the ABCD Study, a large, diverse, population-based sample, estimated the prevalence of BED to be 1.1%.(20) In addition, a population-based study of 14 year old early adolescents found a 14% prevalence rate of subclinical binge-eating behaviors.(21) Second, early adolescents experience racism and discrimination at unacceptable rates; a recent study estimated that 4.8% of 10- to 11-year old children reported being treated unfairly because of their race, ethnicity or color, and 10% of Black children reported experiencing racism.(22) Given the prevalence of perceived discrimination experienced by youth in this age group, it is critical to characterize the public health effects and rapidly implement antiracism practices. Given the significant impact that peers, teachers, and non-caregiver adults have on early adolescent development,(23–25) we further sought to characterize the impacts of expressed by these varying groups. The purpose of the current study was to examine the associations between racial discrimination and BED among a large, diverse cohort of early-adolescents ages 9–14 years old (primarily ages 10–11). Methods Study population This study uses survey data from the Adolescent Brain Cognitive Development (ABCD) Study to determine the association between racial discrimination and BED among US early adolescents. The ABCD Study is a large, prospective cohort study of brain development, health, and health behaviors among US adolescents across 21 recruitment sites.(26) Study design and recruitment strategies have previously been described.(27) We included data collected between 2018–2020, corresponding to Year one of the ABCD Study, the first year adolescent-reported racial discrimination was assessed. Participants missing data for i) sociodemographic characteristics or ii) all discrimination questions (n = 710) were excluded, yielding the total sample of 11,075. All participants gave assent, and parents/caregivers provided signed informed consent. The ABCD Study protocol was approved by the Institutional Review Board of the University of California, San Diego and at each respective study site. Exposure: Racial discrimination Racial discrimination was measured using the Perceived Discrimination Scale,(28,29) which was developed to measure adolescents’ perception of being unaccepted in society or being unwanted based on their racial or ethnic background or skin color. Adolescents were asked, “In the past 12 months, have you felt discriminated against: because of your race, ethnicity, or color?” In addition, adolescents were asked how often they had been treated unfairly or negatively because of their ethnic background by each of the following groups: teachers, adults outside the school, and students (1 = almost never; 2 = rarely; 3 = sometimes; 4 = often; 5 = very often). Outcome: Binge-eating disorder BED diagnosis and behaviors were assessed that the one-year follow-up through parent/caregiver responses to the Kiddie Schedule for Affective Disorders and Schizophrenia (KSADS-5), a computerized tool developed to categorize child and adolescent mental health based on the DSM-5. (1,30) Parents/caregivers completed all modules of the KSADS-5 to characteristics, frequency, and duration of their child’s binge-eating behaviors as well as associated distress. The presence of binge-eating behaviors was assessed by asking if their child experienced a loss of control of their eating and ate way more than he/she needed. BED was determined using the KSADS-5 computerized scoring system, where responses to the survey questions were extrapolated into their respective diagnosis based on reported behaviors corresponding to the DSM-5. Although bulimia nervosa (BN) also consists of binge eating symptoms, the prevalence of BN in the sample was low and therefore this study focused on binge-eating behaviors and BED. Covariates We selected potential confounders for the association between racial discrimination and BED based on prior literature and theory.(16,31) Age, sex (male, female), race/ethnicity (White, Latino/Hispanic, Black, Asian, Native American, Other), nativity (youth born in US or outside of the US), household income ($24,999 or less, $25,000 through $49,999, $50,000 through $74,999, $75,000 through $99,999, $100,000 through $199,999, $200,000 and greater), and highest parent education (high school or less vs. college or more) were selected from parent self-report data at baseline. ABCD Study site (21 total sites) was also included to adjust for potential regional variation. Statistical analyses Unadjusted and adjusted logistic regressions were conducted using Stata 17.0 (StataCorp, College Station, TX) to estimate associations between past year experiences of racial/ethnic discrimination and BED diagnosis and behaviors. In addition, unadjusted and adjusted logistic regression analyses estimated the association between frequency of ethnic discrimination (teachers, adults outside school, students), and BED diagnosis and behaviors. The ABCD study sociodemographic varaiabeles were standardized to match the distribution American Community Survey from the U.S. Census.(32) Results The 11,075 adolescent respondents were racially and ethnically diverse (53.4% White, 19.6% Latino/Hispanic, 16.5% Black, 5.6% Asian, 3.2% Native American, 1.4% Other, Table 1). Binge-eating behaviors and diagnosis were relative rare at 7.9% and 1.1%, respectively. Approximately one in twenty youth reported experiencing racial or ethnic discrimination in the past year. In addition, adolescents reported higher rates of students perpetuating ethnic discrimination that teachers or other adults outside of school. In both the adjusted and unadjusted models, racial/ethnic discrimination was associated with increased binge-eating behaviors and binge-eating disorder diagnosis (Table 2). In the adjusted models, adolescents who reported perceived discrimination had 3.31 higher odds of BED (95% CI 1.66 – 6.63). Increased frequency of ethnic discrimination by students was also significantly associated with a higher odds of BED diagnosis and behaviors. In addition, respondents who reported more frequent ethnic discrimination by adults outside of school had significantly higher odds of BED diagnosis. Discussion In this national, sociodemographically diverse sample of early adolescents in the U.S., we found that experiencing racial/ethnic discrimination was associated with binge-eating behaviors and diagnosis, even when adjusting for confounding factors including race, sex, nativity, parental education, and socioeconomic status. The relationship between discrimination and binge-eating is consistent with prior studies in minoritized adult populations that have demonstrated associations between experiencing racial/ethnic discrimination and binge-eating behaviors in Latino and Black the general adult population and young adults.(14, 16) Our findings contribute to the literature by demonstrating that perceived discrimination is significantly associated with higher odds of binge-eating behaviors and diagnosis in a national, diverse population of US early adolescents; importantly, early-adolescents represent an under-researched age group whose developmental period is vulnerable to developing health-related risk behaviors.(19, 33) As BED is associated with significant distress, morbidity, and mortality, it is critical to investigate risk factors in this age group to design primary and secondary prevention interventions.(34, 35) A potential mechanism of this relationship is experiencing discrimination may impact adolescents’ self-esteem and increases risk for depression symptoms, both of which are associated with increased rates of BED.(28, 36, 37) In addition, several theoretical models have conceptualized racial/ethnic discrimination as an important stressor that drives binge eating from maladaptive coping responses from the resultant increased stress and changes in cortisol levels.(38, 39) Our study further adds to the literature by exploring how unique groups of perpetrators influence the association between discrimination and BED. In our study, adolescents reported students to be the most common perpetuators of ethnic discrimination with one if four adolescents reporting experiencing ethnic discrimination by students rarely or more frequently; in addition, reporting ethnic discrimination perpetuated by students was significantly associated with increased odds of binge-eating disorder behaviors and diagnosis. The significant impact of peer discrimination on adolescent’s mental health has been supported in prior literature.(24, 40) From a developmental perspective, peer discrimination may be particularly impactful for early adolescents as they increasing spend time outside of the home and rely more on peers for psychosocial acceptance, self-concept, socialization, and identity formation.(41, 42) Several studies have shown that peer victimization in early adolescence is predictive of subsequent development of depressive symptoms.(43, 44) Our study builds upon these studies by highlighting that discrimination is also associated with BED in a national, diverse sample of early adolescents. Of note, experiencing discrimination by other adults outside school was also associated with a significantly higher odds of BED diagnosis. This is consistent with literature that shows the important influence that nonparental adults, such as mentors and police, can have on adolescent mental health.(45–47) Importantly, in the adjusted models, discrimination perpetuated by teachers was not significantly associated with binge-eating behaviors or diagnosis and discrimination by adults outside school was not associated with binge eating symptoms. However, this may be partially due to the smaller sample size in these groups. Prior studies have found that teachers play a critical role in adolescent development and mental health.(48, 49) Further studies should continue to investigate the relationships between ethnic discrimination perpetuated by teachers and adults outside school and binge eating. This study has several limitations. This study used parent report of binge eating diagnosis and behaviors. While parents are important reporters for eating disorders in early adolescents as children have less insight into their eating behaviors,(26, 50) parent and child reports of binge-eating behaviors have a tendency of low concordance.(26, 50, 51) In addition, this study did not explore the setting where the discrimination occurred so we are unable to assess how the location (ie school, virtual settings, community) estimate the impacts BED behaviors. Conclusions This study demonstrates that experiencing racial or ethnic discrimination in early adolescence is associated with BED diagnosis and symptoms, which has important school, clinical, and public policy implications. For example, schools may consider implementing curricula focused on anti-racist practices that foster environments where all youth to thrive.(52) In addition, minoritized populations, for instance, have historically received inadequate access to eating disorder care and inclusion in eating disorders research, which increases the risk of delayed and poorer outcomes.(53) Policy changes that that target these systemic issues, such as increased education about BED among diverse populations, increased food assistance among marginalized communities,(54) and increased access to eating disorder trained mental health professionals,(53) may profoundly impact the risk of BED. The US Preventive Services Task Force (USPSTF) recently reviewed eating disorder screening in asymptomatic adolescents and adults and determined there to be insufficient evidence to recommend routine screening in this population, especially among racial/ethnic minority populations (55). However, clinicians may still consider screening for eating disorder behaviors in early adolescents with significant risk factors, such as racial discrimination, given the significant physician and mental health consequences of eating disorders.(56) Acknowledgements The ABCD Study was supported by the Nacional Institutes of Health and additional federal partners under award numbers U01DA041022, U01DA041025, U01DA041028, U01DA041048, U01DA041089, U01DA041093, U01DA041106, U01DA041117, U01DA041120, U01DA041134, U01DA041148, U01DA041156, U01DA041174, U24DA041123, and U24DA041147. A full list of supporters is available at https://abcdstudy.org/federal-partners/. A listing of participating sites and a complete listing of the study investigators can be found at https://abcdstudy.org/principal-investigators.html. ABCD consortium investigators designed and implemented the study and/or provided data but did not necessarily participate in the analysis or writing of this report. Funding J.M.N. was funded by the National Institutes of Health (K08HL159350), the American Heart Association Career Development Award (CDA34760281), and the Doris Duke Charitable Foundation (2022056). The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Availability of data and materials Data used in the preparation of this article were obtained from the ABCD Study (https://abcdstudy.org), held in the NIMH Data Archive (NDA). Investigators can apply for data access through the NDA (https://nda.nih.gov/). Abbreviations ACEs Adverse Childhood Experiences ABCD Adolescent Brain Cognitive Development study BED Binge-eating disorder KSADS-5 Kiddie Schedule for Affective Disorders and Schizophrenia BN Bulimia Nervosa SD Standard Deviation Table 1: Sociodemographic characteristics of participants in the Adolescent Brain Cognitive Development Study (ABCD) Study, 2018–2020, (n=11,075) Mean (SD) or % Demographic characteristics Age 12.0 (0.7) Sex Female 48.8% Male 51.2% Race/ethnicity White 53.4% Latino / Hispanic 19.6% Black 16.5% Asian 5.6% Native American 3.2% Other 1.4% Nativity Youth born in U.S. 96.3% Youth born outside U.S. 3.7% Highest parental education High school education or less 15.6% College education or more 84.4% Household income $24,999 or less 16.9% $25,000 to $49,999 20.2% $50,000 to $74,999 18.2% $75,000 to $99,999 16.0% $100,000 to $199,999 21.7% $200,000 and greater 7.0% Type of discrimination reported Discrimination because of race, ethnicity, or color 4.7% Been treated unfairly or negatively because of your ethnic background by: Teachers Almost never 92.0% Rarely 4.3% Sometimes 2.1% Often 0.9% Very Often 0.7% Adults outside school Almost never 90.2% Rarely 6.5% Sometimes 2.2% Often 0.6% Very Often 0.5% Students Almost never 74.9% Rarely 13.1% Sometimes 7.8% Often 2.2% Very Often 2.0% Binge eating Binge-eating behaviors 7.9% Binge-eating disorder diagnosis 1.1% ABCD Study sociodemographic variables were standardized to match the distribution American Community Survey from the U.S. Census. Table 2: Associations between discrimination and binge-eating behaviors and diagnosis (N=11,075) Panel A: Bivariate Model Binge-eating behaviors Binge-eating disorder diagnosis OR OR (95% CI) (95% CI) Discrimination because of race, ethnicity, or color 2.22*** 4.31*** (1.62 – 3.05) (2.40 – 7.74) Been treated unfairly or negatively because of your ethnic background by: Teachers 1.20** 1.44** (1.06 – 1.36) (1.16 – 1.79) Adults outside school 1.19* 1.55** (1.03 – 1.37) (1.19 – 2.01) Students 1 18*** 1 54*** (1.08 – 1.28) (1.28 – 1.86) Panel B: With Confounding Variables OR OR (95% CI) (95% CI) Discrimination because of race, ethnicity, or color 2 12*** 3.31* (1.50 – 3.00) (1.66 – 6.63) Been treated unfairly or negatively because of your ethnic background by:  Teachers 1.12 1.26 (0.97–1.23) (0.98 – 1.60)  Adults outside school 1.11 1.42* (0.95 – 1.30) (1.06 – 1.90)  Students 1.12* 1.36** (1.02 – 1.23) (1.10 – 1.68) * indicates p<0.05 ** indicates p<0.01 *** indicates significant at <0.001. ABCD Study sociodemographic variables were standardized to match the distribution American Community Survey from the U.S. Census. Panel B models include sex, race/ethnicity, household income, parent education, and site. Declarations Ethics approval and consent to participate Written informed consent and assent were obtained from the parent/guardian and adolescent, respectively, to participate in the ABCD Study. The University of California, San Diego provided centralized institutional review board (IRB) approval and each participating site received local IRB approval: Children’s Hospital Los Angeles, Los Angeles, California Florida International University, Miami, Florida Laureate Institute for Brain Research, Tulsa, Oklahoma Medical University of South Carolina, Charleston, South Carolina Oregon Health and Science University, Portland, Oregon SRI International, Menlo Park, California University of California San Diego, San Diego, California University of California Los Angeles, Los Angeles, California University of Colorado Boulder, Boulder, Colorado University of Florida, Gainesville, Florida University of Maryland at Baltimore, Baltimore, Maryland University of Michigan, Ann Arbor, Michigan University of Minnesota, Minneapolis, Minnesota University of Pittsburgh, Pittsburgh, Pennsylvania University of Rochester, Rochester, New York University of Utah, Salt Lake City, Utah University of Vermont, Burlington, Vermont University of Wisconsin—Milwaukee, Milwaukee, Wisconsin Virginia Commonwealth University, Richmond, Virginia Washington University in St. Louis, St. Louis, Missouri Yale University, New Haven, Connecticut All methods were carried out in accordance with relevant guidelines and regulations. Consent for publication Not applicable Competing interests The authors declare that they have no competing interests. ==== Refs References 1. American Psychiatric Association D, American Psychiatric Association. Diagnostic and statistical manual of mental disorders: DSM-5. Vol. 5 . Washington, DC: American psychiatric association; 2013. 2. Hudson JI , Hiripi E , Pope HG , Kessler RC . The Prevalence and Correlates of Eating Disorders in the National Comorbidity Survey Replication. Biol Psychiatry. 2007 Feb 1;61 (3 ):348–58.16815322 3. Alegria M , Woo M , Cao Z , Torres M , Meng X li, Striegel-Moore R. Prevalence and Correlates of Eating Disorders in Latinos in the U.S. Int J Eat Disord. 2007 Nov;40 (Suppl ):S15–21.17584870 4. Marques L , Alegria M , Becker AE , Chen C nan , Fang A , Chosak A , Comparative Prevalence, Correlates of Impairment, and Service Utilization for Eating Disorders across U.S. Ethnic Groups: Implications for Reducing Ethnic Disparities in Health Care Access for Eating Disorders. Int J Eat Disord. 2011 Jul;44 (5 ):412–20.20665700 5. Rosenbaum DL , White KS . The relation of anxiety, depression, and stress to binge eating behavior. J Health Psychol. 2015 Jun 1;20 (6 ):887–98.26032804 6. Swanson SA , Crow SJ , Le Grange D , Swendsen J , Merikangas KR . Prevalence and Correlates of Eating Disorders in Adolescents: Results From the National Comorbidity Survey Replication Adolescent Supplement. Arch Gen Psychiatry. 2011 Jul 1;68 (7 ):714–23.21383252 7. Gee GC , Walsemann KM , Brondolo E . A Life Course Perspective on How Racism May Be Related to Health Inequities. Am J Public Health. 2012 May;102 (5 ):967–74.22420802 8. Pachter LM , Coll CG . Racism and Child Health: A Review of the Literature and Future Directions. J Dev Behav Pediatr. 2009 Jun;30 (3 ):255.19525720 9. Trent M , Dooley DG , Dougé J , SECTION ON ADOLESCENT HEALTH, COUNCIL ON COMMUNITY PEDIATRICS, COMMITTEE ON ADOLESCENCE, The Impact of Racism on Child and Adolescent Health. Pediatrics. 2019 Aug 1;144 (2 ):e20191765.31358665 10. Jones CP . Invited Commentary: “Race,” Racism, and the Practice of Epidemiology. Am J Epidemiol. 2001 Aug 15;154 (4 ):299–304.11495851 11. Stevens R , Gilliard-Matthews S , Dunaev J , Woods M , Brawner BM . The Digital Hood: Social Media Use among Youth in Disadvantaged Neighborhoods. New Media Soc. 2017 Jun;19 (6 ):950–67.28694736 12. Njoroge WFM , Forkpa M , Bath E . Impact of Racial Discrimination on the Mental Health of Minoritized Youth. Curr Psychiatry Rep. 2021 Oct 14;23 (12 ):81.34648076 13. Cohen S , Janicki-Deverts D , Doyle WJ , Miller GE , Frank E , Rabin BS , Chronic stress, glucocorticoid receptor resistance, inflammation, and disease risk. Proc Natl Acad Sci. 2012 Apr 17;109 (16 ):5995–9.22474371 14. Beccia AL , Jesdale WM , Lapane KL . Associations between perceived everyday discrimination, discrimination attributions, and binge eating among Latinas: results from the National Latino and Asian American Study. Ann Epidemiol. 2020 May;45 :32–9.32340835 15. Johnson SN , Forbush KT , Swanson TJ . The impact of discrimination on binge eating in a nationally representative sample of Latine individuals. Int J Eat Disord. 2022 Aug;55 (8 ):1120–9.35796067 16. Assari S. Perceived Discrimination and Binge Eating Disorder; Gender Difference in African Americans. J Clin Med. 2018 Apr 24;7 (5 ):89.29695062 17. Brown KL , Graham AK , Perera RA , LaRose JG . Eating to cope: Advancing our understanding of the effects of exposure to racial discrimination on maladaptive eating behaviors. Int J Eat Disord. 2022 Dec;55 (12 ):1744–52.36205358 18. Hagiwara N , Green TL , Moreno O , Smith D , Corona R . Ethnic discrimination and weight outcomes among Latinx emerging adults: Examinations of an individual-level mediator and cultural moderators. Cultur Divers Ethnic Minor Psychol. 2021;27 :189–200.32212740 19. Dahl RE , Allen NB , Wilbrecht L , Suleiman AB . Importance of investing in adolescence from a developmental science perspective. Nature. 2018 Feb;554 (7693 ):441–50.29469094 20. Chu J , Raney JH , Ganson KT , Wu K , Rupanagunta A , Testa A , Adverse childhood experiences and binge-eating disorder in early adolescents. J Eat Disord. Forthcoming; 21. Derks IPM , Harris HA , Staats S , Gaillard R , Dieleman GC , Llewellyn CH , Subclinical binge eating symptoms in early adolescence and its preceding and concurrent factors: a population-based study. J Eat Disord. 2022 Nov 23;10 (1 ):180.36424658 22. Nagata JM , Ganson KT , Sajjad OM , Benabou SE , Bibbins-Domingo K . Prevalence of Perceived Racism and Discrimination Among US Children Aged 10 and 11 Years. JAMA Pediatr. 2021 Aug;175 (8 ):861–3.33999104 23. Bennett M , Roche KM , Huebner DM , Lambert SF . School Discrimination and Changes in Latinx Adolescents’ Internalizing and Externalizing Symptoms. J Youth Adolesc. 2020 Oct;49 (10 ):2020–33.32447566 24. Bennett M , Roche KM , Huebner DM , Lambert SF . Peer Discrimination, Deviant Peer Affiliation, and Latino/a Adolescent Internalizing and Externalizing Symptoms: A Prospective Study. J Clin Child Adolesc Psychol. 2022 Jul 19;0 (0 ):1–17. 25. McFarland L , Murray E , Phillipson S . Student–teacher relationships and student self-concept: Relations with teacher and student gender. Aust J Educ. 2016 Apr 1;60 (1 ):5–25. 26. Barch DM , Albaugh MD , Avenevoli S , Chang L , Clark DB , Glantz MD , Demographic, physical and mental health assessments in the adolescent brain and cognitive development study: Rationale and description. Dev Cogn Neurosci. 2017 Nov 3;32 :55–66.29113758 27. Garavan H , Bartsch H , Conway K , Decastro A , Goldstein RZ , Heeringa S , Recruiting the ABCD sample: Design considerations and procedures. Dev Cogn Neurosci. 2018 Apr 16;32 :16–22.29703560 28. Cano MÁ , Schwartz SJ , Castillo LG , Romero AJ , Huang S , Lorenzo-Blanco EI , Depressive Symptoms and Externalizing Behaviors among Hispanic Immigrant Adolescents: Examining Longitudinal Effects of Cultural Stress. J Adolesc. 2015 Jul;42 :31–9.25899132 29. Phinney JS , Madden T , Santos LJ . Psychological variables as predictors of perceived ethnic discrimination among minority and immigrant adolescents. J Appl Soc Psychol. 1998;28 :937–53. 30. Townsend L , Kobak K , Kearney C , Milham M , Andreotti C , Escalera J , Development of Three Web-Based Computerized Versions of the Kiddie Schedule for Affective Disorders and Schizophrenia Child Psychiatric Diagnostic Interview: Preliminary Validity Data. J Am Acad Child Adolesc Psychiatry. 2020 Feb 1;59 (2 ):309–25.31108163 31. Berry OO , Londoño Tobón A , Njoroge WFM . Social Determinants of Health: the Impact of Racism on Early Childhood Mental Health. Curr Psychiatry Rep. 2021 Mar 12;23 (5 ):23.33712922 32. Heeringa SG , Berglund PA . A Guide for Population-based Analysis of the Adolescent Brain Cognitive Development (ABCD) Study Baseline Data [Internet]. 2020 Feb [cited 2022 Jan 6] p. 2020.02.10.942011. Available from: 10.1101/2020.02.10.942011v1 33. Skinner AC , Perrin EM , Moss LA , Skelton JA . Cardiometabolic Risks and Severity of Obesity in Children and Young Adults. N Engl J Med. 2015 Oct;373 (14 ):1307–17.26422721 34. van Hoeken D , Hoek HW . Review of the burden of eating disorders: mortality, disability, costs, quality of life, and family burden. Curr Opin Psychiatry. 2020 Nov;33 (6 ):521–7.32796186 35. O’Brien D , Fisher J . 5 ways bosses can reduce the stigma of mental health at work.:6 . 36. Cheng HL . Disordered Eating Among Asian/Asian American Women: Racial and Cultural Factors as Correlates. Couns Psychol. 2014 Aug 1;42 (6 ):821–51. 37. Paradies Y , Ben J , Denson N , Elias A , Priest N , Pieterse A , Racism as a Determinant of Health: A Systematic Review and Meta-Analysis. PLoS ONE. 2015 Sep 23;10 (9 ):e0138511.26398658 38. Heatherton TF , Baumeister RF . Binge eating as escape from self-awareness. Psychol Bull. 1991;110 :86–108.1891520 39. Zeiders KH , Doane LD , Roosa MW . Perceived Discrimination and Diurnal Cortisol: Examining Relations among Mexican American Adolescents. Horm Behav. 2012 Apr;61 (4 ):541–8.22342577 40. Dulin-Keita A , Hannon Iii L , Fernandez JR , Cockerham WC . The defining moment: children’s conceptualization of race and experiences with racial discrimination. Ethn Racial Stud. 2011 Apr 1;34 (4 ):662–82.21532908 41. Hughes DL , Watford JA , Del Toro J . Chapter One - A Transactional/Ecological Perspective on Ethnic–Racial Identity, Socialization, and Discrimination. In: Horn SS , Ruck MD , Liben LS , editors. Advances in Child Development and Behavior [Internet]. JAI; 2016 [cited 2023 Mar 29]. p. 1–41. (Equity and Justice in Developmental Science: Implications for Young People, Families, and Communities; vol. 51 ). Available from: https://www.sciencedirect.com/science/article/pii/S0065240716300209 42. DiClemente RJ , Santelli JS , Crosby R . Adolescent Health: Understanding and Preventing Risk Behaviors. John Wiley & Sons; 2009. 610 p. 43. Delgado MY , Nair RL , Updegraff KA , Umaña-Taylor AJ . Discrimination, Parent-Adolescent Conflict, and Peer Intimacy: Examining Risk and Resilience in Mexican-Origin Youths’ Adjustment Trajectories. Child Dev. 2019 May;90 (3 ):894–910.28940221 44. Stein GL , Castro-Schilo L , Cavanaugh AM , Mejia Y , Christophe NK , Robins R . When Discrimination Hurts: The Longitudinal Impact of Increases in Peer Discrimination on Anxiety and Depressive Symptoms in Mexican-origin Youth. J Youth Adolesc. 2019 May;48 (5 ):864–75.30879164 45. Hurd NM , Stoddard SA , Bauermeister JA , Zimmerman MA . Natural Mentors, Mental Health, and Substance Use: Exploring Pathways via Coping and Purpose. Am J Orthopsychiatry. 2014 Mar;84 (2 ):190–200.24826935 46. Jindal M , Mistry KB , Trent M , McRae A , Thornton RLJ . Police Exposures and the Health and Well-being of Black Youth in the US: A Systematic Review. JAMA Pediatr. 2022 Jan 1;176 (1 ):78–88.34491292 47. Szapocznik J , Coatsworth JD . An ecodevelopmental framework for organizing the influences on drug abuse: A developmental model of risk and protection. In: Drug abuse: Origins & interventions. Washington, DC, US: American Psychological Association; 1999. p. 331–66. 48. Maene C , D’hondt F , Van Lissa CJ , Thijs J , Stevens PAJ . Perceived Teacher Discrimination and Depressive Feelings in Adolescents: The Role of National, Regional, and Heritage Identities in Flemish Schools. J Youth Adolesc. 2022;51 (12 ):2281–93.35987976 49. Butler-Barnes ST , Leath S , Inniss-Thompson MN , Allen PC , D’Almeida MEDA , Boyd DT . Racial and gender discrimination by teachers: Risks for Black girls’ depressive symptomatology and suicidal ideation. Cultur Divers Ethnic Minor Psychol. 2022 Oct;28 (4 ):469–82. 50. Braet C , Soetens B , Moens E , Mels S , Goossens L , Van Vlierberghe L . Are two informants better than one? Parent–child agreement on the eating styles of children who are overweight. Eur Eat Disord Rev. 2007;15 (6 ):410–7.17960860 51. Tanofsky-Kraff M , Yanovski SZ , Yanovski JA . Comparison of child interview and parent reports of children’s eating disordered behaviors. Eat Behav. 2005 Jan 1;6 (1 ):95–9.15567115 52. Priest N , Alam O , Truong M , Sharples R , Nelson J , Dunn K , Promoting proactive bystander responses to racism and racial discrimination in primary schools: a mixed methods evaluation of the ‘Speak Out Against Racism’ program pilot. BMC Public Health. 2021 Jul 21;21 :1434.34289830 53. Moreno R , Buckelew SM , Accurso EC , Raymond-Flesch M . Disparities in access to eating disorders treatment for publicly-insured youth and youth of color: a retrospective cohort study. J Eat Disord. 2023 Jan 24;11 :10.36694235 54. Food insecurity and binge-eating disorder in early adolescence - Nagata - International Journal of Eating Disorders - Wiley Online Library [Internet]. [cited 2023 Apr 27]. Available from: 10.1002/eat.23944 55. Feltner C , Peat C , Reddy S , Riley S , Berkman N , Middleton JC , Screening for Eating Disorders in Adolescents and Adults: Evidence Report and Systematic Review for the US Preventive Services Task Force. Vol. 327 , JAMA - Journal of the American Medical Association. 2022. 56. Nagata JM , Golden NH . New US Preventive Services Task Force Recommendations on Screening for Eating Disorders. Vol. 182 , JAMA Internal Medicine. 2022.

PMC010xxxxxx/PMC10312985.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398093 10.21203/rs.3.rs-2986616/v1 10.21203/rs.3.rs-2986616 preprint 1 Article The Superior Colliculus Projection Upon the Macaque Inferior Olive May Paul J. University of Mississippi Medical Center Warren Susan University of Mississippi Medical Center Kojima Yoshiko University of Washington Contributions: The anatomy experiments were designed by Paul J. May and were carried out by Susan Warren and Paul J. May. The physiology experiments were designed and carried out by Yoshiko Kojima. The data was analyzed by all authors. Figures were prepared by Paul J. May and Susan Warren. The manuscript was initially drafted by Paul J. May and edited by all three authors. ✉ pauljmay55@gmal.com 30 5 2023 rs.3.rs-2986616https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs2986616v1.pdf Saccade accommodation is a productive model for exploring the role of the cerebellum in behavioral plasticity. In this model, the target is moved during the saccade, gradually inducing a change in the saccade vector as the animal adapts. The climbing fiber pathway from the inferior olive provides a visual error signal generated by the superior colliculus that is believed to be crucial for cerebellar adaptation. However, the primate tecto-olivary pathway has only been explored using large injections of the central portion of the superior colliculus. To provide a more detailed picture, we have made injections of anterograde tracers into various regions of the macaque superior colliculus. As shown previously, large central injections primarily label a dense terminal field within the C subdivision at caudal end of the contralateral medial inferior olive. Several, previously unobserved, sites of sparse terminal labeling were noted: bilaterally in the dorsal cap of Kooy and ipsilaterally in C subdivision of the medial inferior olive. Small, physiologically directed, injections into the rostral, small saccade portion of the superior colliculus produced terminal fields in the same regions of the medial inferior olive, but with decreased density. Small injections of the caudal superior colliculus, where large amplitude gaze changes are encoded, again labeled a terminal field located in the same areas. The lack of a topographic pattern within the main tecto-olivary projection suggests that either the precise vector of the visual error is not transmitted to the vermis, or that encoding of this error is via non-topographic means. Oculomotor Non-human Primate Saccade Adaptation Motor Learning Cerebellum National Eye InstituteEY014263 U.S. National Institutes of Health to Paul J. May & Susan Warren and National Eye InstituteEY023277 U.S. National Institutes of Health to Yoshiko KojimaU.S. National Institutes of HealthOD010425 Washington National Primate Research CenterVision Research Core for the University of WashingtonP30EY001730 ==== Body pmcINTRODUCTION Normal behavior requires movement accuracy. This accuracy must be maintained despite the fact that the body changes both due to growth at early ages and decreased muscle effectiveness later in adulthood. The ability to change motor output in response to sensory feedback indicating an inaccurate movement has been made is termed motor adaptation, and when the inaccuracy is due to an undershooting or overshooting of the target, it is more specifically designated gain adaptation. A common and productive behavioral model for gain adaptation is saccade adaptation (Straube et al., 1997). In this model, the target is moved as the individual is making a saccade. Since visual sensory input is suppressed during a saccade, the perception of the individual is that the saccade over- or under-shot, not that the target moved. In response to repeated trials in which the saccade does not properly obtain the target, the gain of the saccade is changed in an attempt to compensate for missing the target. The same system can be used to demonstrate that the nervous system can also compensate for inaccuracies related to saccade direction (Noto et al., 1999). In order to trigger gain adaptation, the nervous system produces a visual sensory signal indicating the extent of the targeting error, and then uses this error signal to manipulate the gain of the motor output (Wallman & Fuchs, 1998). The cerebellum is known to play a critical role in the process of saccadic gain adaptation. Optican and Robinson (1980) demonstrated that the central nervous system is able to adapt to surgically weakening of the horizontal recti via tenectomy, but this capacity to adapt is dramatically impaired by ablation of the cerebellum, and more specifically, the oculomotor vermis. Stimulation of the oculomotor vermis produces eye movements in a manner that suggests it is topographically organized (Ron & Robinson, 1973; Noda et al., 1987), although there is no evidence of topography based on the simple and complex spike activity of vermal Purkinje cells (Kojima et al., 2010A; 2010B; Kojima, 2019). Inactivation of the fastigial nucleus, which is a necessary node for output by the oculomotor vermis, largely eliminates saccade adaptation (Robinson et al., 2002; 2006). Pharmacologic manipulation of the oculomotor vermis also affects saccadic motor learning, but seems to mainly involve amplitude increase adaptation (Kojima et al., 2010B; 2011). Recordings from Purkinje cells in the oculomotor vermis show that they changed their complex spiking pattern during the adaptation paradigm (Soetedjo & Fuchs, 2006). Initially, the modulation of complex spike activity in Purkinje cells was thought to just signal the direction, but not the magnitude of the required saccade adaptation (Soetedjo & Fuchs, 2006), but further examination suggested that both direction and size were critical to modulating complex spike activity when a visual sensory error was detected, with most cells tuned to errors of less than 5° (Soetedjo et al., 2008). These changes in complex spike activity are appropriate for driving the adaptive changes in the brainstem motor system that generates saccades (Kojima et al., 2010A). The complex spikes observed in Purkinje cells are produced by input from climbing fibers that originate in the inferior olive (IO). The source of the visual error signal that drives IO neurons to change the pattern of complex spikes appears to be the superior colliculus (SC). Evidence for this assertion comes from electrical stimulation of the SC (Kaku et al., 2009; Soetedjo et al., 2009) where stimulation of the rostral superior colliculus at levels that do not produce saccades can still produce saccade adaptation. Furthermore, inactivation of the rostral colliculus, which encodes small saccades, severely impairs saccade adaptation (Kojima & Soetedjo, 2018). This is not due to changes in the saccade-related bursts flowing from the SC to the brainstem gaze centers, as the changes in motor activity that occur during saccade adaptation appear to occur downstream of the SC (Melis & van Gisbergen, 1996; Frens & Van Opstal, 1997; Edelman & Goldberg, 2002; Takeichi et al., 2007). The visual error signal makes its way from the SC to the IO by way of the tecto-olivary projection. This projection is a portion of the crossed predorsal bundle pathway that provides collicular input to the gaze centers in the brainstem and cervical spinal cord (May & Porter, 1992; May, 2006). The main evidence for this pathway in the macaque comes from a study using anterograde transport of tritiated amino acids (Harting, 1977). This study showed an exclusively contralateral projection that terminated within the medial subnucleus of the IO (IOM). However, the animals illustrated for this study had injection sites located in the middle of the SC. As such, they did not indicate whether the tecto-olivary projection is a topographic one, a critical point since saccades of different sizes and directions are mapped in an ordered manner within the intermediate gray layer of the SC. Moreover, they did not include the rostral SC where small visual error signals used to initiate saccade adaptation are produced. A fairly similar terminal distribution of the tecto-olivary projection in the IOM was observed in the cat, although in this species, a sparse ipsilateral projection was present, as well (Weber et al., 1978). More recently, a study of the topography of the collicular projection in the cat has suggested that the rostral SC projects caudally within the IO, while the caudal SC projects rostrally (Kyuhou & Matsuzaki, 1991A). With these points in mind, we set out to investigate the primate tecto-olivary projection by making injections of anterograde tracers into different sectors of the SC using macaque monkeys. METHODS We utilized data from 9 adult and juvenile (2.6–5.0 kg) Macaca fascicularis (7) and Macaca mulata (2) monkeys in this study. Both male (5) and female (4) monkeys were used, but no sex-specific differences were found. All animals were also used in other, non-conflicting studies. The surgical procedures were approved by the Institutional Animal Care and Use Committees of the University of Mississippi Medical Center and the University of Washington. All animal procedures fell within the guidelines put forth by the Guide for Animal Care and Use issued by the USDA. Animals were sedated before surgery through the use of ketamine HCl (10 mg/kg, IM). They were anesthetized during the surgeries with isoflurane (~ 3%). They also received a preemptive analgesic dose of Carprofen (3 mg/kg, IM). They were given dexamethasone (2.5 mg/kg, IV) to avoid edema and atropine sulfate (0.05 mg/kg, IV) to preclude excess airway secretions. Their temperature was regulated during surgery and vital signs were recorded and kept within normal values. At the conclusion of the surgical procedures, wound edges were infused with Sensorcaine and the animals were provided with Buprenex (0.001 mg/kg, IM) to avoid post-surgical discomfort. For stereotaxic injections (n = 7) the animals heads were placed in a head holder (Kopf, Tujunga, CA) and a midline incision was made in the scalp. A craniotomy was placed over the midbrain and the cerebral cortex over the SC was aspirated. For biotinylated dextran amine (BDA) (Molecular Probes; ThermoFisher, Waltham, MA) injections (n = 6), a solution of 10% BDA was held in a Hamilton syringe that was angled 22° tip rostral in the rostrocaudal plane. At 2–3 sites in the SC, tracer (0.1–0.2 μl) was injected 1.5 mm beneath the SC surface. For the Phaseolus vulgaris leukoagglutinin (PhaL) (Vector Laboratories, Newark, CA) injection (n = 1), a glass micropipette with a ~ 35 μm tip held a 2% solution of the tracer in 0.1M, pH 8.0 phosphate buffer (PB). The micropipette was held at a 22° angle, tip rostral in the rostrocaudal plane. The injection was made at a depth of 1.5 mm beneath the SC surface by passing 7 μA positive current with a 50% duty cycle through the tip with a Midguard iontophoresis unit (Harvard Apparatus, Boston, MA). After injection, the aspiration defect was filled with Gelfoam, the scalp was reapproximated and sutured with vicryl. For the physiologically localized injections (n = 2), we implanted each monkey with fixtures to prevent head movements, a scleral search coil (Judge et al., 1980) to measure eye position in space, and a recording chamber that was aimed the SC (see Kojima & May, 2021 for details). After the monkeys had recovered from the surgery, we trained them to track a small visual target in a dimly lit, sound-attenuating booth. To prepare for the injection, we plotted the topographic map of the rostral SC (Robinson, 1972; Sparks and Mays, 1980; Munoz and Wurtz, 1995) by recording unit activity and using electrical stimulation. On the day preceding each injection, we made electrode penetrations into the SC to reveal the optimal vector of the chosen locus (Sparks and Mays, 1980; Munoz and Wurtz, 1995; Soetedjo et al., 2002A; 2002B; Kojima and Soetedjo, 2017) by recording and stimulation (50 μA, 500 Hz, 50 ms trains of 0.1 ms cathodal pulses). On the day of the injection, we advanced the tip of the injectrode (a 35-gauge stainless steel tube which was insulated by epoxylite except for its beveled tip to allow electrical stimulation) towards the same site until we heard neuronal bursts related to pseudo-random (in direction and size) target steps and/or the targeting saccades (Kojima and Soetedjo, 2018). We then stimulated to evoke saccades and took the site’s preferred vector as the average vector of 5 evoked saccades. For one monkey, the stimulation evoked a 4.1° saccade in a 337° direction (right and down). For the other monkey, the stimulation evoked a 1.7° saccade in a 144° direction (left and up). In each animal, we injected 120 nl of 10% BDA by using brief pulses of air pressure (PV830 Pneumatic PicoPump, WPI, Sarasota). Animals survived for 2 (PhaL) to 3 weeks (BDA) following the surgery to allow time for tracer transport. They were once again sedated (ketamine HCl, 10 mg/kg, IM) and then deeply anesthetized (sodium pentobarbital, 50 mg/kg, IP), before being perfused through the heart with a buffered saline rinse followed by a fixative solution containing 1% paraformaldehyde and 1.5% glutaraldehyde in 0.1M, pH 7.2 PB. The brain was blocked in the frontal plane and postfixed for 1 hr in the same fixative solution, then stored in 0.1M, pH7.2 PB at 4° C. Sections were cut at 50 μm (PhaL) or 100 μm (BDA) on a vibratome (Leica VT 1000S, Leica Biosystems, Wetzlar, Germany) and collected serially in 0.1M, pH 7.2 PB. To reveal the BDA, a 1 in 3 series of sections (at a minimum) was incubated overnight at 4° C in an avidin D conjugated horseradish peroxidase (Vector Laboratories, Newark, CA) 1:5000 solution in 0.1M, pH 7.2 PB with 0.05% Triton-X 100. The sections were then reacted with the chromagen diaminobenzidine HCl (DAB) in a 0.1M, pH 7.2 PB 5.0% solution to which 0.05% nickel ammonium sulfate and 0.05% cobalt chloride had been added. The reaction was catalyzed by the addition of 0.011% hydrogen peroxide. To reveal the PhaL, a 1 in 3 series of sections was first pretreated with Triton-X 100 (0.3%) in 0.1M, pH 7.2 PB, then placed in the blocking solution containing 10% normal goat serum. Next, they were placed in a 0.5% solution containing biotinylated goat anti-PhaL (Sigma, St. Louis, MO) overnight at 4°C. The biotinylated anti-PhaL was then tagged using the last solution of an ABC kit (Vector Laboratories, Newark, CA) and the same DAB chromagen procedure as detailed above was used to visualize the reaction product. In all cases, sections were mounted onto gelatinized slides, dried, counterstained with cresyl violet, dehydrated in a graded series of ethanols, cleared in toluene and coverslipped. For illustration of sections, a Wild M8 stereoscope (Leica Biosystems, Wetzlar, Germany) with a drawing tube was employed. The distribution of terminal label was illustrated using an Olympus BH2 microscope (Olympus Life Sciences, Tokyo, Japan) equipped with a drawing tube. Images of the terminal label were captured on a Nikon Eclipse E600 photomicroscope with a Nikon DXM1200F digital camera by use of Nikon Elements software (Nikon, Tokyo, Japan). Images were adjusted in Adobe Photoshop (Adobe, San Jose, CA) for contrast, color and brightness in order to best match the view seen with the eyepieces. RESULTS The cases analyzed included a variety of SC injection sites that varied both in size and location. In the case of the BDA injections, we examined cases where the injection included much of the SC (n = 2), along with injections that were localized in the rostral (n = 2), middle (n = 2) and caudal (n = 2) SC. Among the cases were two in which the injection site was confined to the lateral SC and one in which it was confined to the medial SC. We will not discuss this point further as we were primarily interested in examining the collicular topography with respect to saccade size, which is mapped rostral to caudal in the SC, and because this medial-lateral variable did not seem to be reflected in substantial differences in the IO terminal field. We will first present a case with a large injection, then cases with more circumscribed injections at mid colliculus, rostral colliculus and caudal colliculus levels. The PhaL injection will be used to illustrate transport from middle levels of the SC. In Fig. 1A–C, a large BDA injection of the SC is illustrated. The injection site extends from the rostral midbrain, where the SC abuts the pretectum to the caudal midbrain, where it lies above the inferior colliculus. While the injection site is centered in the intermediate gray layer (SGI), it extends into substantial portions of the superficial gray layer (SGS) and the deep gray layer (SGP). The outer edge of the injection site extended into the adjacent periaqueductal gray, and it also encroached on the posterior pretectal nucleus (not shown). Labeled axon terminal fields were only observed at the caudal end of the IO (Fig. 1D–H). The densest projection was located within the medial subnucleus of the contralateral IO (IOM). The IOM can be divided into four subdivisions, from lateral to medial: A, B, C and β. It should be noted that while various other approaches to subdividing the nucleus exist, we have utilized that of Barmack (2006). The borders between these subdivisions are poorly defined, so we have mainly identified their relative locations. The vast majority of labeled terminal arbors were located within subdivision C, and filled this division. However, some terminal arbors may have extended into adjacent portions of subdivision B and β. Within the terminal field the densest distribution was located at the very caudal end of the IOM (Fig. 1F–H). An easily seen, but sparse distribution of labeled terminal arbors was also present in the caudal end of the contralateral dorsal cap of Kooy (DC) (Fig. 1F–H). Labeled axons could often be seen running between these two terminal fields. Labeled terminal arbors, albeit a much sparser distribution, were also present ipsilaterally in the IOM (Fig. 1E–H). These were confined to the C subdivision of the subnucleus. Just a very few labeled terminals could be observed within the ipsilateral DC (Fig. 1G). The appearance of the BDA labeled terminal arbors in each of these areas can be directly observed in Fig. 2. The terminal field in the contralateral C subdivision from this injection (Fig. 2A) is dense enough to be visible even at low magnification (Fig. 2C). At higher magnification, the BDA labeled terminal arbors fill the neuropil of this part of the IOM, and appear to be more densely clustered in some regions than others (Fig. 2E). The ipsilateral projection is not intense enough to be visible at low magnification (Fig. 2B), but is easily observed at higher magnification in the C subdivision of IOM (Fig. 2D). This large injection produced quite a few labeled axons in DC on the contralateral side (Fig. 2G). These are not just passing axons, as they display numerous boutonal enlargements. Just an occasional labeled axonal arbor was seen in ipsilateral DC (Fig. 2F), but these axons also displayed boutonal enlargements suggestive of synaptic terminals. Figure 3 shows the appearance of labeled axons in IOM drawn at higher magnification. On both the ipsilateral (Fig. 3A–B) and contralateral (Fig. 3C–D) sides, the vast majority of boutons were of the en passant variety, although terminal boutons could occasionally be observed at the end of short branches. Relatively few branch points were noted on the axons, however. Some of the boutons were seen to lie in close association with the somata of the counterstained neurons, but most were located in the neuropil, where they presumably contact the dendrites of the same olivary cells. An injection of PhaL located at mid collicular levels (Fig. 4E) produced a similar pattern of anterograde labeling. Labeled axonal arbors again produced a dense terminal field in the caudal end of the contralateral IOM that was primarily located in the C subdivision of the subnucleus (Fig. 4A-D). More label was found in the B subdivision of this case than observed in the previous case (Fig. 1D–H). Scattered terminal arbors were also present on the ipsilateral side in the C subdivision of IOM (Fig. 4A-C). In addition, PhaL labeled axons were encountered in the DC, both contralaterally (Fig. 4A-D) and ipsilaterally (Fig. 4A-C). The appearance of the PhaL labeled axons in the IOM is shown in Fig. 5. On both the contralateral side (Fig. 5B,D) and ipsilateral side (Fig. 5A,C) fine axons connect numerous labeled puncta that are primarily located within the neuropil. Figure 6E shows a small BDA injection located at a site where electrical stimulation produced a 4.1° saccade to the right and down (at 337°). The pattern of labeling from this injection was similar to that observed with a second case where the injection site produced a 1.7° saccade to the left and up (at 144°). As can be seen here, the BDA injection site was relatively small, and it was confined to SGI at the level shown, with an extension along the micropipette track into superficial gray layer (SGS, not shown). Most of the labeled terminal arbors were found at the caudal end of the IOM on the contralateral side (Fig. 6A–D). The terminal field was densest in the C subdivision of IOM, but extended into both the B and β subdivisions. On the ipsilateral side, sparse labeling of axonal arbors was again found in the C subdivision, but in this rostral injection case more axonal arbors were present in the β subdivision (Fig. 6A–C). Labeled axonal arbors were also present in DC in this case. As with the other cases, more labeled axons were observed contralaterally than ipsilaterally (Fig. 6A–D), but compared to other cases, the rostral injections produced more ipsilateral labeling in the DC. The images shown in Fig. 7 reveal the morphology of the labeling produced by this rostral injection. In the contralateral IOM (Fig. 7A,C,D) the labeled axons are studded with numerous en passant boutons that tend to be arranged in clusters over the regions of the IOM that are more darkly stained by the cresy1 violet, producing a honeycomb like network. The labeled axonal arbors in the ipsilateral IOM show a similar organization, but there are far fewer of them (Fig. 7G,H). Labeled axons passing through the contralateral DC also produced a number of arborizations with numerous puncta suggesting synaptic contacts (Fig. 7A,B). Similar axons were found on the ipsilateral side of the DC (Fig. 7E,F). The excellent level of axonal labeling in this case (Fig. 6–7) allowed detailed drawings of the labeled axons to be made. As shown in Fig. 8D–C, both thicker and thinner labeled axons were present within IOM. The thinner axons exhibited large numbers of en passant boutons and occasional terminal boutons on short branches. They tended to be oriented dorsoventrallly. Some of the boutons on these axons are in close association with olivary neuron somata, but the vast majority are located in the neuropil. On the ipsilateral side (Fig. 8A–B), we were able to follow two long axonal branches that were emitted by a larger parent axon. These extended dorsoventrally for long distances into IOM. Within the nucleus, short collateral branches were present. Boutons were found on both the parent and collateral branches. An example case of the two with injections constrained to the caudal SC is shown in Fig. 9. The injection site was located in the lateral half of the caudal pole of the SC (Fig. 9E). As with the other injections, the main terminal field was located in the caudal end of the contralateral IOM (Fig. 9A–D). Within IOM, most of the labeled arbors were concentrated in the C subdivision, although a few were located in adjacent parts of the B and β subdivision of the subnucleus. A small number of terminal arbors were located in the C subdivision of the ipsilateral IOM (Fig. 9A–D). In this case, labeled axonal arbors were almost exclusively found contralaterally in DC, at its caudal pole (Fig. 9C–D). In the case of the example caudal injection, the labeled terminals were dense enough within the C subdivision of the contralateral IOM that the subdivision was demarcated by the label at low magnification (Fig. 10B). Higher magnification views (Fig. 10D) revealed large numbers of labeled puncta, some in close association with counterstained somata, but most located in the neuropil around the somata. Far fewer labeled axonal arbors were present in the C subdivision of IOM ipsilateral to the injection site (Fig. 10A,C). The DC contralateral to the injection site (Fig. 10B,F) also exhibited a fair number of labeled axons with en passant and terminal puncta, albeit far fewer than were present in the IOM. Only one labeled axon with boutons was observed in the ipsilateral DC, as pictured in Fig. 10A,E. DISCUSSION Comparison of the cases with SC injections of anterograde tracers in this study clearly supports the view that the main target of the tecto-olivary projection is the C subdivision at the caudal end of the contralateral IOM. Variable extension of the terminal field into the medially adjacent β subdivision and laterally adjacent B subdivision of the IOM was observed. In addition to this main projection, a much smaller projection to the same region of the ipsilateral IOM was present. The DC also received a sparse input from the tecto-olivary projection at the same levels. This was primarily a contralateral projection. We did not observe any convincing pattern of topographic projection to the IOM based on comparison of injections located in the rostral, middle and caudal SC. However, projections to the ipsilateral DC appeared to increase with more rostral injections. This lack of topography in the tecto-olivary projection of macaques correlates with the evident lack of topographic organization observed in the complex spike activity of primate vermal Purkinje cells (Kojima, 2019). Neuroanatomical tracers can provide spurious anterograde labeling due to spread outside the target area or uptake by axons passing through the injection site. In the cases we analyzed there was sometimes spread into the periaqueductal gray and pretectum. However, the pattern of termination was largely the same in cases that did not show spread into these structures. To the best of our knowledge, the SC does not generally contain axons that pass through it on the way to the inferior olive. Moreover, the pattern of termination was essentially the same when PhaL was used as a tracer instead of BDA, and PhaL is generally believed to show little axon of passage uptake. Descending axons from the superior colliculus dive directly ventrally, then wrap around the periaqueductal gray on their way to the predorsal bundle (monkey: Moschovakis et al., 1988). Thus, injections at various rostrocaudal levels should not involve axons from neurons located at other levels. In sum, we are confident that the results shown here represent actual labeling from the injection sites. Comparison to Previous Studies The findings presented here are in many ways quite similar to those observed previously. The previous examination of this topic in macaque monkeys showed a terminal field in the same region of the IOM following injection of tritiated amino acids into the middle of the superior colliculus (monkey: Harting, 1977). It should be noted that this paper did not designate a β subdivision, so they describe the terminal field as lying in the B subdivision, instead of the C subdivision. However, their B subdivision is located in the same mediolateral position as the C subdivision of the present study. The lack of reported projections to DC and ipsilateral IOM in the Harting study is most likely due to the limitations of autoradiography for revealing sparse projections. In cats with tritriated amino acid injections of the SC, a terminal field located in the same region of the contralateral IOM was observed (Graham, 1977), with some also observing a sparse projection to the ipsilateral IOM (Weber et al., 1978). This projection and the extent of labeling are very similar to that seen in the present monkey data. Furthermore, that its source lies primarily in the intermediate gray layer of the SC throughout its rostrocaudal extent has been confirmed via retrograde tracing in the cat (Weber et al., 1978; Saint-Cyr & Courville, 1981). The location of the tecto-olivary terminal zone in the rat also appears very similar to that observed for the cat and monkey, although the nucleus containing the terminals is termed the medial accessory olive (Hess, 1982; Akiake, 1992). Only a crossed projection was observed in the rat using tritiated amino acid transport. Thus, the main projection to the contralateral IOM appears to be a conserved mammalian feature. The other projections, ipsilateral to IOM and bilaterally to the DC, may not be found in all mammalian species, or these sparse projections may only be easily observed with more modern tracer techniques. In support of the latter interpretation, a sparse ipsilateral and dense contralateral tecto-olivary projection was observed in hedgehogs by use of BDA (Künzle, 1997), although the much simpler organization of the IO in this species makes more detailed comparison challenging. In bats, however, anterograde transport of wheat germ agglutinin conjugated horseradish peroxidase (WGA-HRP) only revealed a crossed projection to the IO (Covey et al., 1987). To the best of our knowledge, only one previous study has examined the topography of the tecto-olivary projection. In cats, Kyuhou and Matsuzaki (1991) provided convincing anterograde and retrograde evidence using wheat germ agglutinin conjugated horseradish peroxidase (WGA-HRP) that the rostral SC projects more caudally in the contralateral IOM and the caudal SC projects more rostrally in the contralateral IOM. They also suggested that the medial and lateral SC project contralaterally to the medial and lateral IOM, respectively, although this data was less clear cut. While we saw variations in the contralateral IOM projection between cases in the monkey, we did not observe a pattern of readily apparent differences that appeared dependent on injection site location. This may represent a difference between these two species. Alternatively, WGA-HRP, which tends to more specifically label terminals, may provide a clearer picture of their distribution than BDA that labels the entire axonal arbor. Arguing against the later interpretation is the fact that most of the boutons we observed were en passant in nature, so the boutons and arbors should occupy essentially the same territory. To the best of our knowledge, projections by the SC to the DC have not been previously reported. However, in the figures from the Harting (1977) study in macaques labeled axons are plotted immediately surrounding the contralateral DC. This nucleus is generally believed to be targeted by the nucleus of the optic tract, which lies rostrally adjacent to the SC, and this projection plays an important role in optokinetic eye movements (monkey: Hoffman & Distler, 1989; Fuchs et al., 1992; Mustari et al., 1994; Büttner-Ennever et al., 1996). Furthermore, the nucleus of the optic tract projection to the DC is predominantly ipsilateral, in contradistinction to the predominantly contralateral projection we have observed. The ipsilateral projection was more evident with our more rostral injections, and this difference could be due to involvement of the nucleus of the optic tract, but it was still evident with injections of the middle of the SC that did not reach the pretectum. Thus, we believe we have good evidence of a previously unobserved, sparse projection from the SC to the DC. The role of this projection is currently unknown. However, more recent studies have shown that the SC may be involved in eye movements outside of its cardinal role in saccades. For example, the rostral, but not the caudal, SC shows modulation in neuronal activity with respect to pursuit goals (monkey: Hafed & Krauzlis, 2008). In view of the relationship between pursuit and optokinetic eye movements, the projection we have observed, that comes primarily from the more rostral portion of the SC, may make sense. Functional Considerations During saccadic gain adaptation, a signal indicating the fact the target has been missed is relayed to the vermal portion of the cerebellar cortex via the inferior olive. This is observed as a change in the complex spike activity of Purkinje cells (monkey: Soetedjo & Fuchs, 2006). The complex spikes modulate the simple spike activity in a manner that correlates with the needed change in gain to reduce the target error and properly acquire the target (monkey: Kojima et al., 2010A). Furthermore, these simple spike changes in Purkinje cell activity modulate fastigial nucleus activity, which in turn affects the activity of the brainstem premotor neurons that produce saccades (monkey: Optican & Robinson, 1980; Robinson et al., 2002; Kojima et al., 2008; 2014). There is clear evidence from recording, stimulation and inactivation experiments that the source of the target error signal comes from the SC, by way of the IO (monkey: Fens & Van Opstal, 1997; Kaku et al., 2009; Soetedjo et al., 2009; Kojima & Soetedjo, 2017; 2018). In saccade adaptation paradigms, the target error that initiates the gain change is generally less than 5°. In this study, we have provided direct evidence that the rostral SC, where such small target errors are encoded, projects to the IOM. It is of course this exact tectorecipient part of the IO that projects to the oculomotor vermis (rat: Hess, 1982; Akaike, 1992; cat: Kyuhou & Mitsuzaki, 1991). The lack of apparent topography in the macaque tecto-olivary projection is surprising in this regard. Considering that saccade size is encoded by the rostrocaudal location of the locus of burst activity in the SC (monkey: Wurtz & Goldberg, 1972; Sparks, 1975), one would presume that target error size following a target miss should be relayed to the vermis by the IO through topographic means. It is possible that the differences in terminal distribution from different loci within the SC are too subtle to be observed within the tight region that is made up of the C subdivision of IOM. These topographic differences are certainly not readily apparent within the paramedian pontine reticular formation, where the topographic map of saccade vectors must be translated into saccade gain appropriate bursts in premotor neurons (monkey: Moschovakis et al., 1998). In the horizontal gaze center, differences in synaptic weighting encode the size of the saccade, not terminal distribution topography (monkey: Grantyn et al., 2002). It is also possible that the more critical information provided by the tecto-olivary-climbing fiber circuit is the fact that a saccade error has been made. Since all parts of the SC provide input to the IO, and the tecto-olivary projection is a subpart of the general predorsal bundle projection to the brainstem gaze circuitry, it follows that a climbing fiber evoked saccade signal is presented to Purkinje cells before every saccade, regardless of amplitude and direction. What differs about the target error signal for saccade adaptation is the timing; it comes immediately after a saccade has been made, not before, and it does not produce a saccade. (The activity that induces a corrective saccade comes later.) Indeed, it appears that only climbing fiber signals that occur in this window of time after an inaccurate saccade is made, are capable of inducing saccade gain adaptation (monkey: Soetedjo & Fuchs, 2006). Furthermore, any error size will induce adaptation in the correct direction, but not necessarily of the appropriate size (Robinson et al., 1973). On the other hand, adaptation is precisely sensitive to error direction (Kojima & Soetedjo, 2017, Noto et al., 1999). So with respect to amplitude, it is possible that a detailed topographic tecto-olivary projection is not necessary, and that the cerebellum only needs to know whether the saccade was too short or too long (i.e., the direction of the gain error), and does not need to know the precise amplitude of the gain error. Nevertheless, adaptation must ultimately appropriately compensate for the amplitude of the gain error. This may simply occur once the target is achieved by successive small gain changes and the error signal is no longer manufactured. Another factor to consider is that many SC neurons are of the so-called “open field” type, and do not encode saccade amplitude in their firing rates. Their activity decreases to zero for saccades that are smaller than a unit’s preferred amplitude, but it remains relatively constant for larger saccades (monkey: Takeichi et al., 2007). These characteristics are also found in the complex spikes of some Purkinje cells (monkey: Soetedjo et al., 2008), suggesting there are IO neurons with open field firing characteristics. Topographic projections from the SC would not be needed for a circuit connecting open field SC neurons to the cerebellum by way of the IO. However, a second population of neurons with “closed field” responses is also present in the SC. Similar characteristics have been observed in the complex spikes of some Purkinje cells, so presumably IO neurons that exhibit closed field characteristics are present. In closed field cells, activity quickly decreases for saccades that are smaller, larger or oriented in a different direction than a unit’s preferred amplitude and direction (monkey: Soetedjo et al., 2019). As the neural activity of closed field neurons is sensitive to both the direction and amplitude of the error, one might expect that these IO neurons would receive a topographic projection from the SC and convey it to the cerebellum. It is possible that a topographic tecto-olivary projection by SC closed field neurons is obscured by the fact it is overlain by a non-topographic tecto-olivary projection by open field SC neurons. Alternatively, as noted above, it may be that direction and amplitude specificity are conferred upon IO neurons by non-topographic means. In summary, the lack of topography in the macaque tecto-olivary projection may be understandable, albeit unexpected. The presence of an ipsilateral tecto-olivary projection to the C subdivision of the IOM was surprising. It is generally assumed that the two sides of the SC work as antagonists, with one side encoding saccades to the right and the other to the left, and that this dichotomy is enforced by inhibitory tectotectal connections (cat: Behan, 1985; Appel & Behan, 1990; monkey: Olivier et al., 1998; Munoz & Istvan, 1998); however, more recent work suggests the interaction between the two sides of the SC is more complex (monkey: Takahashi et al., 2007; 2010; Takahashi, 2019). Perhaps the presence of this sparse projection from the ipsilateral SC represents the anatomical presentation of this complexity. It should also be noted that a sparse projection to the ipsilateral paramedian pontine reticular formation is also observed following the placement of anterograde tracers in the SC (monkey: Moschovakis et al., 1998). The role of these ipsilateral projections remains to be identified. Acknowledgements: We would like to thank Jinrong Wei who assisted us in surgeries and undertook the histological processing. Grant support: This work was supported by National Eye Institute grant EY014263 from the U.S. National Institutes of Health to Paul J. May & Susan Warren and National Eye Institute grant EY023277 from the U.S. National Institutes of Health to Yoshiko Kojima. This work was also made possible by U.S. National Institutes of Health grants OD010425 for the Washington National Primate Research Center, and P30EY001730 for the Vision Research Core for the University of Washington. Data Sharing: The data supporting this study are available for loan upon reasonable request to the authors. Figure 1 Distribution of BDA labeled tecto-olivary terminal arbors following a large injection in the superior colliculus. A-C: The BDA injection involved all collicular layers and extended from the rostral end to the caudal pole of the colliculus. D-H: Most of the terminal label (stipple) was located in the C subdivision of the four subdivisions (A, B, C & β) that are located within the contralateral medial inferior olive (IOM). Lesser terminal fields were observed in the contralateral dorsal cap of Kooy (DC) and the C subdivision of the ipsilateral medial inferior olive subnucleus. Just a few labeled terminals were present in the ipsilateral DC. Figure Abbreviations: BDA – biotinylated dextran amine, DC – dorsal cap of Kooy, IC – inferior colliculus, IOD – dorsal inferior olive, IOM – medial inferior olive, MdRF – medullary reticular formation, OPt – olivary pretectal nucleus, P – pyramid, PAG – periaqueductal gray, PhaL – Phaseolus vulgarus leukoagglutinin, Pul – pulvinar, SGI – intermediate gray layer, SGS – superficial gray layer Figure 2 Morphology of BDA labeled tecto-olivary terminal arbors in the case charted in figure 1. A. Example section through the BDA injection site in the superior colliculus. B-C. Low magnification views showing the A, B, C & β subdivisions within the ipsilateral (B) and contralateral (C) medial inferior olivary subnucleus (IOM) and the dorsal cap of Kooy (DC). Boxes in B indicate areas shown at higher magnifications in D and F. Boxes in C indicate areas shown at higher magnifications in E and G. BDA labeled axonal arbors (arrows) are most common in the contralateral IOM (E). Fewer labeled arbors are present in the contralateral DC (G) and ipsilateral IOM (D). Only a very few labeled arbors were present in the ipsilateral DC (F). Scale in B = C, F = D, E & G Figure 3 Illustration of BDA labeled axonal arbors in the ipsilateral (A, B) and contralateral (C, D) medial inferior olive (IOM) from the case illustrated in figure 1. Shading indicates counterstained somata Figure 5 Morphology of PhaL labeled tecto-olivary terminal arbors in the case charted in figure 4. A-B. Low magnification views showing the B, C & β subdivisions within the ipsilateral (A) and contralateral (B) medial inferior olivary subnucleus (IOM). Boxes in A and B indicate areas of subnucleus C shown at higher magnifications in C and D, respectively. PhaL labeled axonal arbors (arrows) are most common in the contralateral IOM (D). Fewer labeled arbors are present in the ipsilateral IOM (C). Scale in B = A, D = C Figure 6 Distribution of BDA labeled tecto-olivary terminal arbors following a small, physiologically guided, injection into the rostral superior colliculus. Electrical stimulation at this location produced a 4.1° saccade. E: The BDA injection primarily centered in the intermediate gray layer (SGI). A-D: Most of the terminal label (stipple) was located in the C subdivision of the four subdivisions (A, B, C & β) that are located within the contralateral medial inferior olive (IOM), although it extended into adjacent parts of the B and β subdivisions. A sparser terminal field was observed in the C and β subdivision of the ipsilateral medial inferior olive subnucleus. Labeled arbors were also present in the dorsal cap of Kooy (DC), bilaterally Figure 7 Morphology of BDA labeled tecto-olivary terminal arbors in the case charted in figure 6. A, E & G. Low magnification views showing the A, B, C & β subdivisions within the contralateral (A) and ipsilateral (E,G) medial inferior olivary subnucleus (IOM) and the dorsal cap of Kooy (DC). Boxes in A indicate areas shown at higher magnifications in B and D. Boxes in E and G indicate areas shown at higher magnifications in F and H, respectively. BDA labeled axons (arrows) and clusters of labeled boutons (arrowheads) are most common in the contralateral IOM (D). The boxed area in D is shown at higher magnification in C, where clusters of labeled boutons are very evident. Fewer labeled arbors are present in the contralateral DC (B), ipsilateral IOM (F) and ipsilateral DC (H). Scale in E = A & G; F = B & H Figure 8 Illustration of BDA labeled axonal arbors in the ipsilateral (A, B) and contralateral (C, D) medial inferior olive (IOM) from the small rostral injection case illustrated in figure 6. Shading indicates counterstained somata Figure 9 Distribution of BDA labeled tecto-olivary terminal arbors following a small injection in the caudal superior colliculus. E: The BDA injection involved all collicular layers and was constrained to the caudal pole of the colliculus. A-D: Most of the terminal label (stipple) was located in the C subdivision of the four subdivisions (A, B, C & β) that are located within the contralateral medial inferior olive (IOM). Lesser terminal fields were observed in the contralateral dorsal cap of Kooy (DC) and the C subdivision of the ipsilateral medial inferior olive subnucleus. Just a few labeled terminals were present in the ipsilateral DC (C) Figure 10 Morphology of BDA labeled tecto-olivary terminal arbors in the case charted in figure 9. A-B. Low magnification views showing the A, B, C & β subdivisions within the ipsilateral (A) and contralateral (B) medial inferior olivary subnucleus (IOM), along with the dorsal cap of Kooy (DC). Boxes in A indicate areas shown at higher magnifications in C and E. Boxes in B indicate areas shown at higher magnifications in D and F. BDA labeled terminals heavily infiltrate the contralateral IOM (D). Fewer labeled arbors (arrows) are present in the contralateral DC (F) and ipsilateral IOM (C). The arrow in E points to the one labeled axonal arbor observed in the ipsilateral DC. Scale in A = B; F = C-E Conflict of Interest: None of the authors has any conflicts of interest, financial or otherwise, with respect to the work described in this manuscript. Statements & Declarations Ethical Use of Animals Statement: All applicable international, national, and/or institutional guidelines for the care and use of animals were followed. All procedures performed in studies involving animals were in accordance with the ethical standards of the institution at which the studies were conducted. Specifically, they were undertaken under protocols approved by the Institutional Animal Care and Use Committee of the University of Mississippi Medical Center (USDA Animal Welfare Assurance # D16-00174) and the University of Washington (USDA Animal Welfare Assurance # D16-00292). ==== Refs References 1. Akaike T (1992) The tectorecipient zone in the inferior olivary nucleus in the rat. J Comp Neurol 320 :398–414. doi: 10.1002/cne.903200311.1377203 2. Appell PP , Behan M (1990) Sources of subcortical GABAergic projections to the superior colliculus in the cat. J Comp Neurol 302 :143–58. doi: 10.1002/cne.903020111.2086611 3. Barmack NH (2006) Inferior olive and oculomotor system. Prog Brain Res 151 :269–91. doi: 10.1016/S0079-6123(05)51009-4.16221592 4. Behan M (1985) An EM-autoradiographic and EM-HRP study of the commissural projection of the superior colliculus in the cat. J Comp Neurol 234 :105–16. doi: 10.1002/cne.902340108.3980783 5. Büttner-Ennever JA , Cohen B , Horn AK , Reisine H (1996) Efferent pathways of the nucleus of the optic tract in monkey and their role in eye movements. J Comp Neurol 373 :90–107. doi: 10.1002/(SICI)1096-9861(19960909)373:1<90::AID-CNE8>3.0.CO;2–8. 8876465 6. Covey E , Hall WC , Kobler JB (1987) Subcortical connections of the superior colliculus in the mustache bat, Pteronotus parnellii. J Comp Neurol 263 :179–97. doi: 10.1002/cne.902630203.3667975 7. Edelman JA , Goldberg ME (2002) Effect of short-term saccadic adaptation on saccades evoked by electrical stimulation in the primate superior colliculus. J Neurophysiol 87 :1915–23. doi: 10.1152/jn.00805.2000.11929911 8. Frens MA , Van Opstal AJ (1997) Monkey superior colliculus activity during short-term saccadic adaptation. Brain Res Bull 43 :473–83. doi: 10.1016/s0361-9230(97)80001-9.9250621 9. Fuchs AF , Mustari MJ , Robinson FR , Kaneko CR (1992) Visual signals in the nucleus of the optic tract and their brain stem destinations. Ann N Y Acad Sci 656 :266–76. doi: 10.1111/j.1749-6632.1992.tb25214.x.1599148 10. Graham J (1977) An autoradiographic study of the efferent connections of the superior colliculus in the cat. J Comp Neurol 173 :629–54. doi: 10.1002/cne.901730403.864027 11. Grantyn A , Brandi A-M , Dubayle D , Graf W , Ugolini G , Hadjidimitrakis K , Moschovakis A (2002) Density gradients of trans-synaptically labeled collicular neurons after injections of rabies virus in the lateral rectus muscle of the rhesus monkey. J Comp Neurol 12. 451 :346–61. doi: 10.1002/cne.10353. 13. Hafed ZM , Krauzlis RJ (2008) Goal representations dominate superior colliculus activity during extrafoveal tracking. J Neurosci 28 :9426–39. doi: 10.1523/JNEUROSCI.1313-08.2008.18799675 14. Harting JK (1977) Descending pathways from the superior collicullus: an autoradiographic analysis in the rhesus monkey (Macaca mulatta). J Comp Neurol 173 :583–612. doi: 10.1002/cne.901730311.404340 15. Hess DT (1982) The tecto-olivo-cerebellar pathway in the rat. Brain Res 250 :143–8. doi: 10.1016/0006-8993(82)90960-x.6182946 16. Hoffmann KP , Distler C (1989) Quantitative analysis of visual receptive fields of neurons in nucleus of the optic tract and dorsal terminal nucleus of the accessory optic tract in macaque monkey. J Neurophysiol 62 :416–28. doi: 10.1152/jn.1989.62.2.416.2769338 17. Judge SJ , Richmond BJ , Chu FC (1980) Implantation of magnetic search coils for measurement of eye position: an improved method. Vision Res 20 :535–8. doi: 10.1016/0042-6989(80)90128-5 6776685 18. Kaku Y , Yoshida K , Iwamoto Y (2009) Learning signals from the superior colliculus for adaptation of saccadic eye movements in the monkey. J Neurosci 29 :5266–75. doi: 10.1523/JNEUROSCI.0661-09.2009.19386923 19. Kojima Yoshiko (2019) A neuronal process for adaptive control of primate saccadic system. Prog Brain Res 249 :169–181. doi: 10.1016/bs.pbr.2019.03.029.31325976 20. Kojima Y , May PJ (2021) The substantia nigra pars reticulata modulates error-based saccadic learning in monkeys. eNeuro 8:ENEURO.0519-20.2021. doi: 10.1523/ENEURO.0519-20.2021. 21. Kojima Y , Soetedjo R (2017) Change in sensitivity to visual error in superior colliculus during saccade adaptation. Sci Rep 7 :9566. doi: 10.1038/s41598-017-10242-z.28852092 22. Kojima Y , Soetedjo R (2018) Elimination of the error signal in the superior colliculus impairs saccade motor learning. Proc Natl Acad Sci U S A 115 :E8987–E8995. doi: 10.1073/pnas.1806215115.30185563 23. Kojima Y , Iwamoto Y , Robinson FR , Noto CT , Yoshida K (2008) Premotor inhibitory neurons carry signals related to saccade adaptation in the monkey. J Neurophysiol 99 :220–30. doi: 10.1152/jn.00554.2007.17977929 24. Kojima Y , Soetedjo R , Fuchs AF (2010A) Changes in simple spike activity of some Purkinje cells in the oculomotor vermis during saccade adaptation are appropriate to participate in motor learning. J Neurosci 30 :3715–27. doi: 10.1523/JNEUROSCI.4953-09.2010.20220005 25. Kojima Y , Soetedjo R , Fuchs AF (2010B) Effects of GABA agonist and antagonist injections into the oculomotor vermis on horizontal saccades. Brain Res 1366 :93–100. doi: 10.1016/j.brainres.2010.10.027.20951682 26. Kojima Y , Soetedjo R , Fuchs AF (2011) Effect of inactivation and disinhibition of the oculomotor vermis on saccade adaptation. Brain Res 1401 :30–9. doi: 10.1016/j.brainres.2011.05.027.21679930 27. Kojima Y , Robinson FR , Soetedjo R (2014) Cerebellar fastigial nucleus influence on ipsilateral abducens activity during saccades. J Neurophysiol 111 :1553–63. doi: 10.1152/jn.00567.2013.24478158 28. Künzle H (1997) Connections of the superior colliculus with the tegmentum and the cerebellum in the hedgehog tenrec. Neurosci Res 28 :127–45. doi: 10.1016/s0168-0102(97)00034-5.9220470 29. Kyuhou SI , Matsuzaki R (1991A) Topographical organization of the tecto-olivo-cerebellar projection in the cat. Neuroscience 41 :227–41. doi: 10.1016/0306-4522(91)90212-7.1711650 30. Kyuhou SI , Matsuzaki R (1991B) Topographical organization of climbing fiber pathway from the superior colliculus to cerebellar vermal lobules VI-VII in the cat. Neuroscience 45 :691–9. doi: 10.1016/0306-4522(91)90281-r.1775242 31. May PJ (2006) The mammalian superior colliculus: laminar structure and connections. Prog Brain Res 151 :321–78. doi: 10.1016/S0079-6123(05)51011-2.16221594 32. May PJ , Porter JD (1992) The laminar distribution of macaque tectobulbar and tectospinal neurons. Vis Neurosci 8 :257–76. doi: 10.1017/s0952523800002911.1372175 33. Melis BJ , van Gisbergen JA (1996) Short-term adaptation of electrically induced saccades in monkey superior colliculus. J Neurophysiol 76 :1744–58. doi: 10.1152/jn.1996.76.3.1744.8890289 34. Moschovakis AK , Karabelas AB , Highstein SM (1988) Structure-function relationships in the primate superior colliculus. I. Morphological classification of efferent neurons. J Neurophysiol 60 :232–62. doi: 10.1152/jn.1988.60.1.232.3404219 35. Moschovakis AK , Kitama T , Dalezios Y , Petit J , Brandi AM , Grantyn AA (1998) An anatomical substrate for the spatiotemporal transformation. J Neurosci 18 :10219–29. doi: 10.1523/JNEUROSCI.18-23-10219.1998.9822775 36. Munoz DP , Istvan PJ (1998) Lateral inhibitory interactions in the intermediate layers of the monkey superior colliculus. J Neurophysiol 79 :1193–209. doi: 10.1152/jn.1998.79.3.1193.9497401 37. Munoz DP , Wurtz RH (1995) Saccade-related activity in monkey superior colliculus. I. Characteristics of burst and buildup cells. J Neurophysiol 73 :2313–33. doi: 10.1152/jn.1995.73.6.2313.7666141 38. Mustari MJ , Fuchs AF , Kaneko CR , Robinson FR (1994) Anatomical connections of the primate pretectal nucleus of the optic tract. J Comp Neurol 349 :111–28. doi: 10.1002/cne.903490108.7852621 39. Noda H , Fujikado T (1987) Topography of the oculomotor area of the cerebellar vermis in macaques as determined by microstimulation. J Neurophysiol 58 :359–78. doi: 10.1152/jn.1987.58.2.359.3655873 40. Noto CT , Watanabe S , Fuchs AF (1999) Characteristics of simian adaptation fields produced by behavioral changes in saccade size and direction. J Neurophysiol 81 :2798–813. doi: 10.1152/jn.1999.81.6.2798.10368398 41. Olivier E , Porter JD , May PJ (1998) Comparison of the distribution and somatodendritic morphology of tectotectal neurons in the cat and monkey. Vis Neurosci 15 :903–22. doi: 10.1017/s095252389815513x.9764533 42. Optican LM , Robinson DA (1980) Cerebellar-dependent adaptive control of primate saccadic system. J Neurophysiol 44 :1058–76. doi: 10.1152/jn.1980.44.6.1058.7452323 43. Robinson DA (1972) Eye movements evoked by collicular stimulation in the alert monkey. Vision Res 12 :1795–808. doi: 10.1016/0042-6989(72)90070-3.4627952 44. Robinson FR , Fuchs AF , Noto CT (2002) Cerebellar influences on saccade plasticity. Ann N Y Acad Sci 956 :155–63. doi: 10.1111/j.1749-6632.2002.tb02816.x.11960801 45. Robinson FR , Noto CT , Bevans SE (2003) Effect of visual error size on saccade adaptation in monkey. J Neurophysiol 90 :1235–44. doi: 10.1152/jn.00656.2002. Epub 2003 Apr 23. 12711711 46. Robinson FR , Soetedjo R , Noto C (2006) Distinct short-term and long-term adaptation to reduce saccade size in monkey. J Neurophysiol96:1030–41. doi: 10.1152/jn.01151.2005. 47. Ron S , Robinson DA (1973) Eye movements evoked by cerebellar stimulation in the alert monkey. J Neurophysiol 36 :1004–22. doi: 10.1152/jn.1973.36.6.1004.4202613 48. Saint-Cyr JA , Courville J (1981) Sources of descending afferents to the inferior olive from the upper brain stem in the cat as revealed by the retrograde transport of horseradish peroxidase. J Comp Neurol 198 :567–81. doi: 10.1002/cne.901980403.7251931 49. Soetedjo R , Fuchs AF (2006) Complex spike activity of purkinje cells in the oculomotor vermis during behavioral adaptation of monkey saccades. J Neurosci 26 :7741–55. doi: 10.1523/JNEUROSCI.4658-05.2006.16855102 50. Soetedjo R , Kaneko CR , Fuchs AF (2002A) Evidence against a moving hill in the superior colliculus during saccadic eye movements in the monkey. J Neurophysiol 87 :2778–89. doi: 10.1152/jn.2002.87.6.2778.12037180 51. Soetedjo R , Kaneko CR , Fuchs AF (2002B) Evidence that the superior colliculus participates in the feedback control of saccadic eye movements. J Neurophysiol 87 :679–95. doi: 10.1152/jn.00886.2000.11826037 52. Soetedjo R , Kojima Y , Fuchs A (2008) Complex spike activity signals the direction and size of dysmetric saccade errors. Prog Brain Res 171 :153–9. doi: 10.1016/S0079-6123(08)00620-1.18718294 53. Soetedjo R , Fuchs AF , Kojima Y (2009) Subthreshold activation of the superior colliculus drives saccade motor learning. J Neurosci 29 :15213–22. doi: 10.1523/JNEUROSCI.4296-09.2009.19955374 54. Soetedjo R , Kojima Y , Fuchs A (2019) How cerebellar motor learning keeps saccades accurate. J Neurophysiol 121 :2153–2162. doi: 10.1152/jn.00781.2018.30995136 55. Sparks DL (1975) Response properties of eye movement-related neurons in the monkey superior colliculus. Brain Res 90 :147–52. doi: 10.1016/0006-8993(75)90690-3.1131686 56. Sparks DL Mays LE (1980) Movement fields of saccade-related burst neurons in the monkey superior colliculus. Brain Res 190 :39–50. doi: 10.1016/0006-8993(80)91158-0.6769538 57. Straube A , Fuchs AF , Usher S , Robinson FR (1997) Characteristics of saccadic gain adaptation in rhesus macaques. J Neurophysiol 77 :874–95. doi: 10.1152/jn.1997.77.2.874.9065856 58. Takahashi M (2019) Morphological and electrophysiological characteristics of the commissural system in the superior colliculi for control of eye movements. Prog Brain Res 249 :105–115. doi: 10.1016/bs.pbr.2019.04.027.31325971 59. Takahashi M , Sugiuchi Y , Shinoda Y (2007) Commissural mirror-symmetric excitation and reciprocal inhibition between the two superior colliculi and their roles in vertical and horizontal eye movements. J Neurophysiol 98 :2664–82. doi: 10.1152/jn.00696.2007.17728384 60. Takahashi M , Sugiuchi Y , Shinoda Y (2010) Topographic organization of excitatory and inhibitory commissural connections in the superior colliculi and their functional roles in saccade generation. J Neurophysiol 104 :3146–67. doi: 10.1152/jn.00554.2010.20926614 61. Takeichi N , Kaneko CR , Fuchs AF (2007) Activity changes in monkey superior colliculus during saccade adaptation. J Neurophysiol 97 :4096–107. doi: 10.1152/jn.01278.2006.17442764 62. Wallman J , Fuchs AF (1998) Saccadic gain modification: visual error drives motor adaptation. J Neurophysiol 80 :2405–16. doi: 10.1152/jn.1998.80.5.2405.9819252 63. Weber JT , Partlow GD , Harting JK (1978) The projection of the superior colliculus upon the inferior olivary complex of the cat: an autoradiographic and horseradish peroxidase study. Brain Res 144 :369–77. doi: 10.1016/0006-8993(78)90163-4.346175 64. Wurtz RH , Goldberg ME (1972) Activity of superior colliculus in behaving monkey. 3. Cells discharging before eye movements. J Neurophysiol 35 :575–86. doi: 10.1152/jn.1972.35.4.575.4624741

PMC010xxxxxx/PMC10312990.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398238 10.21203/rs.3.rs-3053775/v1 10.21203/rs.3.rs-3053775 preprint 1 Article Literacy, but not memory, is associated with hippocampal connectivity in illiterate adults de Paula França Resende Elisa Universidade Federal de Minas Gerais Lara Vivian P. Faculdade de Ciências Médicas de Minas Gerais Santiago Ana Luisa C. Universidade Federal de Minas Gerais Friedlaender Clarisse V. Universidade Federal de Minas Gerais Rosen Howard J. University of California, San Francisco Brown Jesse A. University of California, San Francisco Cobigo Yann University of California, San Francisco Silva Lênio L. G. Axial Inteligência Diagnóstica de Souza Leonardo Cruz Universidade Federal de Minas Gerais Rincon Luciana Universidade Federal de Minas Gerais Grinberg Lea T. University of California, San Francisco Maciel Francisca I.P. Universidade Federal de Minas Gerais Caramelli Paulo Universidade Federal de Minas Gerais Author’s Contributions EPFR: Conceived the study, contributed to the study design, supervised data collection and analysis, interpreted the results, and wrote the manuscript, VLP, ALCS: Participated in data collection and contributed to data analysis, CVF: Contributed to the study design, and assisted in data collection, HJR: Contributed to the study design, and critically revised the manuscript, JAB: Assisted in neuroimaging data analysis, contributed to the interpretation of results, and provided critical feedback on the manuscript, YC: Assisted in neuroimaging data preprocessing, analysis and interpretation. LLG: Assisted with neuroimaging acquisition and preprocessing the data, LCS: Assisted in data interpretation and contributed to the critical discussion of results, and reviewed the manuscript, LR: Assisted in data collection and interpretation, LTG: contributed to the study design, supervised data collection and analysis, interpreted the results, and critically revised the manuscript, FIPM: Assisted in data collection, contributed to the literature review, and provided critical revisions to the manuscript, PC: Conceived and designed the study, provided overall supervision, contributed to data interpretation, and critically revised the manuscript. All authors have read and approved the final version of the manuscript. ✉ elisa.resende@gbhi.org 16 6 2023 rs.3.rs-3053775https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs3053775v1.pdf Background The influence of hippocampal connectivity on memory performance is well established in individuals with high educational attainment. However, the role of hippocampal connectivity in illiterate populations remains poorly understood. Methods Thirty-five illiterate adults were administered a literacy assessment (Test of Functional Health Literacy in Adults - TOFHLA), structural and resting state functional MRI and an episodic memory test (Free and Cued Selective Reminding Test). Illiteracy was defined as a TOFHLA score below 53. We evaluated the correlation between hippocampal connectivity at rest and both free recall and literacy scores. Results Participants were mostly female (57.1%) and Black (84.8%), with a median age of 50 years. The median TOFHLA literacy score was 28.0 [21.0;42.5] out of 100 points and the median free recall score was 30.0 [26.2;35] out of 48 points. The median gray matter volume of both the left and right hippocampi was 2.3 [2.1; 2.4] cm3. We observed a significant connectivity between both hippocampi and the precuneus and the ventral medial prefrontal cortex. Interestingly, the right hippocampal connectivity positively correlated with the literacy scores (β = 0.58, p = 0.008). There was no significant association between episodic memory and hippocampal connectivity. Neither memory nor literacy scores correlated with hippocampal gray matter volume. Conclusions Low literacy levels correlate with hippocampal connectivity in illiterate adults. The lack of association with memory scores might be associated with low brain reserve in illiterate adults. illiteracy cognitive reserve hippocampal connectivity episodic memory Alzheimer’s AssociationGBHI_ALZ-18-534892 World Federation of NeurologyNational Institute on Aging of the National Institutes of HealthR21AG0692502 ==== Body pmcIntroduction Life expectancy is increasing in low- and middle-income countries (LMIC), and consequently, the prevalence of dementia is rapidly rising. Between 2015 and 2050, the prevalence of dementia is expected to increase by 138% in countries like Brazil, compared to an increase of 56% in high-income countries (HIC) (Prince et al., 2015). Whereas most causes of dementia have no curative treatments so far, disease-modifying drugs for Alzheimer’s disease have a high cost and a controversial efficacy (Cummings et al., 2022). Moreover, co-pathologies in dementia are very common (Suemoto et al., 2017). Therefore, preventing dementia is a powerful strategy to mitigate the high burden of the disease on patients, caregivers, and society. Research shows that in LMIC 48% of dementia cases could be prevented if 12 modifiable dementia-risk factors were controlled (Suemoto et al., 2022). Low educational level ranks highest among these factors. The prevalence and incidence of dementia in illiterate older adults is two and five times higher than in literate adults (César-Freitas et al., 2022; Nitrini et al., 2009; Ribeiro et al., 2022), respectively. Up to 19% of dementia cases can be attributed to low educational attainment in HIC and up to 30% in LMIC, where low education is more prevalent (Mukadam et al., 2019; Norton et al., 2014). An increase in educational attainment in HIC is believed to have contributed to the recently observed decline in dementia incidence(Wu et al., 2017). In the Framingham study, the incidence of dementia is declining only among persons who have at least a high-school degree (Satizabal et al., 2016). Although other socioeconomic determinants of health associated with high education may play a role in the apparent protective trends, several studies support education as an independent factor leading to lower risk of dementia (Lu et al., 2019; Sharp & Gatz, 2011; Wang et al., 2012). Understanding how the illiterate brain process verbal and non-verbal cognitive tasks can help develop tailored strategies towards improving literacy skills to increase one’s memory abilities to mitigate the symptoms of dementia. Brain activation regarding letters and face recognition have different patterns in literate vs. illiterate individuals (Dehaene & Cohen, 2007; Dehaene et al., 2015; Dehaene et al., 2010). Additionally, regions of the brain involved in language processing have better white matter integrity in literate individuals compared to illiterates (Resende et al., 2017; Thiebaut de Schotten et al., 2014). However, it is unknown whether episodic memory correlates with hippocampal volumes and connectivity in illiterate individuals, classical neural substrates related to episodic memory in persons with high educational level. Previous research in low literate older adults (mean of four years of formal education) showed that episodic memory correlated with the integrity of white matter bundles that connect the hippocampus with the precuneus and with hippocampal volume (Resende et al., 2017). However, this correlation was significant only amongst the group with more than four years of formal education (Resende, Rosen, et al., 2018). The role of hippocampal connectivity on episodic memory performance in adults is still a matter of debate (Aggleton & Brown, 1999; Bhattacharyya, 2017; Eichenbaum, 2004; Sidhu et al., 2013; van Kesteren et al., 2010) and there are no studies in illiterate adults. The default mode network is important for memory processing (Staffaroni et al., 2018) and it is affected in patients with dementia of the amnestic type (Malotaux et al., 2022; Seeley et al., 2009; Zhou et al., 2010). In the present study, we used structural and functional MRI to define whether there was an association between episodic memory and hippocampal volumes and functional connectivity in illiterates. Understanding the brain mechanisms involved in episodic memory processing in illiterates can help unveil possible markers of successful interventions to improve memory in these populations, to mitigate the memory problems caused by neurodegenerative process that comes with aging. Methods Population We used a community-based participatory research approach to collaborate with a basic-literacy training program for adults that is sponsored by the government. This program targets illiterate adults that did not have the opportunity to go to school when they were at the school age and want to learn how to read and write later in life. Adults aged 40 to 80 years-old that spontaneously enrolled in those late-life educational programs in the city of Belo Horizonte, Brazil, from February to July 2019 were invited to participate in the present research. Forty-three persons signed the informed consent and agreed to participate in the research. Sociodemographic and smoking habits were collected through a structured questionnaire. The level of physical activity was assessed with the Baecke scale (Baecke et al., 1982; Rocha et al., 1992). Depression, anxiety and alcohol abuse were investigated by the Mini International Neuropsychiatric Interview (Sheehan et al., 1998). All evaluations were conducted upon entry in the late-life literacy program before any literacy training. The socioeconomic levels were determined using the ABEPE (Brazilian Association of Research Companies) framework that categorize households into different socioeconomic levels. This classification considers various factors such as income, education, and ownership of goods to determine the living standards of households. The level A category represents the highest socioeconomic level with high income levels, advanced education, and ownership multiple properties and luxury goods. The level B category includes households with a relatively high socioeconomic status, although slightly lower than those in level A. These households generally have good incomes, tertiary education, and own properties and durable goods. The level C category encompasses households with a middle socioeconomic status. They usually have moderate incomes, secondary education, and may own a house or apartment. The levels D and E represent households with a lower socioeconomic status that often have low incomes, limited education, and may live in rented accommodations or informal settlements. They may face significant economic challenges and lack of access to basic services. They often live in poverty, struggling to meet their basic needs and relying on government assistance programs. Literacy and cognitive assessment Participants that enroll in those late-life government sponsored programs have various degrees of reading and writing skills. Some never attended formal school while others attended for few years. Their reading abilities vary from inability to recognize letters to some reading capacity, without comprehending the meaning of the text. Therefore, we used the Test of Functional Health Literacy in Adults (TOFHLA) (Parker et al., 1995), validated for Brazilian Portuguese (Maragno et al., 2019), to evaluate the participant’s literacy skills across different levels. Previous studies determined that a score equal or lower than 53 defines illiteracy (Apolinario et al., 2015). Global cognition was assessed by the Mini Mental State Examination (Brucki et al., 2003; Folstein et al., 1975) Episodic memory was assessed with the visual form (pictures) of the Free and Cued Selective Reminding test (FCSRT) (Grober et al., 2010; Zibetti et al., 2014). The FCSR-IR Free Recall sum-of-attempts was considered the proxy for episodic memory. Non-verbal intelligence was assessed by the Beta-3 test (Rabelo & Pacanaro, 2011), attention with the digit span test (de Paula et al., 2013), reading abilities with the Human Frontier Science Program reading test (Martins et al., 2023), words and sentence repetition with the Boston Diagnostic Aphasia Examination (Goodglass & Kaplan, 1983; Miotto et al., 2010) and verbal comprehension with the Token test (de Paula et al., 2013). Finally, participants performed the rapid naming of colors, letters, numbers, and objects(da Silva et al., 2020) and the Ekman’s facial emotion recognition test (Passarelli et al., 2018). Global cognitive reserve was assessed with a structured questionnaire available in Portuguese, that includes years of education, leisure activities and occupational attainment (Nucci et al., 2012). Neuroimaging acquisition and analysis Brain MRIs were acquired in a 3 Tesla Siemens Verio scanner with 3D-T1 and resting-state functional MRI (rsfMRI) acquisitions. The acquisition parameters were as follows. For 3D-T1: Field of View of 208×240×256 mm at reconstructed resolution of 1×1×1 mm, TE = min full echo, TR 2300 ms, TI 900 ms. For rsfMRI: Voxel resolution 2×2×2mm, Field of View of 220×220×163 mm, TE = 30 ms, TR 3000 ms, FA = 90°, time for acquisition 10 minutes. Before any prepossessing of the images, all T1-weighted images were visually inspected for quality control. One image was excluded because of a large artifact. T1-weighted images undergone bias field correction using N3 algorithm, the segmentation was performed using SPM12 unified segmentation (Ashburner & Friston, 2005). A customized group template was generated from the segmented gray and white matter tissues and cerebrospinal fluid (CSF) by non-linear registration template generation using Large Deformation Diffeomorphic Metric Mapping framework (Ashburner & Friston, 2011). Native subjects’ space gray and white matter were geometrically normalized to the group template, modulated, and then smoothed in the group template. The applied smoothing used a Gaussian kernel with 8 ~ mm full width half maximum. Every step of the transformation was carefully inspected from the native space to the group template. For statistical purposes, linear and non-linear transformations between the group template space and International Consortium of Brain Mapping (ICBM) (Mazziotta et al., 1995) were applied. The Harvard-Oxford atlas (Desikan et al., 2006) was used to calculate the hippocampal volumes for each participant. The rsfMRI analyses were done using the CONN (Whitfield-Gabrieli & Nieto-Castanon, 2012) release 20.b toolbox and SPM12 (Penny et al., 2011). First, functional and anatomical data were preprocessed using a flexible preprocessing pipeline (Nieto-Castanon, 2020) including realignment with correction of susceptibility distortion interactions, slice timing correction, outlier detection, direct segmentation and MNI-space normalization, smoothing, and band-pass filtering. Functional data were realigned using SPM realign & unwarp procedure (Andersson et al., 2001), where all scans were coregistered to a reference image (first scan of the first session) using a least squares approach and a 6 parameter (rigid body) transformation (Friston et al., 1995), and resampled using b-spline interpolation to correct for motion and magnetic susceptibility interactions. Temporal misalignment between different slices of the functional data (acquired in interleaved Siemens order) was corrected following SPM slice-timing correction procedure (Henson et al., 1999; Sladky et al., 2011), using sinc temporal interpolation to resample each slice BOLD timeseries to a common mid-acquisition time. Potential outlier scans were identified using ART (Whitfield-Gabrieli, 2009) as acquisitions with framewise displacement above 0.9 mm or global BOLD signal changes above 5 standard deviations (Power et al., 2014). A reference BOLD image was computed for each subject by averaging all scans excluding outliers. Functional and anatomical data were normalized into standard MNI space, segmented into grey matter, white matter, and CSF tissue classes, and resampled to 2 mm isotropic voxels following a direct normalization procedure (Calhoun et al., 2017) using SPM unified segmentation and normalization algorithm (Ashburner & Friston, 2005) with the default IXI-549 tissue probability map template. Functional data were smoothed using spatial convolution with a Gaussian kernel of 8 mm full width half maximum. Last, BOLD signal timeseries were bandpass filtered between 0.01 Hz and 0.1 Hz. In addition, functional data were denoised using a standard denoising pipeline(Friston et al., 1996) including the regression of potential confounding effects characterized by white matter timeseries (5 CompCor noise components), CSF timeseries (5 CompCor noise components), motion parameters and their first order derivatives (12 factors) (Friston et al., 1996), outlier scans (below 13 factors) (Power et al., 2014), session effects and their first order derivatives (2 factors), and linear trends (2 factors) within each functional run, followed by bandpass frequency filtering of the BOLD timeseries (Hallquist et al., 2013) between 0.008 Hz and 0.09 Hz. CompCor stands for Component-based noise correction method (Behzadi et al., 2007) that computes the average BOLD signal as well as the largest principal components orthogonal to the BOLD average, motion parameters, and outlier scans within each subject’s eroded segmentation masks. Those CompCor noise were estimated within the white matter and CSF. Seed-based connectivity maps and ROI-to-ROI connectivity matrices were estimated characterizing the patterns of functional connectivity with 164 HPC-ICA networks (Whitfield-Gabrieli & Nieto-Castanon, 2012) and Harvard-Oxford atlas ROIs (Desikan et al., 2006). Functional connectivity strength was represented by Fisher-transformed bivariate correlation coefficients from a weighted general linear model (weighted-GLM (Nieto-Castanon, 2020)), defined separately for each pair of seed and target areas, modeling the association between their BOLD signal timeseries. To compensate for possible transient magnetization effects at the beginning of each run, individual scans were weighted by a step function convolved with an SPM canonical hemodynamic response function and rectified. The seed-based connectivity analyses were done placing a seed in each hippocampi using the Harvard-Oxford automated atlas (Desikan et al., 2006). The ROI-to-ROI connectivity matrices analyzed were the ones between each hippocampus and the ventral medial pre-frontal (VMPFC), each hippocampus (HC) and the Precuneus (PCC) and between the VMPFC and PCC. Finally, the group-level analyses were performed using a GLM. For each individual voxel a separate GLM was estimated, with first-level connectivity measures at this voxel as dependent variables (one independent sample per subject), and groups as independent variables. Voxel-level hypotheses were evaluated using multivariate parametric statistics with random-effects across subjects and sample covariance estimation across multiple measurements. Inferences were performed at the level of individual clusters (groups of contiguous voxels). Cluster-level inferences were based on parametric statistics from Gaussian Random Field theory (Worsley et al., 1996). Results were thresholded using a combination of a cluster-forming p < 0.001 voxel-level threshold, and a familywise corrected p-FDR < 0.05 cluster-size threshold (Chumbley et al., 2010) Demeaned age was used as a covariate in all neuroimaging analyses. Statistical analyses Continuous variables were depicted in median and interquartile intervals; categorical variables were depicted in frequencies. GLM considering age, sex and total intracranial volume as covariates were used to calculate the correlation between episodic memory, literacy levels, brain connectivity and hippocampal volumes. In the first model, the FCSRT free-recall sum of attempts was the dependent variable, and the predictors were the functional connectivity between each HC separately and the VMPFC, between each HC and precuneus, and between the VMPFC and PCC, as well as with each hippocampal volume. In the second model, the literacy level measured by the TOFHLA total score was the dependent variable and the predictors were the same depicted above. Results The final sample had 35 participants. We excluded three participants that had claustrophobia and did not tolerate the brain MRI, one participant whose scan had artifacts that precluded the analysis, three that were left-handed and one that scored 98 in the TOFLHA and was, therefore considered literate. The median age was 50 years, 57.1% (n = 20) of participants were women and 84.8% (n = 28) were Blacks (Table 1). The median TOFHLA score was 28 with an interquartile interval of 21.0 to 42.5. The seed-based connectivity analysis at rest showed a significant connectivity between both HC and the VMPFC and PCC, and other brain regions (Fig. 1). However, we failed to find a significant association between the HC-VMPFC connectivity and episodic memory measured by the FCSRT free recall sum of attempts (Table 2). On the other hand, we found significant associations between the low TOFHLA scores and the HC-VMPFC connectivity (Table 3). Interestingly, the association was in opposite directions in each hippocampus. On the right side, the stronger the HC-VMPFC connectivity, the better the literacy scores (β = 0.58, p = 0.004), whilst on the left side, the stronger the connectivity, the worse the literacy scores (β=−0.39, p = 0.041). Age and sex did not significantly correlate with the association between HC connectivity and memory or literacy scores. Discussion In a group of middle-aged adults, the performance on a literacy test, even low enough to be considered illiterate per the literature (Apolinario et al., 2015), correlated with the HC-VMPFC connectivity. The association between low literacy levels and HC-VMPFC may suggest the role of even some literacy on cognitive reserve mechanisms. In contrast, we can speculate that the lack of association between episodic memory performance and hippocampal connectivity might reflect that this reserve is not enough to strengthen the role of hippocampal connectivity in memory abilities. Cognitive reserve refers to distinct cognitive mechanisms, developed across the lifespan, that make a person more resilient or resistant to cognitive decline caused by brain damage (Stern et al., 2023). A higher level of cognitive reserve equips the brain to compensate through more efficient brain activation patterns that are more flexible and resilient to neurodegeneration or other forms of brain injury (Stern et al., 2023). Because higher cognitive reserve is associated with more tolerance to hippocampal atrophy (Murray et al., 2011), neurodegeneration (Hoenig et al., 2017; Wirth et al., 2014), and cerebrovascular (Fernandez-Cabello et al., 2016) pathologies, we believe that improving literacy levels might increase the HC-VMPFC connectivity and eventually prevent cognitive impairment in this population. Our finding may substantiate the hypothesis that improved hippocampal efficiency, reflected in stronger connections between the hippocampus and critical areas for memory processing such as the prefrontal cortex, may impact cognitive reserve even with some schooling. However, because our study was cross-sectional, we cannot demonstrate causality. The TOFHLA test has been widely used to measure literacy level (Fan et al., 2021). Low literacy measured by the TOFHLA is associated with poor health outcomes (Apolinario et al., 2015; Fan et al., 2021). Although it is well established that the literate brain has different structural and functional properties (Dehaene & Cohen, 2007; Dehaene et al., 2015; Dehaene et al., 2010; Resende, Tovar-Moll, et al., 2018), the neural correlates of literacy measured by literacy tests, and not years of education, is less studied. A previous study showed that higher literacy skills measured by the REALM-SF test correlated with brain structural connectivity, but not with hippocampal volumes (Resende et al., 2022). Interestingly, we found that the very low literacy levels measured by the TOFHLA in our sample was significantly associated with the HC-VMPFC connectivity. We speculated that this finding may reflect how even low levels of literacy can relate to brain functioning, shedding light on a possible mechanism of cognitive reserve in this illiterate population. In terms of episodic memory and brain connectivity, there is still a debate in the literature. The FCSRT is a traditional episodic memory test that has two versions (verbal and visual). The neural correlates of the verbal version have been more explored, while the visual version was less studied. Because the participants were illiterate, the visual version of the FCSRT was more appropriate. The few studies that explored the neural basis of the visual FCSRT test were conducted in persons with high educational level. One study with 14 participants compared the brain activation by the visual FCSRT between novel and repeated stimuli and showed that activations in left superior temporal and left prefrontal cortices were significantly associated with episodic memory (Diamond et al., 2007). Other brain areas activated through the FCSRT stimuli were the inferior parietal lobule, precuneus, hippocampus and parahippocampal gyrus (McLaren et al., 2012) as well as the posterior cingulate cortex and the precuneus connections (Edde et al., 2020). In our study, the lack of association between episodic memory measured by the visual version of the FCSRT and the HC-VMPFC connectivity might be explained by the fact that we did not use task-based functional MRI as the previous studies used, but resting state functional MRI, which might be less sensitive to cognitive-brain correlations (Rasero et al., 2018). Another possibility is that illiterates use less their HC-VMPFC connectivity for memory processing, which might suggest a low cognitive reserve in this group. The fact that we found a significant relationship between literacy levels and the HC-VMPFC connectivity may support this theory, because, as the literacy levels increase, the association becomes stronger. In terms of structural neural correlates of the visual version of the FCSRT, the hippocampal volumes (Slachevsky et al., 2018) and brain areas involving visual processing (Arighi et al., 2018) have been implicated. The verbal version of the FCSRT, however, has been more studied. The hippocampal gray matter volume, mainly the left, has been consistently associated with the verbal version of FCSRT (Arighi et al., 2018; Epelbaum et al., 2018; Ezzati et al., 2016; Frank et al., 2022) in persons with high educational level. This association seems to be more evident in patients with AD (Novellino et al., 2018; Quenon et al., 2016; Sánchez-Benavides et al., 2010) and bvFTD (Bertoux et al., 2018; Poos et al., 2021) than in controls. The very low educational level of our sample combined with the lack of participants with dementia may explain why we did not find an association between episodic memory and hippocampal volumes. Indeed, two previous studies showed that the relationship between episodic memory and hippocampal volumes was moderated by educational level (O’Shea et al., 2018; Resende, Rosen, et al., 2018). Our study has strengths and limitations. It is one of the first studies to look at the associations between the FCSRT visual version and hippocampal functional connectivity and gray matter volumes. The main limitation is the fact that it is cross sectional; therefore, not suitable for demonstrating causality. However, considering the scarcity of studies in illiterate adult populations, we consider it is an important first step into demonstrating whether late life literacy-training might have an impact on cognitive reserve. Nearly all current data available on the cognitive reserve field relate to formal education received in early life, but whether formal education provided during adulthood increases cognitive reserve with downstream benefits on dementia risk it is not known. Even considering the most recent drop in youth illiteracy due to LMIC efforts to provide formal education to school-age children, generations of adults remain illiterate and at higher risk of developing cognitive impairment later in life. If literacy-training in adulthood also improves cognitive reserve, even the current generation of low-educated adults could have benefits, an extremely important issue in LMIC where adult illiteracy rates often exceed 50% (Caribbean et al., 2022). Our next goal is to explore the effects of adult-literacy training in brain structural and functional connectivity as well as in cognitive abilities, to determine whether adult-literacy acquisition might have a beneficial effect on dementia prevention. Eventually, we will be able to inform public policies to increase educational attainment in adulthood with a substantial impact on lowering dementia burden worldwide. Acknowledgments We thank the Alzheimer’s Association and World Federation of Neurology for the funding support. We thank Mrs. Laura Suvalsky Vieira and Mr. Sérgio Martins Duarte for their support with recruiting participants at the Imaculada School, where the participants were recruited. We thank the participants for dedicating their time to research. Funding This work was funded by the Alzheimer’s Association GBHI_ALZ-18-534892 and World Federation of Neurology. Research reported in this publication was partially supported by the National Institute on Aging of the National Institutes of Health under Award Number R21AG0692502. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. PC is funded by CNPq, Brazil (bolsa de produtividade em pesquisa). Availability of data and materials The data that supports the findings are available upon reasonable request. Aggregated and anonymized data, as well as additional information related to the study methodology, can be made available to interested researchers. Requests for data access should be addressed to the corresponding author, Dr. Elisa de Paula França Resende (elisaresende@gbhi.org), who will assess each request on a case-by-case basis in consultation with the research team and in compliance with applicable data protection regulations and institutional policies. Figure 1 Correlation between hippocampal connectivity and low literacy levels. The statistical map is displayed on an inflated brain image. The heat maps represent the T statistical value for the connectivity between the right and left hippocampal seed and the other clusters. Blue means anticorrelation and red means positive correlation. The graph depicts the correlation between literacy levels measured by the Test of Functional Health Literacy Assessment (TOFHLA) and the right HC-VMPFC connectivity. Table 1 Participants characteristics Characteristics n = 35 Age (years) 50.0 [42.5; 58.0] Sex female n(%) 20 (57.1%) Self-reported race Blacks 28 (84.8%) Whites 3 (9.1%) Indigenous 2 (6.1%) Unknown 2 (5.4%) Socioeconomic level B 6 (17.1%) C 11 (32.4%) D-E 17 (50.0%) Current anxiety 5 (14.3%) Current depression 6 (17.1%) Baecke physical exercise scale 3.0 [2.2; 5.8] Cognitive Reserve Index 73.0 [70.0; 79.0] MMSE 22.0 [21.0; 25.5] Animals’ fluency/min 14.0 [12.0; 16.5] Brief cognitive battery delayed recall 8.0 [7.5; 9.0] TOFHLA total 28.0 [21; 42.5] FCSRT free recall sum of attempts 30.0 [26.2; 35.0] FCSRT cue efficiency 0.98 [0.96; 1.0] FCSRT delayed free recall 11.0 [9.0; 13.0] Word reading test 40.0 [0.0; 66.5] Token verbal comprehension 27.0 [21.5; 29.0] Rapid naming colors (seg) 45.5 [42.2; 57.7] Rapid naming letters (seg) 41.5 [32.2; 54.2] Rapid naming numbers (seg) 35.5 [31.2; 42.0] Rapid naming objects (seg) 55.0 [47.2; 62.0] Non-verbal intelligence Beta III test 6.0 [5.0; 7.7] Right Hippocampal volume (mm3) 2.3 [2.1; 2.5] Left Hippocampal volume (mm3) 2.3 [2.2; 2.4] Values depicted in median and Interquartile interval. See the text for more details about the socioeconomic levels. Table 2 General linear models showing the association between FCSRT free recall sum-of-attempts scores and functional connectivity and hippocampal volume. β t p Sex (Male) 0.09 0.23 0.822 Age −0.38 −1.34 0.193 Right HC - VMPFC connectivity 0.15 0.63 0.535 Left HC - VMPFC connectivity 0.24 0.86 0.400 Right HC - PCC connectivity −0.39 −1.67 0.109 Left HC - PCC connectivity 0.23 0.97 0.343 VMPFC - PCC connectivity 0.25 1.22 0.235 Left hippocampus volume 0.11 0.31 0.762 Right hippocampus volume −0.37 −0.98 0.336 TIV 0.04 0.23 0.822 FCSRT: Free and Cued Selective Reminding Test, TIV: total intracranial volume, HC: hippocampus, VMPFC: Ventral medial pre-frontal cortex, PCC: precuneus Table 3 General linear models showing the association between TOFHLA scores (literacy) and functional connectivity and hippocampal volume. Names β t p Sex (Male) 0.25 0.65 0.522 Age −0.16 −0.76 0.456 Right HC - VMPFC connectivity 0.58 2.90 0.008 Left HC - VMPFC connectivity −0.35 −1.5 0.145 Right HC - PCC connectivity 0.10 0.51 0.616 Left HC - PCC connectivity 0.16 0.82 0.419 VMPFC - PCC connectivity 0.18 1.0 0.320 Left hippocampus volume 0.39 1.26 0.219 Right hippocampus volume −0.35 −1.08 0.288 TIV 0.22 1.3 0.208 FCSRT: Free and Cued Selective Reminding Test, TIV: total intracranial volumes, HC: hippocampus, VMPFC: Ventral medial pre-frontal cortex, PCC: precuneus. Declarations Competing interests The authors declare that there are no conflicts of interest that could have influenced the design, conduct, or reporting of the study. Financial or personal relationships that could potentially bias the research findings were disclosed and managed appropriately. Ethical Approval The present study adheres to the ethical standards and guidelines in research, and it was approved by the Institutional Ethical Review Board – Comitê de Ética em Pesquisa da Universidade Federal de Minas Gerais. Approval number 2.955.960, CAAE number: 89764918.2.0000.5149. Informed Consent was obtained prior to data collection, informed consent was obtained from all participants or their legal representatives. The purpose, procedures, potential risks, and benefits of the study were clearly explained, ensuring that participants understood their rights and had the opportunity to ask questions. All personal information and data collected from participants were treated with confidentiality. Identifying information was anonymized and stored securely, limiting access to authorized researchers only. Any data presented in the manuscript has been de-identified to ensure the privacy and confidentiality of participants. The study protocol was reviewed and approved by the relevant institutional or independent ethics committee, ensuring that it complied with ethical guidelines and safeguarded the welfare and rights of participants. ==== Refs References 1. Aggleton J. P. , & Brown M. W. (1999). Episodic memory, amnesia, and the hippocampal-anterior thalamic axis. Behav Brain Sci, 22 (3 ), 425–444; discussion 444–489. 11301518 2. Andersson J. L. , Hutton C. , Ashburner J. , Turner R. , & Friston K. (2001). Modeling geometric deformations in EPI time series. NeuroImage, 13 (5 ). 10.1006/nimg.2001.0746 3. Apolinario D. , Mansur L. L. , Carthery-Goulart M. T. , Brucki S. M. , & Nitrini R. (2015). Cognitive predictors of limited health literacy in adults with heterogeneous socioeconomic backgrounds. J Health Psychol, 20 (12 ), 1613–1625. 10.1177/1359105313520337 24496057 4. Arighi A. , Carandini T. , Mercurio M. , Carpani G. , Pietroboni A. M. , Fumagalli G. , … Scarpini E . (2018). Word and Picture Version of the Free and Cued Selective Reminding Test (FCSRT): Is There Any Difference? Journal of Alzheimer’s disease : JAD, 61 (1 ). 10.3233/JAD-170712 5. Ashburner J. , & Friston K. J. (2005). Unified segmentation. NeuroImage, 26 (3 ). 10.1016/j.neuroimage.2005.02.018 6. Ashburner J. , & Friston K. J. (2011). Diffeomorphic registration using geodesic shooting and Gauss-Newton optimisation. NeuroImage, 55 (3 ). 10.1016/j.neuroimage.2010.12.049 7. Baecke J. A. , Burema J. , & Frijters J. E. (1982). A short questionnaire for the measurement of habitual physical activity in epidemiological studies. The American journal of clinical nutrition, 36 (5 ). 10.1093/ajcn/36.5.936 8. Behzadi Y. , Restom K. , Liau J. , & Liu T. T. (2007). A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. NeuroImage, 37 (1 ). 10.1016/j.neuroimage.2007.04.042 9. Bertoux M. , Flanagan E. C. , Hobbs M. , Ruiz-Tagle A. , Delgado C. , Miranda M. , … Hornberger M . (2018). Structural Anatomical Investigation of Long-Term Memory Deficit in Behavioral Frontotemporal Dementia. J Alzheimers Dis, 62 (4 ), 1887–1900. 10.3233/jad-170771 29614645 10. Bhattacharyya K. B. (2017). James Wenceslaus Papez, His Circuit, and Emotion. In Ann Indian Acad Neurol (Vol. 20 , pp. 207–210). 10.4103/aian.AIAN_487_16 28904449 11. Brucki S. M. , Nitrini R. , Caramelli P. , Bertolucci P. H. , & Okamoto I. H. (2003). Sugestões para o uso do Mini-Exame do Estado Mental no Brasil. Arq Neuropsiquiatr., 61 , 777–781. 10.1590/S0004-282X2003000500014 14595482 12. Calhoun V. D. , Wager T. D. , Krishnan A. , Rosch K. S. , Seymour K. E. , Nebel M. B. , … Kiehl K . (2017). The impact of T1 versus EPI spatial normalization templates for fMRI data analyses. Human brain mapping, 38 (11 ). 10.1002/hbm.23737 13. Caribbean U. O. S. a. R. B. f. E. i. L. A. a. t., (Panama), U. R. O. f. L. A. a. t. C., & Caribbean, E. C. f. L. A. a. t. (2022). Education in Latin America and the Caribbean at a crossroads: regional monitoring report SDG4 - Education 2030. 14. Chumbley J. , Worsley K. , Flandin G. , & Friston K. (2010). Topological FDR for neuroimaging. NeuroImage, 49 (4 ). 10.1016/j.neuroimage.2009.10.090 15. Cummings J. , Lee G. , Nahed P. , Kambar M. E. Z. N. , Zhong K. , Fonseca J. , & Taghva K. (2022). Alzheimer’s disease drug development pipeline: 2022. Alzheimer’s & dementia (New York, N.Y.), 8 (1 ). 10.1002/trc2.12295 16. César-Freitas K. G. , Suemoto C. K. , Power M. C. , Brucki S. M. D. , & Nitrini R. (2022). Incidence of dementia in a Brazilian population: The Tremembé Epidemiologic Study. Alzheimer’s & dementia : the journal of the Alzheimer’s Association, 18 (4 ). 10.1002/alz.12423 17. da Silva P. B ., Engel de Abreu P. M. J , Laurence P. G , Nico M. Â. N , Simi L. G. V ., Tomás R. C , & Macedo . (2020). Rapid Automatized Naming and Explicit Phonological Processing in Children With Developmental Dyslexia: A Study With Portuguese-Speaking Children in Brazil. Frontiers in psychology, 11 . 10.3389/fpsyg.2020.00928 18. de Paula J. J. , Bertola L. , Avila R. T. , Moreira L. , Coutinho G. , de Moraes E. N. , … Malloy-Diniz L. F . (2013). Clinical applicability and cutoff values for an unstructured neuropsychological assessment protocol for older adults with low formal education. PLoS One, 8 (9 ), e73167. 10.1371/journal.pone.0073167 24066031 19. Dehaene S. , & Cohen L. (2007). Cultural Recycling of Cortical Maps. Neuron., 25 (2 ), 384–398. 10.1016/j.neuron.2007.10.004 20. Dehaene S. , Cohen L. , Morais J. , & Kolinsky R. (2015). Illiterate to literate: behavioural and cerebral changes induced by reading acquisition. Nature Reviews Neuroscience., 16 , 234–244. 10.1038/nrn3924 25783611 21. Dehaene S. , Pegado F. , Braga L. W. , Ventura P. , Nunes-Filho G. , Jobert A. , … Cohen L . (2010). How Learning to Read Changes the Cortical Networks for Vision and Language. Science., 330 , 1359–1364. 10.1126/science.1194140 21071632 22. Desikan R. S. , Ségonne F. , Fischl B. , Quinn B. T. , Dickerson B. C. , Blacker D. , … Killiany R. J . (2006). An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage, 31 (3 ), 968–980. 10.1016/j.neuroimage.2006.01.021 16530430 23. Diamond E. L. , Miller S. , Dickerson B. C. , Atri A. , DePeau K. , Fenstermacher E. , … Sperling R., A . (2007). Relationship of fMRI activation to clinical trial memory measures in Alzheimer disease. Neurology, 69 (13 ). 10.1212/01.wnl.0000277292.37292.69 24. Edde M. , Dilharreguy B. , Theaud G. , Chanraud S. , Helmer C. , Dartigues J. F. , … Catheline G . (2020). Age-related change in episodic memory: role of functional and structural connectivity between the ventral posterior cingulate and the parietal cortex. Brain structure & function, 225 (7 ). 10.1007/s00429-020-02121-7 25. Eichenbaum H. (2004). Hippocampus: cognitive processes and neural representations that underlie declarative memory. Neuron, 44 (1 ), 109–120. 10.1016/j.neuron.2004.08.028 15450164 26. Epelbaum S. , Bouteloup V. , Mangin J. F. , La Corte V. , Migliaccio R. , Bertin H. , … Dufouil C . (2018). Neural correlates of episodic memory in the Memento cohort. Alzheimer’s & dementia (New York, N.Y.), 4 . 10.1016/j.trci.2018.03.010 27. Ezzati A. , Katz M. J. , Zammit A. R. , Lipton M. L. , Zimmerman M. E. , Sliwinski M. J. , & Lipton R. B. (2016). Differential association of left and right hippocampal volumes with verbal episodic and spatial memory in older adults. Neuropsychologia, 93 (Pt B ). 10.1016/j.neuropsychologia.2016.08.016 28. Fan Z. Y. , Yang Y. , & Zhang F. (2021). Association between health literacy and mortality: a systematic review and meta-analysis. Archives of public health = Archives belges de sante publique, 79 (1 ). 10.1186/s13690-021-00648-7 29. Fernandez-Cabello S. , Valls-Pedret C. , Schurz M. , Vidal-Pineiro D. , Sala-Llonch R. , Bargallo N. , … Bartres-Faz D . (2016). White matter hyperintensities and cognitive reserve during a working memory task: a functional magnetic resonance imaging study in cognitively normal older adults. Neurobiol Aging., 48 , 23–33. 10.1016/j.neurobiolaging.2016.08.008 27636672 30. Folstein M. F. , Folstein S. E. , & McHugh P. R. (1975). ‘Mini-Mental State’. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res, 12 , 189–198. 10.1016/0022-3956(75)90026-6 1202204 31. Frank D. , Garo-Pascual M. , Velasquez P. A. R. , Frades B. , Peled N. , Zhang L. , & Strange B. A. (2022). Brain structure and episodic learning rate in cognitively healthy ageing. NeuroImage, 263 . 10.1016/j.neuroimage.2022.119630 32. Friston K. J. , Ashburner J. , Frith C. D. , Poline J. B. , Heather J. D. , & Frackowiak R. S. J. (1995). Spatial Registration and Normalization of Images. Human Brain Mappin, 2 , 165–189. 33. Friston K. J. , Williams S. , Howard R. , Frackowiak R. S. , & Turner R. (1996). Movement-related effects in fMRI time-series. Magnetic resonance in medicine, 35 (3 ). 10.1002/mrm.1910350312 34. Goodglass H. , & Kaplan E. (1983). Boston Diagnostic Aphasia Examination (BDAE). Lea and Febiger. 35. Grober E. , Sanders A. E. , Hall C. , & Lipton R. B. (2010). Free and cued selective reminding identifies very mild dementia in primary care. Alzheimer Dis Assoc Disord, 24 (3 ), 284–290. 10.1097/WAD.0b013e3181cfc78b 20683186 36. Hallquist M. N. , Hwang K. , & Luna B. (2013). The nuisance of nuisance regression: spectral misspecification in a common approach to resting-state fMRI preprocessing reintroduces noise and obscures functional connectivity. NeuroImage, 82 . 10.1016/j.neuroimage.2013.05.116 37. Henson R. N. A. , Buechel C. , Josephs O. , & Friston K. J. (1999). The slice-timing problem in event-related fMRI. NeuroImage(9 ), 125.10417246 38. Hoenig M. C. , Bischof G. N. , Hammes J. , Faber J. , Fliessbach K. , van Eimeren T. , & Drzezga A. (2017). Tau pathology and cognitive reserve in Alzheimer’s disease. Neurobiol Aging, 57 , 1–7. 10.1016/j.neurobiolaging.2017.05.004 28577411 39. Lu K. , Nicholas J. M. , Collins J. D. , James S. N. , Parker T. D. , Lane C. A. , … Schott J. M . (2019). Cognition at age 70: Life course predictors and associations with brain pathologies. Neurology. 10.1212/wnl.0000000000008534 40. Malotaux V. , Dricot L. , Quenon L. , Lhommel R. , Ivanoiu A. , & Hanseeuw B. (2022). Default-Mode Network Connectivity Changes During the Progression Toward Alzheimer’s Dementia: A Longitudinal Functional Magnetic Resonance Imaging Study. Brain connectivity. 10.1089/brain.2022.0008 41. Maragno C. A. D. , Mengue S. S. , Moraes C. G. , Rebelo M. V. D. , Guimaraes A. M. M. , & Pizzol T. (2019). Test of health Literacy for Portuguese-speaking Adults. Rev Bras Epidemiol, 22 , e190025. 10.1590/1980-549720190025 (Teste de letramento em saude em portugues para adultos.)30942331 42. Martins M. A. , Begeny J. C. , & Capellini S. A. (2023). Translation and cultural adaptation of the HELPS Reading Fluency Program into Brazilian Portuguese: A report of systematic adaptation processes and initial evidence of efficacy. Frontiers in psychology, 14 . 10.3389/fpsyg.2023.1034749 43. Mazziotta J. C. , Toga A. W. , Evans A. , Fox P. , & Lancaster J. (1995). A probabilistic atlas of the human brain: theory and rationale for its development. The International Consortium for Brain Mapping (ICBM). NeuroImage, 2 (2 ). 10.1006/nimg.1995.1012 44. McLaren D. G. , Sreenivasan A. , Diamond E. L. , Mitchell M. B. , Van Dijk K. R. , Deluca A. N. , … Atri A . (2012). Tracking cognitive change over 24 weeks with longitudinal functional magnetic resonance imaging in Alzheimer’s disease. Neuro-degenerative diseases, 9 (4 ). 10.1159/000335876 45. Miotto E. , Sato J. , Lucia M. , Camargo C. , & Scaff M. (2010). Development of an adapted version of the Boston Naming Test for Portuguese speakers. Revista brasileira de psiquiatria (Sao Paulo, Brazil : 1999), 32 (3 ). 10.1590/s1516-44462010005000006 46. Mukadam N. , Sommerlad A. , Huntley J. , & Livingston G. (2019). Population attributable fractions for risk factors for dementia in low-income and middle-income countries: an analysis using cross-sectional survey data. Lancet Glob Health, 7 (5 ), e596–e603. 10.1016/s2214-109x(19)30074-9 31000129 47. Murray A. D. , Staff R. T. , McNeil C. J. , Salarirad S. , Ahearn T. S. , Mustafa N. , & Whalley L. J. (2011). The balance between cognitive reserve and brain imaging biomarkers of cerebrovascular and Alzheimer’s diseases. Brain, 134 (Pt 12 ), 3687–3696. 10.1093/brain/awr259 22102649 48. Nieto-Castanon A. (2020). FMRI minimal preprocessing pipeline. In Press H . (Ed.), Handbook of functional connectivity Magnetic Resonance Imaging methods in CONN (pp. 3–16). Hilbert Press. 49. Nitrini R. , Bottino C. M. , Albala C. , Custodio Capunay N. S. , Ketzoian C. , Llibre Rodriguez J. J. , … Caramelli P . (2009). Prevalence of dementia in Latin America: a collaborative study of population-based cohorts. Int Psychogeriatr, 21 (4 ), 622–630. https://doi.org/S1041610209009430 [pii] 10.1017/S1041610209009430 19505354 50. Norton S. , Matthews F. E. , Barnes D. E. , Yaffe K. , & Brayne C. (2014). Potential for primary prevention of Alzheimer’s disease: an analysis of population-based data. Lancet Neurol. 13 (8 ), 788–794. 10.1016/S1474-4422(14)70136-X.25030513 51. Novellino F. , Vasta R. , Sarica A. , Chiriaco C. , Salsone M. , Morelli M. , … Quattrone A . (2018). Relationship between Hippocampal Subfields and Category Cued Recall in AD and PDD: A Multimodal MRI Study. Neuroscience, 371 . 10.1016/j.neuroscience.2017.12.028 52. Nucci M. , Mapelli D. , & Mondini S. (2012). Cognitive Reserve Index questionnaire (CRIq): a new instrument for measuring cognitive reserve. Aging Clin Exp Res, 24 (3 ), 218–226. 10.3275/7800 21691143 53. O’Shea D. M. , Langer K. , Woods A. J. , Porges E. C. , Williamson J. B. , O’Shea A. , & Cohen R. A. (2018). Educational Attainment Moderates the Association Between Hippocampal Volumes and Memory Performances in Healthy Older Adults. Front Aging Neurosci, 10 , 361. 10.3389/fnagi.2018.00361 30467475 54. Parker R. M. , Baker D. W. , Williams M. V. , & Nurss J. R. (1995). The test of functional health literacy in adults: a new instrument for measuring patients’ literacy skills. J Gen Intern Med., 10 (10 ), 537–541.8576769 55. Passarelli M. , Masini M. , Bracco F. , Petrosino M. , & Chiorri C. (2018). Development and validation of the Facial Expression Recognition Test (FERT). Psychological assessment, 30 (11 ). 10.1037/pas0000595 56. Penny W. D. , Friston K. J. , Ashburner J. T. , Kiebel S. J. , & Nichols T. E. (2011). Statistical parametric mapping: the analysis of functional brain images. Elsevier. 57. Poos J. M. , Russell L. L. , Peakman G. , Bocchetta M. , Greaves C. V. , Jiskoot L. C. , … Rohrer J. D . (2021). Impairment of episodic memory in genetic frontotemporal dementia: A GENFI study. Alzheimer’s & dementia (Amsterdam, Netherlands), 13 (1 ). 10.1002/dad2.12185 58. Power J. D. , Mitra A. , Laumann T. O. , Snyder A. Z. , Schlaggar B. L. , & Petersen S. E. (2014). Methods to detect, characterize, and remove motion artifact in resting state fMRI. NeuroImage, 84 . 10.1016/j.neuroimage.2013.08.048 59. Prince M. , Wimo A. , Guerchet M. , Ali G.-C. , Wu Y.-T. , & Prina M. (2015). World Alzheimer Report 2015 - The Global Impact of Dementia: an analysis of prevalence, incidence, cost and trends. A. s. D. I. (ADI). 60. Quenon L. , Dricot L. , Woodard J. , Hanseeuw B. , Gilis N. , Lhommel R. , & Ivanoiu A. (2016). Prediction of Free and Cued Selective Reminding Test Performance Using Volumetric and Amyloid-Based Biomarkers of Alzheimer’s Disease. Journal of the International Neuropsychological Society : JINS, 22 (10 ). 10.1017/S1355617716000813 61. Rabelo I. S. A. , & Pacanaro S. V. d. S. L ., Irene F . Alves Almeida , da Silva Ambiel Gisele Aparecida , Matteo Rodolfo Augusto . (2011). Teste Não Verbal de Inteligência Geral - BETA-III: subtestes raciocínio matricial e códigos / C. E. Kellogg, N. W. Morton; Padronização brasileira. Editora Casa do Psicólogo. 62. Rasero J. , Aerts H. , Ontivero O.M , Cortes J. M ., Stramaglia S ., & Marinazzo D . (2018). Predicting functional networks from region connectivity profiles in task-based versus resting-state fMRI data. PloS one, 13 (11 ). 10.1371/journal.pone.0207385 63. Resende E. , Xia F. , Sidney S. , Launer L. J. , Schreiner P. J. , Erus G. , … Yaffe K . (2022). Higher literacy is associated with better white matter integrity and cognition in middle age. Alzheimer’s & dementia (Amsterdam, Netherlands), 14 (1 ). 10.1002/dad2.12363 64. Resende E. P. F. , Rosen H. J. , Chiang K. , Staffaroni A. M. , Allen I. , Grinberg L. T. , … Caramelli P . (2018). Primary School Education May Be Sufficient to Moderate a Memory-Hippocampal Relationship. Front Aging Neurosci, 10 , 381. 10.3389/fnagi.2018.00381 30515091 65. Resende E. P. F. , Tovar-Moll F. F. , Ferreira F. M. , Bramati I. , de Souza L. C. , Carmona K. C. , … Caramelli P . (2017). Integrity of white matter structure is related to episodic memory performance in the low-educated elderly. Arq Neuropsiquiatr, 75 (11 ), 778–784. 10.1590/0004-282x20170158 29236820 66. Resende E. P. F. , Tovar-Moll F. F. , Ferreira F. M. , Bramati I. , de Souza L. C. , Carmona K. C. , … Caramelli P . (2018). White Matter Microstructure in Illiterate and Low-Literate Elderly Brazilians: Preliminary Findings. Cogn Behav Neurol, 31 (4 ), 193–200. 10.1097/wnn.0000000000000173 30562228 67. Ribeiro F. , Teixeira-Santos A. C. , Caramelli P. , & Leist A. K. (2022). Prevalence of dementia in Latin America and Caribbean countries: Systematic review and meta-analyses exploring age, sex, rurality, and education as possible determinants. Ageing research reviews, 81 . 10.1016/j.arr.2022.101703 68. Rocha D. S. , Dibai-Filho A. V. , Pinheiro J. S. , Azevedo L. F. S. , Falcai A. , Tavarez R. R. J. , … Bassi-Dibai D . (1992). The Baecke Habitual Physical Activity Questionnaire (BHPAQ): a valid internal structure of the instrument to assess healthy Brazilian adults. Rev Assoc Med Bras, 68 (7 ). 10.1590/1806-9282.20211374 69. Satizabal C. L. , Beiser A. S. , Chouraki V. , Chene G. , Dufouil C. , & Seshadri S. (2016). Incidence of Dementia over Three Decades in the Framingham Heart Study. N Engl J Med., 374 (6 ), 523–532. 10.1056/NEJMoa1504327 26863354 70. Seeley W. W. , Crawford R. K. , Zhou J. , Miller B. L. , & Greicius M. D. (2009). Neurodegenerative diseases target large-scale human brain networks. Neuron, 62 (1 ), 42–52. https://doi.org/S0896-6273(09)00249-9 [pii] 10.1016/j.neuron.2009.03.024 19376066 71. Sharp E. S. , & Gatz M. (2011). Relationship between education and dementia: an updated systematic review. Alzheimer Dis Assoc Disord, 25 (4 ), 289–304. 10.1097/WAD.0b013e318211c83c 21750453 72. Sheehan D. V. , Lecrubier Y. , Sheehan K. H. , Amorim P. , Janavs J. , Weiller E. , … Dunbar G . (1998). The Mini-International Neuropsychiatric Interview (M.I.N.I.): the development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry, 59 (Suppl 20 ), 22–33;quiz 34–57. 73. Sidhu M. K. , Stretton J. , Winston G. P. , Bonelli S. , Centeno M. , Vollmar C. , … Duncan J. S . (2013). A functional magnetic resonance imaging study mapping the episodic memory encoding network in temporal lobe epilepsy. Brain, 136 (Pt 6 ), 1868–1888. 10.1093/brain/awt099 23674488 74. Slachevsky A. , Barraza P. , Hornberger M. , Muñoz-Neira C. , Flanagan E. , Henríquez F. , … Delgado C . (2018). Neuroanatomical Comparison of the “Word” and “Picture” Versions of the Free and Cued Selective Reminding Test in Alzheimer’s Disease. Journal of Alzheimer’s disease : JAD, 61 (2 ). 10.3233/JAD-160973 75. Sladky R. , Friston K. J. , Tröstl J. , Cunnington R. , Moser E. , & Windischberger C. (2011). Slice-timing effects and their correction in functional MRI. NeuroImage, 58 (2 ). 10.1016/j.neuroimage.2011.06.078 76. Staffaroni A. M. , Brown J. A. , Casaletto K. B. , Elahi F. M. , Deng J. , Neuhaus J. , … Kramer J. H . (2018). The Longitudinal Trajectory of Default Mode Network Connectivity in Healthy Older Adults Varies As a Function of Age and Is Associated with Changes in Episodic Memory and Processing Speed. The Journal of neuroscience : the official journal of the Society for Neuroscience, 38 (11 ). 10.1523/JNEUROSCI.3067-17.2018 77. Stern Y. , Albert M. , Barnes C. A. , Cabeza R. , Pascual-Leone A. , & Rapp P. R. (2023). A framework for concepts of reserve and resilience in aging. Neurobiology of aging, 124 . 10.1016/j.neurobiolaging.2022.10.015 78. Suemoto C. K. , Ferretti-Rebustini R. E. , Rodriguez R. D. , Leite R. E. , Soterio L. , Brucki S. M. , … Grinberg L. T . (2017). Neuropathological diagnoses and clinical correlates in older adults in Brazil: A cross-sectional study. PLoS Med, 14 (3 ), e1002267. 10.1371/journal.pmed.1002267 28350821 79. Suemoto C. K. , Mukadam N. , Brucki S. M. D.P, C., R, N., J, L., … P, F. C ,. (2022). Risk factors for dementia in Brazil: Differences by region and race. Alzheimer’s & dementia : the journal of the Alzheimer’s Association. 10.1002/alz.12820 80. Sánchez-Benavides G. , Gómez-Ansón B. , Molinuevo J. L. , Blesa R. , Monte G. C. , Buschke H. , & Peña-Casanova J. (2010). Medial temporal lobe correlates of memory screening measures in normal aging, MCI, and AD. Journal of geriatric psychiatry and neurology, 23 (2 ). 10.1177/0891988709355271 81. Thiebaut de Schotten M. , Cohen L. , Amemiya E. , Braga L. W. , & Dehaene S. (2014). Learning to read improves the structure of the arcuate fasciculus. Cereb Cortex, 24 (4 ), 989–995. 10.1093/cercor/bhs383 23236205 82. van Kesteren M. T. , Fernández G. , Norris D. G. , & Hermans E. J. (2010). Persistent schema-dependent hippocampal-neocortical connectivity during memory encoding and postencoding rest in humans. Proc Natl Acad Sci U S A, 107 (16 ), 7550–7555. 10.1073/pnas.0914892107 20363957 83. Wang H. X. , Gustafson D. R. , Kivipelto M. , Pedersen N. L. , Skoog I. , Windblad B. , & Fratiglioni L. (2012). Education halves the risk of dementia due to apolipoprotein epsilon4 allele: a collaborative study from the Swedish brain power initiative. Neurobiol Aging, 33 (5 ), 1007.e1001–1007. 10.1016/j.neurobiolaging.2011.10.003 84. Whitfield-Gabrieli S. , & Nieto-Castanon A. (2012). Conn: a functional connectivity toolbox for correlated and anticorrelated brain networks. Brain connectivity, 2 (3 ). 10.1089/brain.2012.0073 85. Whitfield-Gabrieli S. , Nieto-Castanon A. , Ghosh S. (2009). Artifact detection tools (ART). In (Version Release Version 7 (19), 11.) Cambridge, MA. 86. Wirth M. , Villeneuve S. , La Joie R. , Marks S. M. , & Jagust W. J. (2014). Gene-environment interactions: lifetime cognitive activity, APOE genotype, and β-amyloid burden. J Neurosci, 34 (25 ), 8612–8617. 10.1523/JNEUROSCI.4612-13.2014 24948815 87. Worsley K. J. , Marrett S. , Neelin P. , Vandal A. C. , Friston K. J. , & Evans A. C. (1996). A unified statistical approach for determining significant signals in images of cerebral activation. Human brain mapping, 4 (1 ). 10.1002/(SICI)1097-0193(1996)4:1&lt;58::AID-HBM4&gt;3.0.CO;2-O 88. Wu Y. T. , Beiser A. S. , Breteler M. M. B. , Fratiglioni L. , Helmer C. , Hendrie H. C. , … Brayne C . (2017). The changing prevalence and incidence of dementia over time - current evidence. Nat Rev Neurol, 13 (6 ), 327–339. 10.1038/nrneurol.2017.63 28497805 89. Zhou J. , Greicius M. D. , Gennatas E. D. , Growdon M. E. , Jang J. Y. , Rabinovici G. D. , … Seeley W. W . (2010). Divergent network connectivity changes in behavioural variant frontotemporal dementia and Alzheimer’s disease. Brain : a journal of neurology, 133 (Pt 5 ). 10.1093/brain/awq075 90. Zibetti M. R. , Bordignon S. , & Trentini C. M. (2014). Memória e aprendizagem no procedimentode recordação seletiva livre e com pistas. Temas em Psicologia, 22 (4 ), 771–782. 10.9788/TP2014.4-08

PMC010xxxxxx/PMC10312991.txt

==== Front Res Sq ResearchSquare Research Square American Journal Experts 37398196 10.21203/rs.3.rs-3036166/v1 10.21203/rs.3.rs-3036166 preprint 1 Article High-Content Small Molecule Screen Identifies a Novel Compound That Restores AP-4-Dependent Protein Trafficking in Neuronal Models of AP-4-Associated Hereditary Spastic Paraplegia Saffari Afshin Boston Children’s Hospital, Harvard Medical School Brechmann Barbara Boston Children’s Hospital, Harvard Medical School Boeger Cedric Boston Children’s Hospital, Harvard Medical School http://orcid.org/0000-0001-9488-7893 Saber Wardiya Afshar Boston Children’s Hospital/Harvard Medical School http://orcid.org/0000-0003-0495-5317 jumo Hellen http://orcid.org/0000-0002-6895-9627 Whye Dosh Boston Children’s Hospital, Harvard Medical School Wood Delaney Boston Children’s Hospital, Harvard Medical School Wahlster Lara Boston Children’s Hospital, Harvard Medical School Alecu Julian Boston Children’s Hospital, Harvard Medical School http://orcid.org/0000-0002-0504-7815 Ziegler Marvin Harvard Medical School Scheffold Marlene Boston Children’s Hospital, Harvard Medical School Winden Kellen Boston Children’s Hospital Hubbs Jed Boston Children’s Hospital http://orcid.org/0000-0003-3194-5413 Buttermore Elizabeth Boston Children’s Hospital Barrett Lee Boston Children’s Hospital http://orcid.org/0000-0002-3166-3435 Borner Georg Max Planck Institute of Biochemistry http://orcid.org/0000-0002-1594-8780 Davies Alexandra Max Planck Institute of Biochemistry http://orcid.org/0000-0001-7044-2953 Sahin Mustafa Boston Children’s Hospital http://orcid.org/0000-0002-0026-4714 Ebrahimi-Fakhari Darius Boston Children’s Hospital, Harvard Medical School AUTHOR CONTRIBUTIONS A.S., A.K.D., M.S., D.E.F. conceptualized and designed the experiments. A.S., B.B., A.K.D., C.B., W.A.S., H.J., D.Wh., D.Wo., L.W., J.E.A., M.Z., K.W. performed experiments. J.H. designed and supervised the re-synthesis of compound C-01. E.D.B., L.B., S.S., M.A., provided technical assistance and analysis tools. A.S., B.B. and D.E.F. wrote the first draft of the manuscript. All authors contributed to the final draft of the manuscript. D.E.F. wrote the grants that supported this project. M.S. and D.E.F. supervised the project. ✉ darius.ebrahimi-fakhari@childrens.harvard.edu 12 6 2023 rs.3.rs-3036166https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-rs3036166v1.pdf Unbiased phenotypic screens in patient-relevant disease models offer the potential to detect novel therapeutic targets for rare diseases. In this study, we developed a high-throughput screening assay to identify molecules that correct aberrant protein trafficking in adaptor protein complex 4 (AP-4) deficiency, a rare but prototypical form of childhood-onset hereditary spastic paraplegia, characterized by mislocalization of the autophagy protein ATG9A. Using high-content microscopy and an automated image analysis pipeline, we screened a diversity library of 28,864 small molecules and identified a lead compound, C-01, that restored ATG9A pathology in multiple disease models, including patient-derived fibroblasts and induced pluripotent stem cell-derived neurons. We used multiparametric orthogonal strategies and integrated transcriptomic and proteomic approaches to delineate putative molecular targets of C-01 and potential mechanisms of action. Our results define molecular regulators of intracellular ATG9A trafficking and characterize a lead compound for the treatment of AP-4 deficiency, providing important proof-of-concept data for future Investigational New Drug (IND)-enabling studies. ==== Body pmcINTRODUCTION Despite remarkable advances in our ability to delineate the genetic causes of rare neurological diseases, it is estimated that specific therapies exist for less than 5% 1. Thus, there is a significant unmet need for developing and implementing novel platforms for drug discovery. Informed by disease-relevant cellular phenotypes, automated and unbiased cell-based high-throughput small molecule screens have the potential to uncover novel therapeutic targets 2, 3, 4, 5, 6. Adaptor protein complex 4 (AP-4)-related hereditary spastic paraplegia (AP-4-HSP), which comprises AP4B1-associated SPG47 (OMIM #614066), AP4M1-associated SPG50 (OMIM #612936), AP4E1-associated SPG51 (OMIM #613744) and AP4S1-associated SPG52 (OMIM #614067), is a rare but prototypical form of childhood-onset complex hereditary spastic paraplegia (HSP) and an important genetic mimic of cerebral palsy 7, 8. Children with AP-4-HSP present with features of both a neurodevelopmental disorder (e.g., early-onset global developmental delay and seizures, microcephaly, and developmental brain malformations) and a neurodegenerative disease (e.g., progressive spasticity and weakness, loss of ambulation, and extrapyramidal movement disorders) 7, 8, 9, 10. AP-4-HSP is caused by bi-allelic loss-of-function variants in any of the four AP-4 subunits (ε, β4, μ4, σ4), leading to impaired AP-4 assembly and function 11, 12, 13, 14, 15. AP-4 is an obligate heterotetrameric protein complex 16, 17, 18 that mediates transport from the trans-Golgi network (TGN) to the cell periphery, including sites of autophagosome biogenesis 19, 20. Three independent groups identified the core autophagy protein and lipid scramblase ATG9A as a major cargo of AP-4 11, 12, 13, 14, 21, linking loss of AP-4 function to defective autophagy 22, 23. AP-4 deficiency in non-neuronal 11, 12, 21, 24, 25 and neuronal cells 13, 14, 15 leads to an accumulation of ATG9A in the TGN, including in iPSC-derived neurons from AP-4-HSP patients 15. From this body of work, and overlapping neuronal phenotypes of AP-4 13, 14, 26, 27 and Atg9a28 knockout mice, the following working model for AP-4 deficiency emerges: (1) AP-4 is required for trafficking of ATG9A from the TGN; (2) loss-of-function variants in AP-4 subunits lead to a loss of AP-4 function; (3) ATG9A accumulates in the TGN leading to a reduction of axonal delivery of ATG9A; (4) lack of ATG9A at the distal axon impairs autophagy leading to axonal degeneration. Other AP-4 cargo proteins identified to date include the poorly characterized transmembrane proteins SERINC1 and SERINC3 12, and the endocannabinoid producing enzyme DAG lipase beta (DAGLB) 29. In this study, we leverage intracellular ATG9A mislocalization as a cellular readout for AP-4 deficiency to develop a large-scale, automated, multiparametric, unbiased phenotypic small molecule screen for modulators of ATG9A trafficking in patient-derived cellular models. We employed this platform to screen a diversity library of 28,864 novel small molecules in AP-4-deficient patient fibroblasts and identified 503 compounds that re-distribute ATG9A from the TGN to the cytoplasm. Through a series of orthogonal assays in neuronal cells, including differentiated AP4B1KO SH-SY5Y cells and human induced pluripotent stem cell (hiPSC)-derived neurons from AP4-HSP patients, we defined a series of 5 novel compounds that restore neuronal phenotypes of AP-4-deficiency. In a comprehensive multiparametric analysis, a novel small molecule, termed C-01, emerged as a lead compound with an EC50 of ~ 5μM. Target deconvolution strategies using transcriptomic and proteomic profiling revealed that C-01 modulates intracellular vesicle trafficking and increases autophagic flux, potentially through differential expression of several RAB (Ras-associated binding) proteins. Our findings demonstrate the ability of carefully designed high-throughput screens to identify molecular targets for AP-4 deficiency and support the development of C-01 as a novel therapeutic for AP-4-HSP. RESULTS Primary screening of 28,864 compounds in fibroblasts from AP-4-HSP patients identifies 503 active compounds A diversity library of 28,864 novel small molecules was provided by Astellas Pharma Inc. and arrayed in 384-well microplates. The primary screen was conducted in fibroblasts from a well-characterized patient with core clinical features of SPG47 8 and bi-allelic loss-of-function variants in AP4B1 (NM_001253852.3: c.1160_1161del (p.Thr387ArgfsTer30) / c.1345A > T (p.Arg449Ter)) (Fig. 1a,b). Fibroblasts from the sex-matched parent (unaffected heterozygous carrier) served as controls. The assay was fully automated, miniaturized to 384-well microplates, and compounds were added for 24h at a single concentration of 10μM (Fig. 1c). The ATG9A ratio (ATG9A fluorescence intensity inside the TGN vs. in the cytoplasm) was used as the primary assay metric, as established previously 15, 21. The population distributions of the subcellular ATG9A signal inside and outside the TGN, at the level of single cells for negative (bi-allelic loss-of-function, LoF/LoF) and positive (heterozygous carriers, WT/LoF) controls are shown in Fig. 1d and 1e. ATG9A ratios demonstrated symmetrical and approximately normal distributions and robust separation of both groups (Fig. 1f). Cell counts were similar for positive and negative controls, excluding cell death or changes in proliferation rates as possible confounding factors (Fig. 1g). To test for reproducibility across replicates, assay plates were randomly sampled into two sets, and similar positions on the assay plates were plotted against each other (Fig. 1h,i). Random sampling was simulated 100 times, and mean correlation coefficients were calculated. Using the ATG9A ratio (Fig. 1i) as a primary readout resulted in higher replicate correlation (mean r = 0.90 ± 0.002 SD), compared to absolute ATG9A intensities (Fig. 1h) (mean r = 0.82 ± 0.0008 SD). ATG9A ratios showed robust discriminative power between positive and negative controls (LoF/LoF mean: 1.1 ± 0.02 SD, n = 1312 wells vs. WT/LoF mean: 1.34 ± 0.05 SD, n = 1312 wells; Mann-Whitney U test, p < 0.0001) (Fig. 1j). The ATG9A ratio as the primary outcome metric was further supported by a generalized linear model, which demonstrated high specificity and sensitivity (Fig. 1k, AUC: 0.96). Source data for assay performance are provided in Supplementary File 1. Throughout the screen, assay performance was monitored using established quality control metrics for cell-based screens (Z’ robust ≥ 0.3, strictly standardized median difference ≥ 3, and an inter-assay coefficient of variation ≤ 10%) 30, 31, 32. All assay metrics were calculated for positive and negative controls of the same assay plate to avoid bias by inter-plate variability. Predefined thresholds were met by all assay plates (Supplementary Fig. 1a and Supplementary File 2). The results of the primary screen are summarized in Fig. 1l and 1m, and the complete dataset is provided in Supplementary File 3. Of the 28,838 compounds, 26 were excluded due to non-quantifiable ATG9A signal, exceptionally low cell counts or imaging artifacts. The remaining 28,812 compounds were evaluated for changes in cell count and the ATG9A ratio. The vast majority (n = 26,961, 93.5%) did not show any significant reduction in the ATG9A ratio (defined as a reduction by at least 3 SD). 1,435 (5.0%) compounds were excluded due to toxicity, defined as a reduction in the mean cell count by at least 2 SD compared to the negative controls. Only a small subset of 503 compounds (1.7%) reduced the ATG9A ratio by 3 or more SD compared to negative controls (Fig. 1m). Of these, 61 (0.2%) also reduced cell counts, while the remaining 442 (1.5%) showed no toxicity. In summary, from this high-throughput primary screen, 503 active compounds were identified and selected for further testing. Counter-screen in fibroblasts from AP-4-HSP patients confirms 16 compounds that lead to a dose-dependent redistribution of ATG9A To validate the 503 active compounds identified in the primary screen, compounds were retested for dose-dependency using an 11-point dose range (range: 40nM to 40μM) (Fig. 2a). Source data for the secondary screen are provided in Supplementary File 4. All concentrations were screened in biological duplicates and subjected to the same quality control metrics as in the primary screen (Supplementary Fig. 1b and Supplementary File 5). Similar to the results from the primary screen, ATG9A ratios for negative and positive controls showed a robust separation (LoF/LoF mean: 1.4 ± 0.07 (SD), n = 269 wells, vs. WT/LoF mean: 1.12 ± 0.02 (SD), n = 269 wells, Mann-Whitney U test, p < 0.0001, Fig. 2b). Activity in the secondary screen was defined as the ability to reduce the ATG9A ratio by at least 3 SD in both replicates and at least 2 different concentrations, without exerting toxicity. 51 compounds (10.1%) met these a priori defined criteria (Supplementary Fig. 2a,b). After manually verifying image quality and validating dose-response relationships, compounds were triaged (Fig. 2a and Supplementary Fig. 2a,b). Seventeen compounds demonstrated a clear and reproducible dose-response relationship, without evidence of image artifacts or autofluorescence. The EC50 for most compounds were in the low micromolar range (median: 4.66μM, IQR: 8.63, Fig. 2). 34 compounds were found to carry autofluorescence or imaging artifacts and were thus excluded from further testing (Supplementary Fig. 2b). One active compound was unavailable from the manufacturer and was removed. In summary, a counter-screen in AP-4-deficient patient fibroblasts confirmed and established dose-dependent effects on intracellular ATG9A distribution for 16 compounds (Fig. 2c). Orthogonal assays in neuronal models of AP-4-deficiency confirm 5 active compounds To validate active compounds from the secondary screen in a human cell line with neuron-like properties, the ATG9A assay was optimized for neuroblastoma-derived SH-SY5Y cells following a 5-day neuronal differentiation protocol with retinoic acid 33 (Fig. 3a). SH-SY5Y cells with stable expression of a AP4B1-targeting CRISPR/Cas9 machinery (AP4B1KO) 12 served as negative controls while AP4B1-wildtype (AP4B1WT) cells were used as positive controls. All 16 active compounds were tested in an 8-point dose range (50nM to 30μM) with a treatment duration of 24h. Quantification of the ATG9A ratio in differentiated SH-SY5Y cells showed a robust separation between control conditions (AP4B1KO: 1.80 ± 0.06 (SD), n = 158 wells vs. AP4B1WT: 1.17 ± 0.03 (SD), n = 160 wells, Mann-Whitney U test, p < 0.0001, Fig. 3b, Supplementary File 6). Compounds were evaluated based on their dose-dependent reduction of the ATG9A ratio and absence of cell toxicity. Eleven of 16 compounds were excluded due to lacking activity (n = 7), suspicion for artefacts or autofluorescence (n = 3), or obvious changes in cellular morphology (n = 1) (Supplementary Fig. 3). Of the five remaining compounds, three restored the ATG9A ratio to levels of wildtype controls (F-01, G-01 and H-01) while two compounds (B-01 and C-01) led to a reduction by at least 3 SD at higher concentrations (Fig. 3c–h). To assess whether these effects were specific to ATG9A or similar effects were also present for other AP-4 cargo proteins, we turned to a second neuronal AP-4 cargo protein, DAGLB 29. Similar to ATG9A, the DAGLB ratio (DAGLB fluorescence intensity in the TGN vs. in the cytoplasm) showed a robust separation between AP4B1WT and AP4B1KO cells (AP4B1KO: 1.80 ± 0.1 (SD), n = 192 wells vs. AP4B1WT: 1.36 ± 0.07 (SD), n = 192 wells, Mann-Whitney U test, p < 0.0001, Fig. 3i, Supplementary File 6). All five active compounds showed activity in the DAGLB assay, suggesting a broader effect on the trafficking of at least 2 AP-4 cargo proteins from the TGN. Again, F-01, G-01 and H-01 (Fig. 3l–o) resulted in normalization of the intracellular DAGLB distribution, while B-01 and C-01 led to a moderate reduction of DAGLB ratios at higher concentrations (Fig. 3j,k,o). Since small molecules can have pleotropic effects on cellular functions and organellar morphology, we adapted a multiparametric morphological profiling approach 34. Eighty-five measurements of the nucleus, cytoskeleton, global cell morphology, the TGN, and ATG9A vesicles were automatically computed for each image, serving as a rich and unbiased source for interrogating biological perturbations induced by compound treatment (Supplementary File 6). Principal component analysis was used to reduce dimensionality and cluster images based on their properties (Fig. 4a and Supplementary Fig. 4). Positive and negative controls clustered closely together and were separated only by the ATG9A signal (Fig. 4b and Supplementary Fig. 4a). B-01, C-01 and G-01 showed properties comparable to positive and negative controls, suggesting little off-target effects (Fig. 4b Supplementary Fig. 4b,c,e). F-01 and H-01, however, changed cellular morphology in a dose-dependent manner (Fig. 4b and Supplementary Fig. 4d,f), with changes mainly driven by the first principal component, accounting for 31.1% of the observed variance (Fig. 4c). To decipher the phenotypic alterations responsible for these changes, the Pearson correlation coefficients of the first principal component with each measurement were calculated (Fig. 4d). Features with a correlation coefficient > 0.75 were selected to define morphological profiles (Fig. 4e). Interestingly, TGN fluorescence intensity and morphology seemed to be the most significant drivers for the separation, suggesting that disruption of TGN integrity potentially biased the assessment of ATG9A ratios in cells treated with compounds F-01 and H-01 (Fig. 4b and Supplementary Fig. 4d,f). Following these analyses, TGN fluorescence intensity and morphological measures such as TGN area and elongation, as well as compactness and roughness, as indicators of the complexity of the TGN, were quantified for cells treated with all five active compounds (Fig. 4f,g). While C-01 showed stable TGN signal and morphology across all assessed measurements, the other compounds depicted some degree of change. Again, F-01 and H-01 seemed to result in TGN changes in a dose-dependent manner while B-01 and G-01 led to only moderate alterations (Fig. 4f,g). Of note, these changes to TGN morphology were not detectable by visual inspection but only delineated through an automated analysis of ~ 600 images containing ~ 30,000 cells per group, showcasing the power of our automated, unbiased, high-throughput platform. C-01 restores ATG9A and DAGLB trafficking in hiPSC-derived neurons from AP-4-HSP patients Informed by the findings in differentiated AP4B1KO SH-SY5Y cells, we next investigated whether these results would translate to human neurons. hiPSCs from patients with AP-4-HSP due to biallelic loss-of-function variants in AP4M1 (NM_004722.4: c.916C > T (p.Arg306Ter) / c.694dupG (p.Glu232GlyfsTer21)) and AP4B1 (NM_001253852.3: c.1160_1161del (p.Thr387ArgfsTer30) / c.1345A > T (p.Arg449Ter)) were generated 35, 36 and differentiated into glutamatergic cortical neurons using established protocols 15, 37, 38. hiPSC-derived neurons from sex-matched parents (unaffected heterozygous carriers) served as controls (Fig. 5a and Supplementary File 7). Baseline quantification of ATG9A ratios in DIV (day in vitro) 14 neurons treated with vehicle for 24h showed robust separation between patient and control lines, exceeding the differences observed in AP-4-deficient fibroblasts and differentiated SH-SY5Y cells (SPG50 patient mean: 4.31 ± 0.4 (SD), n = 60 wells vs. heterozygous control: 1.56 ± 0.12 (SD), n = 60 wells, Mann-Whitney U test, p < 0.0001, Fig. 5b). Neurons were treated for 24h in 8-point dose titration experiments. B-01 and G-01 lacked activity on the ATG9A ratio and were thus excluded (Fig. 5d). C-01, F-01 and H-01, by contrast, showed a robust reduction in the ATG9A ratio (Fig. 5e,f). A multiparametric analysis showed that, similar to observations in AP4B1KO SH-SY5Y cells, only C-01 preserved TGN integrity (Fig. 5f), while F-01 and H-01 impacted TGN morphology, suggesting off-target effects (Fig. 5e). Based on its favorable profile, C-01 was selected as a lead compound and was re-synthesized for further testing (Fig. 5g). Prolonged treatment of C-01 for 72h to test for ATG9A and DAGLB translocation, demonstrated that C-01 was able to restore ratios of both AP-4 cargo proteins to levels close to controls with an EC50 of ~ 5μM, while maintaining a favorable profile (Fig. 5h, Supplementary File 7). This greater effect on ATG9A distribution, compared to the ~ 50% reduction of the ATG9A ratio at 24h treatment, suggests a time- and dose-dependent effect. C-01 changed the ATG9A ratio through simultaneously decreasing ATG9A intensities inside the TGN and increasing cytoplasmic ATG9A levels, suggesting ATG9A translocation as the most likely mechanism of action. No changes in TGN morphology or any other cellular measurements were observed, indicating overall preservation of cellular morphology and little off-target effects. A similar pattern was observed with respect to DAGLB translocation (Fig. 5h). These findings were confirmed in a second set of experiments in hiPSC-derived neurons from a patient with SPG47 (Fig. 5i, Supplementary File 7), demonstrating that findings extend to other forms of AP-4-deficiency. Taken together, C-01 emerged as a robust modulator of ATG9A and DAGLB trafficking in human neurons from patients with AP-4 deficiency. Target deconvolution using transcriptomic and proteomic analyses delineates putative mechanisms of action for C-01 To explore potential mechanisms of action of C-01 in an unbiased manner, we used a multi-omics approach, combining bulk RNA sequencing and unbiased label-free quantitative proteomics (source data are provided in Supplementary Files 8–10). First, RNA sequencing was conducted in differentiated AP4B1WT and AP4B1KO SH-SY5Y cells treated for 72h with either vehicle or compound C-01 (5μM, Supplementary File 8). Analysis of differential gene expression identified few significant transcriptional changes in response to C-01 treatment, suggesting that this compound does not elicit major alterations in gene expression or induce many off-target effects (Supplementary Fig. 5). Since changes in gene expression caused by short-duration small molecule treatments might not reach predefined cut-offs for standard differential expression analyses, and because compounds might affect groups of genes in shared pathways rather that modifying single target genes, we adapted an unbiased and unsupervised network approach to identify groups of co-expressed genes. Hierarchical clustering of samples showed that treatment with C-01, regardless of cell line, was the main differentiator in our dataset (Fig. 6a). To identify the gene networks responsible for these changes, weighted gene co-expression network analysis (WGCNA) 39, 40 was used to group the 18,506 expressed genes into 36 co-expression modules (Fig. 6b). Gene expression profiles within each module were summarized using the “module eigengene” (ME), defined as the first principal component (PC) of a module 41. Within each module, the association of MEs with measured traits were examined by correlation analysis (Fig. 6c). Eight modules that showed an absolute correlation coefficient > 0.5 were selected for further evaluation. For these selected modules, ME based connectivity was determined for every gene by calculating the absolute value of the Pearson correlation between the expression of the gene and the respective ME, producing a quantitative measure of module membership (MM). Similarly, the correlation of individual genes with C-01 treatment was computed, defining gene significance (GS) for C-01. Using the GS and MM, an intramodular analysis was performed, allowing identification of genes that have high significance with treatment as well as high connectivity to their modules (Fig. 6d). Five modules were significantly related to C-01 treatment, defined as showing an absolute correlation coefficient between MM and GS > 0.5 (Fig. 6e). A list of the genes contained in each module along with their module membership is provided in Supplementary File 9. To summarize the biological information contained in these modules of interest, gene ontology (GO) analysis was performed, which demonstrated enrichment in biological pathways in three out of the five assessed modules (Fig. 6f). The ‘blue module’ showed down-regulation of pathways involved in axonogenesis, actin filament organization and proteasome-mediated pathways. The ‘light-yellow module’ contained genes involved in ER stress response, amino acid metabolism and transcription. Finally, the ‘mediumpurple3 module’ depicted upregulation of genes involved in vesicular transport, particularly involving TGN and ER-associated transport, as well as membrane and vesicle dynamics. This last module showed the highest gene ratios (defined as the percentage of total differentially expressed genes in the given GO term) and lowest P-values of all differentially regulated pathways across all modules, suggesting the upregulation of alternative vesicle mediated transport mechanisms by compound C-01 (Fig. 6f). To assess whether similar themes would emerge on the protein level, we next used unbiased quantitative proteomics in both differentiated SH-SY5Y cells (AP4B1KO and AP4B1WT) and hiPSC-derived neurons (patient with AP4B1-associated SPG47 and control) treated for 72h with either vehicle or compound C-01 (5μM). After quality filtering, 8,141 unique proteins in SH-SY5Y cells and 7,386 unique proteins in hiPSC-derived neurons were quantified. Differential enrichment analyses for both cell lines are shown in Fig. 7a,b, and source data are provided in Supplementary File 10. As expected, baseline quantification of differentially expressed proteins in AP4B1KO SH-SY5Y cells showed downregulation of AP-4 subunits, AP4B1, AP4E1 and AP4M1, and increased ATG9A levels, as reported in other models of AP-4 deficiency 11, 12, 13 (Supplementary Fig. 6a). PCA analysis of SH-SY5Y cells demonstrated 4 distinct clusters separated by C-01 treatment (PC1, explaining 12.3% of variance) and genotype (PC2, explaining 8.7% of variance) (Fig. 7a). Testing of vehicle vs. C-01 treated cells showed broadly similar groups of dysregulated proteins in AP4B1WT and AP4B1KO SH-SY5Y cells (Supplementary Fig. 6b-d), suggesting a conserved mechanism of action independent of genotype, which allowed the pooling of cell lines to increase the power of the analysis (Fig. 7a). Similar observations were made for hiPSC-derived neurons (Fig. 7b and Supplementary Fig. 6e-h). Here, cell lines were a stronger discriminator, likely due to heterogeneity of the positive and negative controls, as expected in cell lines derived from different individuals. Again, differentially enriched proteins following C-01 treatment in hiPSC-neurons showed a high degree of similarity between patient and control lines (Supplementary Fig. 6f-h), allowing a combined analysis (Fig. 7b). Despite the heterogeneity in the neuronal samples, significant overlap was observed between the differentially enriched proteins in SH-SY5Y cells and hiPSC-derived neurons. Data sets were thus integrated for a combined analysis, which detected several proteins that were dysregulated across all cell types and genotypes (Supplementary Fig. 6i-l), providing strong evidence that these changes were related to treatment with C-01 (Fig. 7c). Consistent with the overall changes in gene expression, pathway enrichment analysis using the Reactome database 42 highlighted engagement of intracellular trafficking pathways as a potential mechanism of action for C-01 (Fig. 7c). Specifically, modulation of RAB proteins involved in vesicle transport emerged as a consistent theme across cell types and genotypes, with the strongest evidence for the upregulation of RAB1B and downregulation of RAB3C and RAB12. Notably, while C-01 led to a significant change in protein levels of all three RAB protein family members in SH-SY5Y cells, only RAB3C and RAB12 reached significance in neurons (Fig. 7d). This overall pattern of RAB protein modulation was further supported by upregulation of the RAB protein geranylgeranyltransferase components A1 (CHM) in SH-SY5Y cells and A2 (CHML) in both SH-SY5Y cells and neurons. CHM and CHML play a vital role for tethering RAB proteins to intracellular membranes 43, 44. Additionally, upregulation of transferrin receptor protein 1 (TFRC) was observed (Fig. 7c), consistent with prior reports showing that reduction of RAB12 associates with increased protein levels of TFRC 45. Collectively, these findings suggest a potential role of RAB proteins in regulating vesicle transport in response to C-01 treatment. RAB3C and RAB12 knockout are involved in C-01 -mediated vesicle trafficking and autophagy RAB3C and RAB12 displayed the strongest and most consistent protein expression changes in both differentiated SH-SY5Y cells and hiPSC-derived neurons following treatment with C-01 (Fig. 7d) and were therefore selected for further investigation. Correlation analysis revealed a strong correlation (r = 0.93) between the LFQ intensities of these two proteins in both cell types and across different genotypes in response to C-01 (Fig. 7e). To assess whether a correlation was also present on the transcriptional level, mRNA levels of RAB3C and RAB12 in response to C-01 treatment were analyzed in AP4B1WT and AP4B1KO SH-SY5Y cells. While there was a trend toward a reduction of RAB3C and elevation of RAB12 mRNA levels and correlation analysis demonstrated a moderate inverse correlation, none of these changes reached statistical significance (Supplementary Fig. 7). These findings suggest that RAB3C and RAB12 levels are altered through a post-transcriptional mechanism following treatment with C-01. To investigate the potential impact of RAB3C and RAB12 on ATG9A translocation in the AP-4-deficient background, we used CRISPR/Cas9-mediated knockouts of RAB3C and RAB12 in AP4B1KO SH-SY5Y cells (Fig. 8a,b, Supplementary Fig. 8 and Supplementary File 11). We found that knockout of RAB12 did not affect ATG9A translocation, while knockout of RAB3C caused a moderate reduction in the ATG9A ratio (Fig. 8a). Combined knockout of RAB3C and RAB12 in AP4B1KO SH-SY5Y cells did not show an additive effect. Interestingly, however, the effects of C-01 on ATG9A translocation were significantly enhanced by knockout of RAB3C, but not RAB12 alone. Combined knockout of both genes further augmented the effect of C-01. These findings suggest that both RAB3C alone, or in combination with RAB12, play a role in C-01-mediated ATG9A redistribution. A converging theme of ATG9A translocation and alteration of RAB protein expression is autophagy. RAB proteins are known modulators of autophagy with key functions in various steps of the pathway 46, 47. ATG9A, a core autophagy protein, acts as a lipid scramblase and promotes autophagosome formation and elongation 48, 49, 50, 51. To investigate whether C-01 leads to changes in autophagic flux, AP4B1WT and AP4B1KO SH-SY5Y cells were treated with C-01 for 72h and LC3-I to LC3-II conversion was measured by western blotting (Fig. 8c–f and Supplementary Fig. 8a). Levels of LC3-II were significantly elevated in all cell lines treated with C-01, suggesting modulation of the autophagy pathway. Co-treatment with bafilomycin A1, which blocks autophagosome-lysosome fusion, led to further LC3-II accumulation, indicating that C-01 increases autophagic flux (Fig. 8c–f). Blocking the late stages of the autophagy pathway, with either bafilomycin A1 or chloroquine, reversed the effect of C-01 on ATG9A translocation in a dose-dependent manner, suggesting that this process requires intact autophagic flux (Fig. 8g–i). Next, since our data suggested a contribution of RAB3C and RAB12 to the effect of C-01, we investigated the impact of RAB3C and RAB12 knockout in AP4B1KO SH-SY5Y cells with and without C-01 treatment (Fig. 8j–l and Supplementary Fig. 8b-d). Neither RAB3C nor RAB12 knockout alone led to major changes in baseline or C-01-enhanced autophagic flux (Fig. 8j,k). However, combined knockout of RAB3C and RAB12 significantly increased the ratio of LC3-II to LC3-I by approximately 36% (Fig. 8l). Upon treatment with bafilomycin A1, both RAB3C knockout alone and combined knockout of RAB3C and RAB12 further increased C-01-mediated LC3-I to LC3-II conversion (Fig. 8j–l). These findings suggest the possibility that RAB3C and RAB12 modulate C-01-mediated ATG9A trafficking and subsequent autophagy induction. DISCUSSION Identification of novel therapeutic targets for rare neurological diseases represents a major scientific and public health challenge 1, 4. The increasing number of rare genetic diseases 52, the rising rate of diagnoses 53, and the significant burden for patients 54, 55, caregivers 56 and health care systems 57 highlight the urgent need for translational research that moves beyond gene discovery to the identification of disease mechanisms and therapies. Unbiased high-content small molecule screens are a platform for drug-repurposing approaches and a starting point for the rationale development of new compounds 1, 2, 3, 4, 5, 6. Disease-relevant ‘screenable’ phenotypes across cellular models, including patient-derived cells, provide an entry point into developing automated, high-content screening and analysis platforms. In this study, we develop the first high-throughput cell-based phenotypic screening platform for a prototypical form of childhood-onset HSP caused by defective protein trafficking. Our platform allows us to determine the subcellular localization of the AP-4 cargo protein ATG9A in several cellular models of AP-4-deficiency. The hypothesis that ATG9A mislocalization is a key mechanism in the pathogenesis of AP-4-HSP is supported by the independent work of the Robinson 12, Kittler 14 and Bonifacino 11, 13, 58 groups, in addition to our own work 15, 21, 24, 25, and by the overlapping phenotypes of AP-4 13, 14, 26 and Atg9a28 knockout mice. ATG9A is the only conserved autophagy-related transmembrane protein 50 and in mammalian cells cycles between the TGN and ATG9A vesicles, which associate with endosomes 59 and autophagosome formation sites 59, 60. ATG9A has 4 transmembrane domains and forms homotrimers that have lipid scramblase activity 48, 49, 50, postulated to equilibrate lipids in the double-membrane layer of nascent autophagosomes 61, 62. Basal levels of autophagy are essential for neuronal survival, and neuron-specific ablation of the autophagy pathway leads to axonal degeneration and cell death 63, 64, 65. In neurons, autophagosomes form in the distal axon 66, 67 and are subject to active transport 68, 69, 70. Thus, efficient vesicular trafficking and spatial distribution of ATG9A are essential for axonal function as demonstrated in CNS-specific Atg9a knockout mice 28. Having established a robust and dynamic assay that reliably measures intracellular ATG9A distribution, we systematically screened a large library of 28,864 novel small molecules for their ability to restore ATG9A trafficking from the TGN to the cytoplasm. Following this primary screen, a counter-screen and a series of orthogonal experiments identified a novel small molecule, termed C-01, that can restore the intracellular distribution of ATG9A and a second transmembrane AP-4 cargo protein, DAGLB, in neuronal models of AP-4 deficiency, including iPSC-derived neurons from two patients with AP-4-HSP. Compound C-01 has physicochemical properties that are within the parameters that are optimal for CNS drugs 71 and therefore represents a strong candidate for an in vivo tool compound. In addition, the low molecular weight and topological polar surface area create opportunities for compound optimization. Since the molecular targets of C-01 are unknown, we employed a target deconvolution strategy using transcriptomics and proteomics to define the cellular pathways impacted by this novel small molecule. This approach identified two central themes: 1) modulation of Golgi dynamics and vesicular trafficking, and 2) engagement of autophagy. At the core of the putative pathways affected by C-01, we identified the Rab proteins RAB1B, RAB3C and RAB12, as well as the interacting Rab geranyl transferase subunits CHM and CHML. RAB3C and RAB12 showed the strongest and most consistent association with C-01 treatment in both SH-SY5Y cells and iPSC-derived neurons, and our analyses suggest that these two proteins are involved in C-01-mediated redistribution of ATG9A from the TGN and increase of autophagic flux. Rab proteins comprise a large family of small guanosine triphosphate (GTP) binding proteins that act as key regulators of intracellular membrane trafficking in eukaryotic cells at several stages, including cytoplasmic cargo sorting, vesicle budding, docking, fusion and membrane organization 72, 73. Rab GTPases function both as soluble and specifically localized, integral-membrane proteins, the latter being mediated by prenylation. Among the roughly 70 known Rab proteins, more than 20 are primarily associated with the TGN, where they regulate Golgi organization, coordinate vesicle trafficking and interact with various steps of the autophagy pathway 46, 47. Following treatment with C-01, the RAB protein family members RAB3C and RAB12 were consistently downregulated in both SH-SY5Y cells and iPSC-derived neurons. Knockout experiments of these two proteins revealed that their loss potentiates C-01-mediated ATG9A translocation and autophagic flux. RAB3C, which is part of the RAB3 superfamily, is primarily expressed in brain and endocrine tissues, where it localizes to the Golgi and synaptic vesicles and is involved in exocytosis and modulation of neurotransmitter release 74. RAB12 is mainly localized to recycling endosomes where it regulates endosomal trafficking and lysosomal degradation and has been identified as a modulator of autophagy 75. A well-known downstream target of RAB12 is the transferrin receptor (TfR). Knockdown of RAB12 in mouse embryonic fibroblasts increases TfR protein levels, while overexpression leads to its reduction 45. In line with this, we find that treatment with C-01 reduced RAB12 protein levels while, at the same time, robustly elevating transferrin receptor protein 1 (TFRC). To the best of our knowledge, no interaction between RAB3C and RAB12 has been described so far, however, our data suggest that both proteins are involved in C-01-mediated modulation of vesicle trafficking and autophagic flux. Our study has identified the first candidate small molecule drug capable of restoring protein mislocalization in AP-4-deficient cells, including human neurons from patients. We acknowledge several limitations of our approach, some of which are inherent to high-throughput screens and some that are specific to our assay. First, as ATG9A mislocalization is a cellular phenotype of AP-4 deficiency conserved in non-neuronal and neuronal cells both in vitro11, 12, 13, 14, 15, 25, 76 and in vivo13, 14, 27, we decided to conduct the initial screen in patient-derived fibroblasts, as a simple cellular model of AP-4 deficiency. While the use of patient fibroblasts in the primary screen increases translational relevance, compounds that would have the capacity to correct ATG9A trafficking exclusively in neuronal cells could be missed at this stage. We determined that this risk was outweighed by the benefits of a robust assay performance and the fact that mechanisms of AP-4-mediated protein trafficking are conserved across tissues and cell types 11, 12, 13, 14, 15, 35, 76. Second, even though cell-based disease models can, to some extent, mimic the complexity of therapeutic responses in biological systems, the translation to in vivo models is often challenging, particularly for neurodevelopmental and neurodegenerative disease. Considerations such as a lead compound’s ability to cross the blood-brain-barrier, target engagement in the central nervous system, therapeutic responses in complex neuronal networks relying on interactions with glia cells, developmental windows amenable to therapy, as well as in vivo off-target effects and toxicity must be considered and explored in future studies. To mitigate some of these risks, we employed unbiased multiparametric profiling of C-01 which suggested little off-target effects. Future studies are required to exclude pleiotropic effects or off-target toxicity in different cell types or tissues in vivo. Lastly, while C-01 leads to a redistribution of two well-established AP-4 cargo proteins, ATG9A and DAGLB, we are unable to exclude the possibility that other neuron-specific cargos of AP-4 exist and are important for the pathogenesis of AP-4-HSP. Nonetheless, mislocalization of both proteins is proposed as the major contributor to neuronal pathology caused by AP-4 deficiency, through dysregulated autophagy and endocannabinoid signaling, respectively 11, 12, 13, 14, 29. Our automated high-throughput platform would allow for the rapid interrogation of additional AP-4 cargo proteins in the future. In conclusion, our findings provide a solid foundation for lead optimization of C-01 and development in Investigational New Drug (IND)-enabling studies. More broadly, our approach illustrates the development of a small molecule screening platform for a rare neurogenetic disease, leveraging robust cellular phenotypes. We hope this approach will create a paradigm for other rare and more common disorders of protein trafficking. The increase of autophagic flux through C-01 offers the intriguing possibility that this compound could be considered for the treatment of other autophagy-associated diseases. METHODS Clinical data from patients with AP-4-HSP This study was approved by the Institutional Review Board at Boston Children’s Hospital (IRB-P00033016 and IRB-P00016119). Two patients with AP-4-HSP and their clinically-unaffected, sex-matched parents were enrolled in the International Registry and Natural History Study for Early-Onset Hereditary Spastic Paraplegia (ClinicalTrials.gov Identifier: NCT04712812). Both patients had a clinical and molecular diagnosis of AP-4-HSP and presented with core clinical and imaging features 8. Patient 1 was diagnosed with AP4B1-associated SPG47 and carries the following compound-heterozygous variants: NM_001253852.3: c.1160_1161del (p.Thr387ArgfsTer30) / c.1345A > T (p.Arg449Ter). The sex-matched parent carries the heterozygous c.1160_1161del; p.Thr387Argfs*30 variant. Patient 2 was diagnosed with AP4M1-associated SPG50 and carries the following compound-heterozygous variants: NM_004722.4: c.916C > T (p.Arg306Ter) / c.694dupG (p.Glu232GlyfsTer21). The sex-matched parent carries the heterozygous c.694dupG (p.Glu232GlyfsTer21) variant. Antibodies and reagents The following reagents were used: Bovine serum albumin (AmericanBIO, Cat# 9048-46-8), saponin (Sigma, #47036-50G-F), normal goat serum (Sigma-Aldrich, Cat# G9023–10ML), Dulbecco’s phosphate-buffered saline (DPBS) (Thermo Fisher Scientific, Cat# 14190–250), trypsin (Thermo Fisher Scientific, Cat#25200056), 4% paraformaldehyde (4%) (Boston BioProducts, Cat# BM-155), dimethyl-sulfoxide (DMSO) (American Bioanalytical, Cat# AB03091–00100), bafilomycin A1 (Enzo Life Sciences, Cat# BML-CM110-0100), chloroquine (MedChemExpress, Cat# HY-17589A), Molecular Probes Hoechst 33258 (Thermo Fisher Scientific, Cat# H3569) and Alexa Fluor 647-labelled phalloidin (Thermo Fisher Scientific, Cat#A22287). The following primary antibodies were used: Anti-AP4E1 at 1:500 (BD Bioscience, Cat# 612019), anti-ATG9A at 1:500–1000 (Abcam, Cat# ab108338), anti-DAGLB at 1:500 (Abcam, Cat# 191159), anti-TGN46 at 1:800 (Bio-Rad, Cat# AHP500G), anti-Golgi 97 1:500 (Abcam, Cat# 169287), anti-beta-Tubulin III 1:1000 (Synaptic Systems, Cat# 302304 and Sigma, Cat# T8660), anti-beta-Actin 1:10,000 (Sigma, Cat# A1978–100UL), pan-AKT (Cat# 4691), anti-RAB12 (Santa Cruz, Cat# sc-515613), anti-RAB3C (Santa Cruz, Cat# 107 203), anti-LC3B 1:1000 (Novus, Cat#100–2220). Fluorescently labelled secondary antibodies for immunocytochemistry were used at 1:2000 (Thermo Fisher Scientific, Cat# A11005, A-11008, A-11016, A-11073, A-21235, A-21245), for western blotting at 1:5000 (LI-COR Biosciences, Cat# 926–68022, 926–68023, 926–32212, 926–32213). Small molecule library A diversity small molecule library containing 28,864 novel compounds was provided by Astellas Pharma Inc.. Compounds were arrayed in 384-well microplates at a final concentration of 10mM (1000-fold the screening concentration) in DMSO. Assay plates were stored at −80°C and thawed 30 min prior to cell plating. Active compounds from the primary screen were re-screened in a secondary screen, using eleven-point concentrations (range: 0.04μM, 0.08μM, 0.16μM, 0.31μM, 0.63μM, 1.25μM, 2.5μM, 5μM, 10μM, 20μM, 40μM) in two biological replicates. The chemical structure of the lead compound was disclosed by Astellas Pharma Inc. after the screen was completed. Fibroblast cell culture Fibroblast lines were established from routine skin punch biopsies in both patients and their respective sex-matched heterozygous parents 15. Primary human skin fibroblasts were cultured and maintained as previously described 77. Briefly, cells were maintained in DMEM high glucose (Gibco, #11960044) supplemented with 20% FBS (Gibco, #10082147), penicillin 100U/ml and streptomycin 100μg/ml (Gibco, #15140122). Cells were kept in culture for up to 8 passages and routinely tested for the presence of mycoplasma contamination. For high-throughput imaging, fibroblasts were seeded onto 384-well plates (Greiner Bio-One, #781090) at a density of 2 × 103 per well using the Multidrop Combi Reagent Dispenser (Thermo Fisher Scientific, #11388–558). Media changes were done every 2–3 days and drugs administered 24h before fixation. SH-SY5Y cell culture AP4B1 wildtype (AP4B1WT) and AP4B1 knockout (AP4B1KO) SH-SY5Y cells were generated previously 12. Undifferentiated SH-SY5Y cells were maintained in DMEM/F12 (Gibco, Cat# 11320033) supplemented with 10% heat-inactivated fetal bovine serum (Gibco, Cat# 10438026), 100U/ml penicillin and 100μg/ml streptomycin at 37°C under 5% CO2. SH-SY5Y cells were passaged every 2–3 days and differentiated into a neuron-like state using a 5-day differentiation protocol with all-trans-retinoic acid (MedChemExpress, #HY-14649) as described previously 33. For assessment of ATG9A translocation, differentiated SH-SY5Y cells were plated in 96-well plates (Greiner Bio-One, Cat# 655090), at a density of 1 × 104 cells per well. Media changes were done every 2–3 days and drugs administered 24–72h before fixation. Generation of hiPSC lines and neuronal differentiation Fibroblasts were reprogrammed to hiPSCs using non-integrating Sendai virus as described previously 35, 36. Quality control experiments including karyotyping, embryoid body formation, pluripotency marker expression, STR profiling and Sanger sequencing for AP4B1 or AP4M1 variants were reported previously 35, 36. hiPSC-derived neurons were generated using induced NGN2 expression following published protocols with minor modifications 37, 38. hiPSCs were dissociated into single cells with accutase (Innovative Cell Technology, Cat#AT 104–500) and seeded onto Geltrex-coated plates (Thermo Fisher Scientific, Cat#A1413301). hiPSCs were then infected with concentrated rtTA-, and NGN2-expressing lentiviruses (FUW-M2rtTA Addgene #20342, pTet-O-Ngn2-puro Addgene #52047), in the presence of polybrene (8 μg/ml, Sigma-Aldrich, Cat# TR-1003-G). The next day, hiPSCs were fed with supplemented mTeSRPlus and expanded for cryopreservation. In parallel, a kill curve was generated to determine the optimal puromycin concentration needed to eliminate untransduced cells. Successful transduction was established by adding doxycycline (2 μg/ml, Millipore, Cat#324385–1GM) to virus-treated cells for 24 hours, followed by adding the optimized puromycin concentration (Invitrogen, 1 μg/ml, Cat# ant-pr-1) for up to 48 hours. For the generation of glutamatergic neurons, NGN2 transduced hiPSCs were dissociated into single cells using accutase and seeded onto geltrex-coated plates. The following day, NGN2 expression was induced using doxycycline and selected with puromycin. Growth factors BDNF (10 ng/ml, Peprotech, Cat#450–02), NT3 (10 ng/ml, Peprotech, Cat# 450–03), and laminin (0.2 mg/L, Thermo Fisher Scientific, Cat#23017–015) were added in N2 medium for the first 2 days. Cells were then fed with BDNF (10 ng/ml), NT3 (10 ng/ml), laminin (0.2 mg/Lf), doxycycline (2 μg/ml), and Ara-C (4uM, Sigma-Aldrich, Cat# C1768) in B27 media every other day until differentiation day 6. On day 6, cells were dissociated with papain (Worthington, Cat# LK003178) and DNaseI (Worthington, Cat# LK003172) and replated on poly-D-lysine (0.5mg/ml; Sigma Aldrich, Cat#P6407) and laminin (5μg/ml; Thermo Fisher Scientific, Cat #23017–015) coated plates with or without hiPSC-derived astrocytes (Astro.4U, Ncardia). For assessment of ATG9A translocation, neurons were plated in 96-well plates at a density of 4 × 104 cells per well. Media changes were done every 2–3 days and drugs administered 24–72h before fixation. Immunocytochemistry The immunocytochemistry workflow was optimized for high-throughput using automated pipettes and reagent dispensers (Thermo Fisher Scientific Multidrop Combi Reagent Dispenser, Integra VIAFLO 96/384 liquid handler, Integra VOYAGER pipette). Fibroblasts and SH-SY5Y cells were fixed using 3% and 4% PFA, respectively, permeabilized with 0.1% saponin in PBS and blocked in 1% BSA/0.01% saponin (blocking solution) in PBS. iPSC-derived neurons were fixed in 4% PFA, and permeabilized and blocked using 0.1% Triton X-100/2% BSA/0.05% NGS in PBS. Primary antibody (diluted in blocking solution) was added for 1h (fibroblasts and SH-SY5Y cells) at room temperature or overnight (iPSC neurons) at 4°C. Plates were gently washed three times in blocking solution (fibroblasts and SH-SY5Y cells) or in PBS (iPSC neurons), followed by addition of fluorochrome-conjugated secondary antibodies, Hoechst 33258 and phalloidin for 30min (fibroblasts) or Hoechst 33258 for 60min (SH-SY5Y cells and iPSC neurons) at room temperature. Plates were then gently washed three times with PBS and protected from light. High-content imaging and automated image analysis High-throughput confocal imaging was performed on the ImageXpress Micro Confocal Screening System (Molecular Devices) using an experimental pipeline modified from the pipeline described in Behne et al.15. For experiments in fibroblasts, images were acquired using a 20x S Plan Fluor objective (NA 0.45 μM, WD 8.2–6.9 mm). Per well, 4 fields were acquired in a 2×2 format (384-well plates). For experiments in SH-SY5Y cells and iPSC neurons, up to 36 fields were acquired in a 6×6 format (96-well plate) using a 40x S Plan Fluor objective ((NA 0.60 μm, WB 3.6–2.8 mm). The image analysis was performed using a customized image analysis pipeline in MetaXpress (Molecular Devices): Briefly, cells were identified based on the presence of DAPI signal inside a phalloidin (fibroblasts) or TUBB3 (SH-SY5Y cells and hiPSC-neurons)-positive cell body. Sequential masks were generated for (1) the TGN by outlining the area covered by TGN marker TGN46 (TGN46-positive area, in fibroblasts and SH-SY5Y cells) or Golgi 97 (Golgi 97-positive area, in hiPSC neurons) and (2) for the cell area outside the TGN (actin-positive area minus TGN46-positive area). ATG9A fluorescence intensity was measured in both compartments in each cell and the ATG9A ratio was calculated by dividing the ATG9A fluorescence intensity the TGN by the ATG9A fluorescence intensity in the remaining cell body (Fig. 1b): ATG9ARatio=ATG9AFluoresceneIntensity(F.U.)insidetheTGNATG9AFluoresceneIntensity(F.U.)outsidetheTGN Additional masks for the TGN used for morphologic profiling included TGN Roughness (shape factor in the MetaXpress software) and the following calculated metrics: TGNElongation=TGNWithTGNLength TGNCompactness=(TGNPerimeter)24π*TGNArea Z’-factor robust values and strictly standardized median difference (SSMD) 30 were calculated for each plate and only plates that met the predefined quality metrics of a Z’-factor robust ≥ 0.3 and SSMD ≥ 3 were included in subsequent analyses. Western blotting Western blotting was done as previously described 70. Briefly, cells were lysed in RIPA lysis buffer (Thermo Fisher Scientific Cat# 89900) supplemented with cOmplete protease inhibitor (Roche Cat# 04693124001) and PhosSTOP phosphatase inhibitor (Roche Cat# 4906845001). Total protein concentration was determined using a Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Cat# 23225). Equal amounts of protein were solubilized in LDS sample buffer (Thermo Fisher Scientific, Cat# NP0008) under reducing conditions, separated by gel electrophoresis, using 4–12% (Thermo Fisher Scientific, Cat# NW04125BOX) or 12% Bis-Tris gels (Thermo Fisher Scientific, Cat# NP0343BOX) and MOPS or MES buffer (Thermo Fisher Scientific, #NP0001 and #NP0002) and transferred to a PVDF or nitrocellulose membranes (EMD Millipore, #SLHVR33RS). Following blocking with blocking buffer (LI-COR Biosciences, #927–70001), membranes were incubated overnight with the respective primary antibodies. Near-infrared fluorescent-labeled secondary antibodies (IR800CW, IR680LT; LI-COR Biosciences) were used and quantification was done using the Odyssey infrared imaging system and Empiria Studio Software (LI-COR Biosciences). Sample preparation for RNA extraction SH-SY5Y cells were differentiated with retinoic acid as described above and subsequently treated with compounds of interest for 72h, prior to lysis using Qiagen RTL-Buffer supplemented with 1% ß-mercaptoethanol. RNA extraction, library preparation and sequencing were conducted at Azenta Life Sciences (South Plainfield, NJ, USA). Total RNA was extracted from frozen cell pellet samples using Qiagen RNeasy mini kit following manufacturer’s instructions (Qiagen, Cat# 74004). Library preparation with polyA selection and Illumina sequencing RNA samples were quantified using Qubit 4 Fluorometer (Life Technologies) and RNA integrity was checked using Agilent TapeStation 4200 (Agilent Technologies). RNA sequencing libraries were prepared using the NEBNext Ultra II RNA Library Prep Kit for Illumina using manufacturer’s instructions (New England Biolabs). Briefly, mRNAs were initially enriched with Oligod(T) beads. Enriched mRNAs were fragmented for 15 minutes at 94°C. First strand and second strand cDNA were subsequently synthesized. cDNA fragments were end repaired and adenylated at 3’ends, and universal adapters were ligated to cDNA fragments, followed by index addition and library enrichment by PCR with limited cycles. The sequencing library was validated on the Agilent TapeStation (Agilent Technologies), and quantified by using Qubit 4 Fluorometer (Invitrogen) as well as by quantitative PCR (KAPA Biosystems). The sequencing libraries were clustered on 3 lanes of a flowcell. After clustering, the flowcell was loaded on the Illumina instrument (HiSeq 4000 or equivalent) according to manufacturer’s instructions. The samples were sequenced using a 2×150bp Paired End (PE) configuration. Image analysis and base calling were conducted by the Control software. Raw sequence data (.bcl files) generated the sequencer were converted into fastq files and de-multiplexed using Illumina’s bcl2fastq 2.17 software. One mismatch was allowed for index sequence identification. Downstream RNA sequencing analysis Sequencing reads were mapped to the GRCh38 reference genome available on ENSEMBL using the STAR aligner v.2.7.9a. Differential expression analysis was done using the TREAT approach developed by McCarthy and Smyth 78, implemented in the edgeR package in R. Raw counts were obtained using STAR and low expressed genes were excluded using the method described by Chen et al. 79. Expression data were normalized using the Trimmed Mean of M-values method implemented in the edgeR package. Genes were considered as differentially expressed according to default options with a false discovery rate (Benjamini-Hochberg procedure) < 0.05 and a log2 fold change of > 0.3. Gene ontology (GO) enrichment analysis was done using clusterProfiler 80. Pathways were considered differentially expressed with an FDR < 0.05. Network connectivity analysis To identify transcriptional changes in co-expressed groups of genes following compound treatment, a weighted gene co-expression network analysis (WGCNA) was performed. Raw counts were generated, and low expressed genes were removed as described above. Data were normalized using variance stabilizing transformation as described by Anders et al.81. Batch effects were removed using the limma package in R 82. Preprocessed data were then analyzed using the WGCNA package in R 83, 84. In brief, pairwise Pearson correlations were calculated between all genes and genes with a positive correlation were selected to form a “directed” correlation matrix. Next, the correlations were raised to a power to approximate a scale free network. The adequate power was chosen based on soft thresholding aiming for a high scale independence above 0.8 by keeping a mean connectivity between 200 and 500. Genes were then grouped based on topological overlap and clusters were isolated using hierarchical clustering and adaptive branch pruning of the hierarchical cluster dendrogram, giving rise to groups of co-expressed genes, so called modules. Gene expression profiles within each module were summarized using the “module eigengene” (ME), defined as the first principal component of a module. Within each module, association of MEs with measured clinical traits was examined by correlation analysis. For these selected modules, module eigengene based connectivity was determined for every gene by calculating the absolute value of the Pearson correlation between the expression of the gene and the respective ME, producing a quantitative measure of module membership (MM). Similarly, the correlation of individual genes with the trait of interest was computed, defining gene significance (GS). Using the GS and MM, an intramodular analysis was performed, allowing identification of genes that have high significance with treatment as well as high connectivity to their modules. The biological information contained in modules of interest was summarized with gene ontology (GO) enrichment analysis using clusterProfiler 80. Pathways were considered differentially expressed with an FDR < 0.05. Sample preparation for mass spectrometry Cells were lysed for whole proteome analysis in RIPA lysis buffer (Thermo Fisher Scientific, Cat# 89900) supplemented with cOmplete protease inhibitor (Roche Cat# 04693124001) and PhosSTOP phosphatase inhibitor (Roche Cat# 4906845001) and sonicated in a Bioruptor® Pico Sonication System (one single 30 seconds on/off cycle at 4°C). Protein concentrations were determined using a Pierce BCA Protein Assay Kit (Thermo Fisher Scientific Cat# 23225). Lysates were stored at −80° C until further processing. To generate peptide samples for analysis by mass spectrometry, 30–50μg protein were precipitated by overnight incubation in 5 volumes of ice-cold acetone at − 20° C and pelleted by centrifugation at 10,000×g for 5 min at 4° C. All subsequent steps were performed at room temperature. Precipitated protein pellets were air-dried, resuspended for denaturation and reduction in digestion buffer (50 mM Tris pH 8.3, 8M Urea, 1 mM dithiothreitol (DTT)) and incubated for 15 min. Proteins were alkylated by addition of 5 mM iodoacetamide for 20 min in the dark. Following reduction and alkylation, proteins were enzymatically digested by addition of LysC (1 μg per 50 μg of protein; Wako, Cat# 129–02541) for an overnight incubation. Samples were then diluted four-fold with 50 mM Tris pH 8.3 before addition of Trypsin (1μg per 50μg of protein; Sigma-Aldrich, Cat# T6567) for 3 hours. The digestion reaction was stopped by addition of 1% (v/v) trifluoroacetic acid (TFA) and samples were incubated on ice for 5min to precipitate contaminants, which were pelleted by centrifugation at 10,000×g for 5min. Acidified peptides were transferred to new tubes, before purification by solid-phase extraction using poly(styrenedivinylbenzene) reverse-phase sulfonate (SDB-RPS; Sigma-Aldrich, Cat# 66886-U) StageTips 85. StageTips with three SDB-RPS plugs were washed with 100% acetonitrile, equilibrated with StageTip equilibration buffer (30% [v/v] methanol, 1% [v/v] TFA), and washed with 0.2% (v/v) TFA. 20μg of peptides in 1% TFA were then loaded onto the activated StageTips, washed with 100% isopropanol, and then 0.2% (v/v) TFA. Peptides were eluted in three consecutive fractions by applying a step gradient of increasing acetonitrile concentrations: 20μL SDB-RPS-1 (100 mM ammonium formate, 40% [v/v] acetonitrile, 0.5% [v/v] formic acid), then 20μL SDB-RPS-2 (150 mM ammonium formate, 60% [v/v] acetonitrile, 0.5% [v/v] formic acid), then 30μL SDB-RPS-3 (5% [v/v] NH4OH, 80% [v/v] acetonitrile). Eluted peptides were dried in a centrifugal vacuum concentrator, resuspended in Buffer A* (0.1% (v/v) TFA, 2% (v/v) acetonitrile), and stored at − 20° C until analysis by mass spectrometry. Mass spectrometry Mass spectrometry was performed on an Exploris 480 mass spectrometer coupled online to an EASY-nLC 1200, via a nano-electrospray ion source (all from Thermo Fisher Scientific). Per sample, 250 ng of peptides were loaded on a 50 cm by 75μm inner diameter column, packed in-house with ReproSil-Pur C18-AQ 1.9 μm silica beads (Dr Maisch GmbH). The column was operated at 50° C using an in-house manufactured oven. Peptides were separated at a constant flow rate of 300nL/min using a linear 110min gradient employing a binary buffer system consisting of Buffer A (0.1% [v/v] formic acid) and Buffer B (80% acetonitrile, 0.1% [v/v] formic acid). The gradient ran from 5 to 30% B in 84min, followed by an increase to 60% B in 8min, a further increase to 95% B in 4min, a constant phase at 95% B for 4min, and then a washout decreasing to 5% B in 5min, before re-equilibration at 5% B for 5min. The Exploris 480 was controlled by Xcalibur software (v.4.4, Thermo Fisher Scientific) and data were acquired using a data-dependent top-15 method with a full scan range of 300–1650 Th. MS1 survey scans were acquired at 60,000 resolution, with an automatic gain control (AGC) target of 3 × 106 charges and a maximum ion injection time of 25ms. Selected precursor ions were isolated in a window of 1.4 Th and fragmented by higher-energy collisional dissociation (HCD) with normalized collision energies of 30. MS2 fragment scans were performed at 15,000 resolution, with an AGC target of 1 × 105 charges, a maximum injection time of 28ms, and precursor dynamic exclusion for 30s. Raw mass spectrometry data analysis Mass spectrometry raw files were processed in MaxQuant Version 2.1.4.0 86, 87, using the human SwissProt canonical and isoform protein database, retrieved from UniProt (2022_09_26; www.uniprot.org). Label-free quantification was performed using the MaxLFQ algorithm 88. Matching between runs was enabled to match between equivalent and adjacent peptide fractions, within replicates. LFQ minimum ratio count was set to 1 and default parameters were used for all other settings. All downstream analyses were performed on the ‘protein groups’ file output from MaxQuant. Proteomic downstream data analysis Differential enrichment analysis of proteomics data was done using the DEP package in R. Preprocessing and quality filtering was performed separately for SH-SY5Y cells and hiPSC-derived neurons. Proteins that were only identified by a modification site, or matched the reversed part of the decoy database, as well as commonly occurring contaminants were removed. Duplicate proteins were removed based on the corresponding gene names by keeping those with the highest total MS/MS count across all samples. All following steps were done separately for each cell type (SH-SY5Y cells (Fig. 7a and Supplementary Fig. 6a-d) and hiPSC-derived neurons (Fig. 7b and Supplementary Fig. 6e-h) and for the pooled dataset (Fig. 7c and Supplementary Fig. 6i-l). Low quality entries were removed by keeping only those proteins that had valid MS/MS counts in all replicate samples of at least one experimental condition. Finally, only those proteins were kept that had a maximum of one missing LFQ value in at least one experimental condition. Filtered data were normalized using variance stabilizing transformation and missing values were imputed using a manually defined left-shifted Gaussian distribution with a width of 0.3 and a left-shift of 2.2 SD. Batch effects were corrected using the method described by Johnson et al.89. Statistical testing for differential protein enrichment was done using protein-wise linear models and empirical Bayes statistics implemented in the limma package in R. Proteins were considered as differentially enriched with a false discovery rate of < 0.05 and a log2 fold change > 0.3. The biological information contained in differentially enriched proteins was summarized using Reactome pathway annotation in clusterProfiler 80. Pathways were considered differentially expressed with an FDR < 0.05. Nucleofection sgRNAs against NLRP5, RAB3C and RAB12 were purchased as multi-guide knockout kits (v2) from Synthego, diluted to 100 μM stock concentrations and kept at −20°C. Nucleofection was performed under RNAse-free conditions on a Lonza 4D-Nucleofector (Cat# AAF-1003X, AAF-1003B) according to the manufacturer’s protocol. Briefly, SH-SY5Y cells were harvested, and 4 × 105 cells resuspended in 5 μl Nucleofector Solution. 180 pmol sgRNAs were incubated with 20 pmol Cas9 protein in Nucleofector Solution to form ribonucleoprotein complexes (RNPs) according to the manufacturer’s instructions. The cell solution was then incubated with the respective RNPs and transferred into a nucleofection strip (Cat# V4XC-2032). Strips were placed in the 4D-Nucleofector System, and nucleofection was done using the CA-137 program. Following nucleofection, pre-warmed medium was added after 10 min, and cells were plated. Compound treatment was started 48h after nucleofection. Knockout efficiency of sgRNAs was assessed using the Synthego ICE Analysis online tool. Genomic DNA was extracted from nucleofected cells using the Quick-DNA Microprep Kit (Zymo Research, Cat# D3021) according to manufacturer’s instructions and amplified by PCR using the Platinum™ II Hot-Start PCR Master Mix (Thermo Fisher Scientific, Cat# 14000012). After a hot start, a denaturation temperature of 95° C, an annealing temperature of 58° C and an extension temperature of 72°C were chosen and repeated for 40 cycles. For amplification the following primers were used, while for sequencing only the forward primer was used: NLRP5 forward: CTTGAGAATTTGCTGCAAGATCCT, NLRP5 reverse: CGATTCTTCCCTGTTCCCATGAG, RAB3C forward: CCACTCGCCTCCTGAGTGTCTG, RAB3C reverse: GAACAAGGCAGAAAGTTTCTCCC, RAB12 forward: CTGTGCGCATGGGAGTGTTTTC, RAB12 reverse: CTTACCCACGGTGGACTTGC. Statistical analyses Statistical analysis of continuous variables was performed with R version 4.2.1 (2022–06-23) and RStudio (version 2022.07.1; RStudio, Inc.) using either mean and standard deviation (SD) or median and interquartile range (IQR), depending on the distribution of data tested by visualization with histograms, quantile-quantile plots and normality testing using the Shapiro-Wilk test. Sample sizes are indicated (n) for each analysis. The T-Test (for normally distributed variables) and the Mann-Whitney U test (for non-parametric distributions) was performed to test for statistical differences. ACKNOWLEDGEMENTS The authors thank the patients and their families for participating in this study. The authors thank Selva G. Nataraja, PhD, and the teams at Mitobridge Inc. and Astellas Inc. for feedback and for providing small molecule libraries. The authors thank Jen Smith, PhD, Clarence Yapp, PhD, and the team from the ICCB-Longwood Screening Facility for help with designing and conducting screening experiments, Igor Paron, PhD and Tim Heymann, PhD from the Max Planck Institute of Biochemistry for their technical support for mass spectrometry, and the BCH-Astellas Joint Steering Committee members, Thomas Schwarz, PhD, Larry Benowitz, PhD, and Zhigang He, PhD, for critical feedback and guidance. This study was supported by research grants from the CureAP4 Foundation (to D.E.-F.), the Spastic Paraplegia Foundation (to D.E.-F.), the Tom-Wahlig Foundation (to A.S. and D.E.-F.), the Manton Center for Orphan Disease Research (to D.E.-F.), the BCH Office of Faculty Development (to D.E.-F.), the BCH Translational Research Program (to D.E.-F.), the National Institute of Neurological Disorders and Stroke grant (1K08NS123552-01 to D.E.-F.), and a joint research agreement with Mitobridge Inc. and Astellas Pharmaceuticals Inc. (to D.E.-F. & M.S.). A.S. was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) - SA 4171/1-1. Further support is acknowledged from the German National Academic Foundation (to B.B., C.B., J.E.A, M.Z., M.Sch.), the Carl-Duisburg Program of the Bayer Foundation (to B.B., M.Z.), the German National Exchange Service (to C.B., J.E.A, M.Z.), the European Union’s Horizon 2020 research and innovation program under the Marie Sklodowska-Curie grant agreement no. 896725 (to A.K.D), and the Rosamund Stone Zander chair (to M.S.). The IDDRC at Boston Children’s Hospital is supported by National Institutes of Health grant 1U54HD090255. Data availability RNA sequencing data will be made publicly available through the National Center for Biotechnology Information’s Sequence Read Archive (SRA) [accession number: pending]. Mass spectrometry proteomics data will be deposited to the ProteomeXchange Consortium [http://proteomecentral.proteomexchange.org; link to be updated] via the PRIDE partner repository. Data tables with source data are provided in the supplementary material. Source images are available from the author upon reasonable request. All fibroblast and iPSC lines generated in this study are available with a material transfer agreement. Figure 1 Establishment of a cell-based phenotypic small molecule screening platform using ATG9A translocation as a surrogate for AP-4 function and primary screening of 28,864 novel small molecule compounds. (a) Overview of the primary screening of 28,864 novel small molecule compounds in fibroblasts from a patient with AP-4-HSP due to biallelic loss-of-function variants in AP4B1. (b) Illustration of the automated image analysis pipeline. Representative images of fibroblasts from a patient with HSP-SPG47 (negative control, LoF/LoF) and their sex-matched heterozygous parent (positive control, WT/LoF) are shown. Four markers are captured including Phalloidin (grey), DAPI (blue), TGN (red) and ATG9A (green). The TGN and ATG9A channels are additionally depicted in greyscale. Through a series of masks, the intracellular distribution of ATG9A is calculated at the level of individuals cells, with hundreds of thousands to millions of cells per experiment. Scale bar: 20μm. (c) Overview of the high-throughput platform and workflow. The assay is miniaturized to 96- or 384-well microplates. Cells are stained using automated liquid handlers and imaged using an automated high-content confocal microscope, followed by automated image analysis. Primary metric is the ‘ATG9A ratio’, which is calculated by dividing the ATG9A fluorescence intensity (F.U.) inside the TGN by the ATG9A fluorescence intensity in the cytoplasm. (d-f) The distribution of ATG9A fluorescence intensities inside (d) and outside (e) the TGN, as well as the ATG9A ratio (f) are shown on a per cell basis. 99,927 WT/LoF and 119,522 LoF/LoF cells were quantified. (g) Cell counts are measured for each experimental well. 1312 wells were analyzed per condition. (h&i) Replicate plots were generated by random sampling of the 82 plates from the primary screen in two groups. Similar positions on the assay plates were plotted against each other with respect to the ATG9A fluorescence intensity inside the TGN (h) and the ATG9A ratio (i). Replicate correlations for both analysis strategies were assessed by averaging the Pearson correlation coefficients of 100 random sampling tests. The ATG9A ratio shows a mean Pearson correlation coefficient (r) of 0.9, while the ATG9A fluorescence inside the TGN shows an average r of 0.82. (j) To demonstrate the discriminative power of the ATG9A ratio in separating positive and negative controls, statistical testing was done using the Mann-Whitney U test. Quantification was done using per well means. 1312 wells per condition were included. Positive and negative controls showed a robust separation (p < 0.0001). (k) To test the robustness of separation of the ATG9A ratio between positive (WT/LoF) and negative controls (LoF/LoF), a dataset containing measurement for 99,927 WT/LoF and 119,522 LoF/LoF cells was partitioned into a training set (70% of data) and a test set (30%). A generalized linear model was trained using the training set. The performance of the prediction model using the test set is shown in (k). The AUC is 0.96. (l) Impact of 28,864 compounds applied for 24h at a concentration 10μM. Z-scores for the primary metric, the ATG9A ratio, are shown. All data points represent per well means. The mean of the positive control (WT/LoF) is shown as a green dotted line. The green shaded areas represent ± 1 SD. Active compounds were a priori defined as those reducing the ATG9A ratio by at least 3 SD compared to negative controls. Toxicity was defined as a reduction of cell count of at least 2SD compared to the negative control. 501 compounds show activity by reducing the ATG9A ratio by more than 3 SD. (m) Distribution of Z-scores of all non-toxic 27,412 compounds. Active compounds are highlighted in blue. Figure 2 Counter-screen in fibroblasts from AP-4-HSP patients confirms 16 compounds that lead to dose-dependent redistribution of ATG9A. (a) Overview of the counter-screen of the 503 active compounds identified in the primary screen. To assess for dose-dependent effects, compounds were screened in AP-4-HSP patient-derived fibroblasts in 384-well microplates using 11-point titrations ranging from 40nM to 40μM. All concentrations were screened in duplicates. Active compounds were a priori defined as those reducing the ATG9A ratio by at least 3SD compared to negative controls, in more than one concentration. Toxicity was defined as a reduction of the cell count of at least 2 SD compared to negative controls. 51 compounds demonstrated a clear and reproducible dose-response relationship and raised no suspicion for autofluorescence on automated and manual review. 34 compounds showed autofluorescence or resulted in imaging artifacts. One active compound was unavailable from the manufacturer and was therefore excluded from subsequent testing. (b) Baseline differences in the ATG9A distribution in WT/LoF (n=269) vs. LoF/LoF (n=269) fibroblasts. Statistical testing was done using the Mann-Whitney U test. Positive and negative controls showed a robust separation (p < 0.0001). (c) Dose-response curves were fitted using a four-parameter logistic regression model, and EC50 concentrations were calculated. All concentrations were tested in biologic duplicates. Most EC50 were in the low micromolar range (median: 4.66mM, IQR: 8.63). Black dashed lines represent the a priori defined thresholds of +/− 3SD compared to the negative control (LoF/LoF). Red triangles represent toxic concentrations based on the a priori defined threshold of a reduction of cell counts of at least 2 SD compared to the negative control. The salmon-colored dashed line represents the mean of negative controls, while the green-colored dashed line depicts the mean of the positive controls (WT/LoF). Representative images of the EC50 are shown for each active compound. Representative images show a merge of the 4 channels: Phalloidin (grey), DAPI (blue), TGN (red) and ATG9A (green), as well as the TGN and ATG9A channels in greyscale. For better illustration of differences in ATG9A signals, the fluorescence intensities of the ATG9A channel are additionally shown using a color lookup table. Scale bar: 20μm. NA: not available Figure 3 Orthogonal assays in AP4B1KO SH-SY5Y cells confirm 5 active compounds. (a) Overview of the orthogonal screen of 16 active compounds in differentiated AP4B1KO SH-SY5Y cells, a neuronal model of AP-4 deficiency. Active compounds were a priori defined as those reducing the ATG9A ratio by at least 3 SD compared to negative controls. Toxicity was defined as a reduction of cell count of at least 2 SD compared to the negative control. (b) Baseline differences in ATG9A ratios of AP4B1WT vs. AP4B1KO SH-SY5Y cells were quantified from 160 AB4B1WT and 158 AB4B1KO wells from 5 assay plates. Statistical testing was performed using the Mann-Whitney U test. Positive and negative controls showed a robust separation (p < 0.0001). (c-g) Dose-response curves for ATG9A ratios in AB4B1KO cells treated with different compounds. Data points represent per-well means from 3 different assay plates. Dashed lines show mean Z-scores for positive (green) and negative (salmon) controls. Shaded areas represent ± 1 SD. (h) Representative images of the intracellular ATG9A distribution for individual compounds. The merged image shows beta-3 tubulin (grey), DAPI (blue), the TGN (red) and ATG9A (green). The TGN and ATG9A channels are further separately depicted in greyscale. Scale bar: 10μm. (i) Baseline differences of DAGLB ratios in AP4B1WT vs. AP4B1KO cells were quantified from 192 AB4B1WT and 192 AB4B1KO wells from 4 assay plates. Statistical testing was done using the Mann-Whitney U test. Positive and negative controls showed a robust separation (p < 0.0001). (j-n) Dose-response curves for DAGLB ratios in AB4B1KO cells treated with different compounds. All data points represent per-well means from 4 different assay plates. Dashed lines show mean Z-scores for positive (green) and negative (salmon) controls. Shaded areas represent ± 1 SD. (o) Representative images of the intracellular DAGLB distribution for individual compounds. The merge shows beta-3 tubulin (grey), DAPI (blue), the TGN (red) and DAGLB (green). The TGN and DAGLB channels are further separately depicted in greyscale. Scale bar: 10μm. Figure 4 Multiparametric profiling of 5 active compounds in AP4B1KO SH-SY5Y cells. (a) Multiparametric profiling of images of 5373 cells were acquired using 4 fluorescent channels. Scale: 10μm. A total of 90 measurements per cell were generated for the cytoskeleton (beta-3 tubulin), the nucleus (DAPI), the TGN (TNG46) and ATG9A vesicles (ATG9A). The different steps of data preprocessing and phenotypic clustering using principal component analysis (PCA) are shown. (b) PCA shows different clusters of cells based on 85 phenotypic features. Experimental conditions are color-coded. The first two principal components (PC1 and PC2) explain 43.2% of the observed variance. (c) Bar plot summarizing the variance explained by the first 10 principal components (PCs). Most of the variance is explained by PC1 and to a lesser degree PC2. (d) Correlation analysis of PC1 with all 85 features using the Pearson correlation coefficient. Red dashed lines represent cut-offs for correlations >0.75. (e) Zoom-in on selected features of interest showing a correlation with PC1 >0.75. (f) Measurements of TGN intensity and descriptors of TGN shape and network complexity for the individual hit compounds as line graphs and (g) summarized using heatmap visualization. Figure 5 Compound C-01 restores ATG9A and DAGLB trafficking in iPSC-derived neurons from AP-4-HSP patients. (a) Overview of the testing of 5 active compounds in iPSC-derived cortical neurons from a patient with AP4M1-associated SPG50 compared to heterozygous controls (same-sex parent). Active compounds were defined as those reducing the ATG9A ratio by at least 3 SD compared to negative controls (patient-derived iPSC-neurons treated with vehicle). Toxicity was defined as a reduction of cell count of at least 2 SD compared to the negative control. (b) Baseline differences of ATG9A ratios in controls vs. patient-derived iPSC-neurons were quantified using per well means of 60 wells per condition from 5 plates. Statistical testing was done using the Mann-Whitney U test. Positive and negative controls showed a robust separation (p < 0.0001). (c) Representative images of iPSC-neurons from a patient with SPG50 treated with individual compounds at 5μM for 24h (~EC50 in prior experiments). The merge shows beta-3 tubulin (grey), DAPI (blue), the Golgi (red) and ATG9A (green). The Golgi and ATG9A channels are further separately depicted in greyscale. For better illustration of differences in ATG9A signals, the fluorescence intensities of the ATG9A channel are additionally shown using a color lookup table. Scale: 10μm. (d-f) Dose-response curves for ATG9A ratios in iPSC-neurons from a patient with SPG50 treated with individual compounds for 24h, along with their morphological profiles depicted as heatmaps. All data points represent per-well means of 3–4 independent differentiations. Dashed lines show mean Z-scores for positive (green) and negative (salmon) controls. Shaded areas represent ± 1SD. (g) Chemical synthesis and structure of lead compound C-01. (h&i) Dose-response curves for ATG9A and DAGLB ratios in iPSC-neurons from a patient with SPG50 (h) and an additional patient with SPG47 (i) after prolonged treatment with C-01for 72h, along with the morphologic profile depicting changes in cellular ATG9A and DAGLB distribution, TGN intensity and morphology and cell count. All data points represent per-well means of 2 independent differentiations. Dashed lines show mean Z-scores for positive (green) and negative (salmon) controls. Shaded areas represent ± 1 SD. Figure 6 Target deconvolution using bulk RNA sequencing and weighted gene co-expression network analysis in AP4B1KO SH-SY5Y cells treated with C-01. (a) Hierarchical clustering of 12 samples using average linkage showed two main clusters based on treatment with vehicle vs. C-01, irrespective of cell line. (b) Cluster dendrogram of 18,506 expressed genes based on topological overlap. Clusters of co-expressed genes (“modules”) were isolated using hierarchical clustering and adaptive branch pruning. (c) Heatmap visualization of the correlation of gene expression profiles (“module eigengene”, ME) of each module with measured traits. Pearson correlation coefficients are shown for each cell of the heatmap. (d) Intramodular analysis of module membership (MM) and gene significance (GS) for highly correlated modules, allowing identification of genes that have high significance with treatment as well as high connectivity to their modules. (e) ME expression profiles for the top 5 co-expressed modules. (f) Gene ontology enrichment analysis showed enriched pathways in 3/5 modules. Pathways were considered differentially expressed with an FDR < 0.05. Figure 7 Target deconvolution using unbiased quantitative proteomics in AP4B1KO SH-SY5Y cells and AP-4-HSP patient-derived iPSC-neurons treated with C-01. (a – c) Differential protein enrichment analysis. Statistical testing was done using protein-wise linear models and empirical Bayes statistics. Proteins were considered as differentially enriched with a false discovery rate of < 0.05 and a log2 fold change > 0.3. (a) SH-SY5Y cells: 8141 unique proteins were analyzed. PCA of the top 500 variable proteins shows robust separation between experimental conditions. The volcano plot summarizes differential protein enrichment for AP4B1WT and AP4B1KO cells pooled into two groups, vehicle vs. C-01 treated. Differentially enriched proteins are depicted in black. Proteins with the most consistent enrichment profiles across all experimental conditions (see Supplementary Fig. 6a-d) are colored and labeled in red. (b) iPSC-derived neurons: 7386 unique proteins were analyzed. PCA of the top 500 variable proteins shows robust separation between experimental conditions. The volcano plot summarizes differential protein enrichment for control and patient-derived neurons pooled into two groups, vehicle vs. C-01 treated. Differentially enriched proteins are depicted in black. Proteins with the most consistent enrichment profiles across all experimental conditions (see Supplementary Fig. 6e-h) are colored and labeled in red. (c) Integrated analysis of SH-SY5Y cells and iPSC-derived neurons: 5357 unique proteins were analyzed. The volcano plot summarizes differential protein enrichment for control and AP-4-deficient cells pooled into two groups, vehicle vs. C-01. Proteins with the most consistent enrichment profiles across all experimental conditions (see Supplementary Figure 6i-l) are colored and labeled in red. The dot plot summarizes dysregulated Reactome pathways of the pooled analysis. Pathways were considered differentially expressed with an FDR < 0.05. (d) The RAB protein family members RAB1B, RAB3C and RAB12 showed the most consistent profiles in response to C-01 treatment and were selected for further analysis. LFQ intensities in SH-SY5Y cells (AP4B1WT and AP4B1KO pooled) and neurons (control and patient pooled) are shown. Statistical testing was done using pairwise T-tests. P-values were adjusted for multiple testing using the Benjamini-Hochberg procedure. (e) LFQ intensities of RAB3C and RAB12 in AP4B1WT (n = 11 samples) and AP4B1KO (n = 10 samples) SH-SY5Y cells, as well as control (n = 6 samples) and patient-derived (n = 6 samples) iPSC-derived neurons show a high degree of correlation measured by the Pearson correlation coefficient (r). While there was no difference between genotypes (not shown), C-01 treated cells showed reduced protein levels of both RAB3C and RAB12. Figure 8 RAB3C and RAB12 are involved in C-01-mediated vesicle trafficking and enhancement of autophagic flux. (a) AP4B1KO SH-SY5Y cells were transfected for 72h with RNPs targeting RAB3C, RAB12 or both compared to NLRP5 as a non-essential control. Vehicle vs. C-01 treatment at a concentration of 5μM was administered for 24h. Each experimental condition was tested in 18–28 wells from 3–5 independent plates. The dashed line represents a reduction of the ATG9A ratio of −2 SD compared to the negative control (AP4B1KO + sgNLRP5). Knockout of RAB12 did not significantly alter the ATG9A ratio, while RAB3C knockout led to a reduction of −2 SD. Combining the knockout of RAB3C and RAB12 did not result in an additive effect. However, both RAB3C and RAB12 knockout potentiated the effect of C-01 treatment on ATG9A translocation, which was further enhanced by combined knockout. (b) Representative images of the intracellular ATG9A distribution for different conditions. The merged image shows beta-3 tubulin (grey), DAPI (blue), the TGN (red) and ATG9A (green). The TGN and ATG9A channels are further separately depicted in greyscale. Scale bar: 10μm. (c-f) Representative western blot of whole cell lysates. Cells were treated with vehicle vs. C-01 at a concentration of 5μM for 72h. All experiments were performed in four biological replicates. As expected, AP4E1 levels were reduced in AP4B1KO cells, indicating reduced AP-4 complex formation. ATG9A ratios were significantly increased in AP4B1KO cells and were not altered by C-01 treatment. By contrast, the conversion of LC3-I to LC3-II was significantly elevated in response to C-01 in both AP4B1WT and AP4B1KO cells. To confirm that this increase was due to an increase in autophagic flux, autophagosome-lysosome fusion was blocked by adding bafilomycin A1 at a concentration of 100nM for 4 h prior to cell harvest. (g-h) AP4B1KO SH-SY5Y cells were treated for 72h with vehicle, C-01 (5μM) alone, or C-01 in combination with ascending non-toxic doses of either bafilomycin A1 (5nM or 10nM) or chloroquine (1μM or 2μM). Each experimental condition was tested in 16 wells from 2 independent plates. The dashed line represents a reduction of the ATG9A ratio of −2 SD compared to the negative control (AP4B1KO). As expected, C-01 treatment alone led to a considerable reduction of the ATG9A ratio. Block of late stages of autophagy using either bafilomycin A1 or chloroquine reversed the effect of C-01 in a dose dependent manner. (i) Representative images of the intracellular ATG9A distribution for different conditions. The merged image shows beta-3 tubulin (grey), DAPI (blue), the TGN (red) and ATG9A (green). The TGN and ATG9A channels are further separately depicted in greyscale. Scale bar: 10μm. (j-l) Western blots of whole cell lysates of AP4B1KO SH-SY5Y cells transfected for 72h with RNPs against RAB3C, RAB12 or both, compared to NLRP5. Vehicle vs. C-01 treatment was administered for 48h. While neither RAB3C (g) nor RAB12 (h) knockout alone led to an increase in baseline LC3-II, the combined knockout raised the LC3-II to LC3-I ratio to levels achieved with C-01 treatment alone (i). In response to bafilomycin A1 treatment (100nM for 4 h) both RAB3C knockout alone and the combined knockout of RAB3C and RAB12 led to a significant increase in LC3-II to LC3-I ratios. Statistical testing in all experiments was done using pairwise T-tests. P-values were adjusted for multiple testing using the Benjamini-Hochberg procedure. COMPETING INTERESTS This work was supported by a joint research agreement between Boston Children’s Hospital and Mitobridge Inc., now owned by Astellas Pharmaceuticals Inc.. D.E.F. has served as a consultant to Health Advances LLC, has received speaker honoraria from the Movement Disorders Society, and publishing royalties from Cambridge University Press. M.S. reports grant support from Novartis, Biogen, Astellas, Aeovian, Bridgebio, and Aucta unrelated to this project. He has served on Scientific Advisory Boards for Novartis, Roche, Regenxbio, SpringWorks Therapeutics, Jaguar Therapeutics and Alkermes. ==== Refs References 1. Tambuyzer E , Therapies for rare diseases: therapeutic modalities, progress and challenges ahead. Nat Rev Drug Discov 19 , 93–111 (2020).31836861 2. Moffat JG , Vincent F , Lee JA , Eder J , Prunotto M . Opportunities and challenges in phenotypic drug discovery: an industry perspective. Nat Rev Drug Discov 16 , 531–543 (2017).28685762 3. Eder J , Sedrani R , Wiesmann C . The discovery of first-in-class drugs: origins and evolution. Nat Rev Drug Discov 13 , 577–587 (2014).25033734 4. Sun W , Zheng W , Simeonov A . Drug discovery and development for rare genetic disorders. Am J Med Genet A 173 , 2307–2322 (2017).28731526 5. Macarron R , Impact of high-throughput screening in biomedical research. Nat Rev Drug Discov 10 , 188–195 (2011).21358738 6. Vincent F , Developing predictive assays: the phenotypic screening “rule of 3”. Sci Transl Med 7 , 293ps215 (2015). 7. Ebrahimi-Fakhari D , Behne R , Davies AK , Hirst J . AP-4-Associated Hereditary Spastic Paraplegia. In: GeneReviews((R)) (eds Adam MP , et al.) (2018). 8. Ebrahimi-Fakhari D , Defining the clinical, molecular and imaging spectrum of adaptor protein complex 4-associated hereditary spastic paraplegia. Brain 143 , 2929–2944 (2020).32979048 9. Ebrahimi-Fakhari D , Systematic Analysis of Brain MRI Findings in Adaptor Protein Complex 4-Associated Hereditary Spastic Paraplegia. Neurology 97 , e1942–e1954 (2021).34544818 10. Ebrahimi-Fakhari D , Clinical and genetic characterization of AP4B1-associated SPG47. Am J Med Genet A 176 , 311–318 (2018).29193663 11. Mattera R , Park SY , De Pace R , Guardia CM , Bonifacino JS . AP-4 mediates export of ATG9A from the trans-Golgi network to promote autophagosome formation. Proc Natl Acad Sci U S A 114 , E10697–E10706 (2017).29180427 12. Davies AK , AP-4 vesicles contribute to spatial control of autophagy via RUSC-dependent peripheral delivery of ATG9A. Nat Commun 9 , 3958 (2018).30262884 13. De Pace R , Altered distribution of ATG9A and accumulation of axonal aggregates in neurons from a mouse model of AP-4 deficiency syndrome. PLoS Genet 14 , e1007363 (2018).29698489 14. Ivankovic D , Axonal autophagosome maturation defect through failure of ATG9A sorting underpins pathology in AP-4 deficiency syndrome. Autophagy 16 , 391–407 (2020).31142229 15. Behne R , Adaptor protein complex 4 deficiency: a paradigm of childhood-onset hereditary spastic paraplegia caused by defective protein trafficking. Hum Mol Genet 29 , 320–334 (2020).31915823 16. Dell’Angelica EC , Mullins C , Bonifacino JS . AP-4, a novel protein complex related to clathrin adaptors. J Biol Chem 274 , 7278–7285 (1999).10066790 17. Hirst J , Bright NA , Rous B , Robinson MS . Characterization of a fourth adaptor-related protein complex. Mol Biol Cell 10 , 2787–2802 (1999).10436028 18. Gadbery JE , Integrating structural and evolutionary data to interpret variation and pathogenicity in adapter protein complex 4. Protein Sci 29 , 1535–1549 (2020).32285480 19. Sanger A , Hirst J , Davies AK , Robinson MS . Adaptor protein complexes and disease at a glance. J Cell Sci 132 , jcs222992 (2019).31636158 20. Dell’Angelica EC , Bonifacino JS . Coatopathies: Genetic Disorders of Protein Coats. Annu Rev Cell Dev Biol 35 , 131–168 (2019).31399000 21. Ebrahimi-Fakhari D , High-throughput imaging of ATG9A distribution as a diagnostic functional assay for adaptor protein complex 4-associated hereditary spastic paraplegia. Brain Commun 3 , fcab221 (2021).34729478 22. Ebrahimi-Fakhari D , Congenital disorders of autophagy: an emerging novel class of inborn errors of neuro-metabolism. Brain 139 , 317–337 (2016).26715604 23. Teinert J , Behne R , Wimmer M , Ebrahimi-Fakhari D . Novel insights into the clinical and molecular spectrum of congenital disorders of autophagy. J Inherit Metab Dis 43 , 51–62 (2020).30854657 24. D’Amore A , Loss of ap4s1 in zebrafish leads to neurodevelopmental defects resembling spastic paraplegia 52. Ann Clin Transl Neurol 7 , 584–589 (2020).32216065 25. Ziegler M , Blended Phenotype of Silver-Russell Syndrome and SPG50 Caused by Maternal Isodisomy of Chromosome 7. Neurol Genet 7 , e544 (2021).33553621 26. Matsuda S , Accumulation of AMPA receptors in autophagosomes in neuronal axons lacking adaptor protein AP-4. Neuron 57 , 730–745 (2008).18341993 27. Scarrott JM , Ap4b1-knockout mouse model of hereditary spastic paraplegia type 47 displays motor dysfunction, aberrant brain morphology and ATG9A mislocalization. Brain Commun 5 , fcac335 (2023).36632189 28. Yamaguchi J , Atg9a deficiency causes axon-specific lesions including neuronal circuit dysgenesis. Autophagy 14 , 764–777 (2018).28513333 29. Davies AK , AP-4-mediated axonal transport controls endocannabinoid production in neurons. Nat Commun 13 , 1058 (2022).35217685 30. Bray MA , Carpenter A . Advanced Assay Development Guidelines for Image-Based High Content Screening and Analysis, 2004 edn. Eli Lilly & Company and the National Center for Advancing Translational Sciences (2017). 31. Zhang XD . Illustration of SSMD, z score, SSMD*, z* score, and t statistic for hit selection in RNAi high-throughput screens. J Biomol Screen 16 , 775–785 (2011).21515799 32. Malo N , Hanley JA , Cerquozzi S , Pelletier J , Nadon R . Statistical practice in high-throughput screening data analysis. Nat Biotechnol 24 , 167–175 (2006).16465162 33. Kovalevich J , Langford D . Considerations for the use of SH-SY5Y neuroblastoma cells in neurobiology. Methods Mol Biol 1078 , 9–21 (2013).23975817 34. Rohban MH , Systematic morphological profiling of human gene and allele function via Cell Painting. Elife 6 , (2017). 35. Eberhardt K , Generation and characterization of six human induced pluripotent stem cell lines (iPSC) from three families with AP4M1-associated hereditary spastic paraplegia (SPG50). Stem Cell Res 53 , 102335 (2021).34087981 36. Teinert J , Generation and characterization of six human induced pluripotent stem cell lines (iPSC) from three families with AP4B1-associated hereditary spastic paraplegia (SPG47). Stem Cell Res 40 , 101575 (2019).31525725 37. Winden KD , Biallelic Mutations in TSC2 Lead to Abnormalities Associated with Cortical Tubers in Human iPSC-Derived Neurons. J Neurosci 39 , 9294–9305 (2019).31591157 38. Zhang Y , Rapid single-step induction of functional neurons from human pluripotent stem cells. Neuron 78 , 785–798 (2013).23764284 39. Zhang B , Horvath S . A general framework for weighted gene co-expression network analysis. Stat Appl Genet Mol Biol 4 , Article17 (2005). 40. Winden KD , A systems level, functional genomics analysis of chronic epilepsy. PLoS One 6 , e20763 (2011).21695113 41. Langfelder P , Horvath S . Eigengene networks for studying the relationships between co-expression modules. BMC Syst Biol 1 , 54 (2007).18031580 42. Gillespie M , The reactome pathway knowledgebase 2022. Nucleic Acids Res 50 , D687–D692 (2022).34788843 43. Andres DA , cDNA cloning of component A of Rab geranylgeranyl transferase and demonstration of its role as a Rab escort protein. Cell 73 , 1091–1099 (1993).8513495 44. Cremers FP , Armstrong SA , Seabra MC , Brown MS , Goldstein JL . REP-2, a Rab escort protein encoded by the choroideremia-like gene. J Biol Chem 269 , 2111–2117 (1994).8294464 45. Matsui T , Itoh T , Fukuda M . Small GTPase Rab12 regulates constitutive degradation of transferrin receptor. Traffic 12 , 1432–1443 (2011).21718402 46. Lu Q , Wang PS , Yang L . Golgi-associated Rab GTPases implicated in autophagy. Cell Biosci 11 , 35 (2021).33557950 47. Liu S , Storrie B . How Rab proteins determine Golgi structure. Int Rev Cell Mol Biol 315 , 1–22 (2015).25708460 48. Matoba K , Atg9 is a lipid scramblase that mediates autophagosomal membrane expansion. Nat Struct Mol Biol 27 , 1185–1193 (2020).33106658 49. Maeda S , Structure, lipid scrambling activity and role in autophagosome formation of ATG9A. Nat Struct Mol Biol 27 , 1194–1201 (2020).33106659 50. Guardia CM , Structure of Human ATG9A, the Only Transmembrane Protein of the Core Autophagy Machinery. Cell Rep 31 , 107837 (2020).32610138 51. van Vliet AR , ATG9A and ATG2A form a heteromeric complex essential for autophagosome formation. Mol Cell 82 , 4324–4339 e4328 (2022).36347259 52. Bamshad MJ , Nickerson DA , Chong JX . Mendelian Gene Discovery: Fast and Furious with No End in Sight. Am J Hum Genet 105 , 448–455 (2019).31491408 53. Nguengang Wakap S , Estimating cumulative point prevalence of rare diseases: analysis of the Orphanet database. Eur J Hum Genet 28 , 165–173 (2020).31527858 54. Gunne E , McGarvey C , Hamilton K , Treacy E , Lambert DM , Lynch SA . A retrospective review of the contribution of rare diseases to paediatric mortality in Ireland. Orphanet J Rare Dis 15 , 311 (2020).33148291 55. Walker CE , The collective impact of rare diseases in Western Australia: an estimate using a population-based cohort. Genet Med 19 , 546–552 (2017).27657686 56. Sandilands K , Williams A , Rylands AJ . Carer burden in rare inherited diseases: a literature review and conceptual model. Orphanet J Rare Dis 17 , 428 (2022).36494728 57. Yang G , Cintina I , Pariser A , Oehrlein E , Sullivan J , Kennedy A . The national economic burden of rare disease in the United States in 2019. Orphanet J Rare Dis 17 , 163 (2022).35414039 58. Mattera R , Williamson CD , Ren X , Bonifacino JS . The FTS-Hook-FHIP (FHF) complex interacts with AP-4 to mediate perinuclear distribution of AP-4 and its cargo ATG9A. Mol Biol Cell 31 , 963–979 (2020).32073997 59. Orsi A , Dynamic and transient interactions of Atg9 with autophagosomes, but not membrane integration, are required for autophagy. Mol Biol Cell 23 , 1860–1873 (2012).22456507 60. Young AR , Starvation and ULK1-dependent cycling of mammalian Atg9 between the TGN and endosomes. J Cell Sci 119 , 3888–3900 (2006).16940348 61. Orii M , Tsuji T , Ogasawara Y , Fujimoto T . Transmembrane phospholipid translocation mediated by Atg9 is involved in autophagosome formation. J Cell Biol 220 , e202009194 (2021).33439214 62. Chumpen Ramirez S , --Atg9 interactions via its transmembrane domains are required for phagophore expansion during autophagy. Autophagy, 1–20 (2022). 63. Hara T , Suppression of basal autophagy in neural cells causes neurodegenerative disease in mice. Nature 441 , 885–889 (2006).16625204 64. Komatsu M , Loss of autophagy in the central nervous system causes neurodegeneration in mice. Nature 441 , 880–884 (2006).16625205 65. Komatsu M , Essential role for autophagy protein Atg7 in the maintenance of axonal homeostasis and the prevention of axonal degeneration. Proc Natl Acad Sci U S A 104 , 14489–14494 (2007).17726112 66. Bunge MB . Fine structure of nerve fibers and growth cones of isolated sympathetic neurons in culture. J Cell Biol 56 , 713–735 (1973).4347207 67. Hollenbeck PJ . Products of endocytosis and autophagy are retrieved from axons by regulated retrograde organelle transport. J Cell Biol 121 , 305–315 (1993).7682217 68. Maday S , Holzbaur EL . Autophagosome biogenesis in primary neurons follows an ordered and spatially regulated pathway. Dev Cell 30 , 71–85 (2014).25026034 69. Maday S , Wallace KE , Holzbaur EL . Autophagosomes initiate distally and mature during transport toward the cell soma in primary neurons. J Cell Biol 196 , 407–417 (2012).22331844 70. Ebrahimi-Fakhari D , Impaired Mitochondrial Dynamics and Mitophagy in Neuronal Models of Tuberous Sclerosis Complex. Cell Rep 17 , 1053–1070 (2016).27760312 71. Wager TT , Hou X , Verhoest PR , Villalobos A . Central Nervous System Multiparameter Optimization Desirability: Application in Drug Discovery. ACS Chem Neurosci 7 , 767–775 (2016).26991242 72. Stenmark H. Rab GTPases as coordinators of vesicle traffic. Nat Rev Mol Cell Biol 10 , 513–525 (2009).19603039 73. Wandinger-Ness A , Zerial M . Rab proteins and the compartmentalization of the endosomal system. Cold Spring Harb Perspect Biol 6 , a022616 (2014).25341920 74. Schluter OM , Schmitz F , Jahn R , Rosenmund C , Sudhof TC . A complete genetic analysis of neuronal Rab3 function. J Neurosci 24 , 6629–6637 (2004).15269275 75. Matsui T , Fukuda M . Rab12 regulates mTORC1 activity and autophagy through controlling the degradation of amino-acid transporter PAT4. EMBO Rep 14 , 450–457 (2013).23478338 76. Majumder P , Edmison D , Rodger C , Patel S , Reid E , Gowrishankar S . AP-4 regulates neuronal lysosome composition, function, and transport via regulating export of critical lysosome receptor proteins at the trans-Golgi network. Mol Biol Cell 33 , ar102 (2022).35976706 77. Ebrahimi-Fakhari D , Reduction of TMEM97 increases NPC1 protein levels and restores cholesterol trafficking in Niemann-pick type C1 disease cells. Hum Mol Genet 25 , 3588–3599 (2016).27378690 78. McCarthy DJ , Smyth GK . Testing significance relative to a fold-change threshold is a TREAT. Bioinformatics 25 , 765–771 (2009).19176553 79. Chen Y , Lun AT , Smyth GK . From reads to genes to pathways: differential expression analysis of RNA-Seq experiments using Rsubread and the edgeR quasi-likelihood pipeline. F1000Res 5 , 1438 (2016).27508061 80. Wu T , clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation (Camb) 2 , 100141 (2021).34557778 81. Anders S , Huber W . Differential expression analysis for sequence count data. Genome Biol 11 , R106 (2010).20979621 82. Ritchie ME , limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43 , e47 (2015).25605792 83. Langfelder P , Horvath S . WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics 9 , 559 (2008).19114008 84. Langfelder P , Horvath S . Fast R Functions for Robust Correlations and Hierarchical Clustering. J Stat Softw 46 , i11 (2012).23050260 85. Kulak NA , Pichler G , Paron I , Nagaraj N , Mann M . Minimal, encapsulated proteomic-sample processing applied to copy-number estimation in eukaryotic cells. Nat Methods 11 , 319–324 (2014).24487582 86. Cox J , Mann M . MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat Biotechnol 26 , 1367–1372 (2008).19029910 87. Sinitcyn P , MaxDIA enables library-based and library-free data-independent acquisition proteomics. Nat Biotechnol 39 , 1563–1573 (2021).34239088 88. Cox J , Hein MY , Luber CA , Paron I , Nagaraj N , Mann M . Accurate proteome-wide label-free quantification by delayed normalization and maximal peptide ratio extraction, termed MaxLFQ. Mol Cell Proteomics 13 , 2513–2526 (2014).24942700 89. Johnson WE , Li C , Rabinovic A . Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics 8 , 118–127 (2007).16632515

PMC010xxxxxx/PMC10313903.txt

==== Front Gates Open Res Gates Open Res Gates Open Research 2572-4754 F1000 Research Limited London, UK 37398911 10.12688/gatesopenres.13338.3 Method Article Articles Developing and deploying an efficient genotyping workflow for accelerating maize improvement in developing countries [version 3; peer review: 3 approved, 1 approved with reservations] Offornedo Queen Conceptualization Investigation Methodology Writing – Original Draft Preparation https://orcid.org/0000-0002-1317-219X 1 Menkir Abebe Resources 1 Babalola Deborah Writing – Review & Editing 1 Gedil Melaku Conceptualization Funding Acquisition Supervision Writing – Review & Editing https://orcid.org/0000-0002-6258-6014 a1 1 Bioscience Center and Maize Improvement Program, International Institute of Tropical Agriculture (IITA) Headquarters, Ibadan, Oyo State, 200001, Nigeria a M.Gedil@cgiar.org No competing interests were disclosed. 3 8 2022 2022 6 328 7 2022 Copyright: © 2022 Offornedo Q et al. 2022 https://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background: Molecular breeding is an essential tool for accelerating genetic gain in crop improvement towards meeting the need to feed an ever-growing world population. Establishing low-cost, flexible genotyping platforms in small, public and regional laboratories can stimulate the application of molecular breeding in developing countries. These laboratories can serve plant breeding projects requiring low- to medium-density markers for marker-assisted selection (MAS) and quality control (QC) activities. Methods: We performed two QC and MAS experiments consisting of 637 maize lines, using an optimised genotyping workflow involving an in-house competitive allele-specific PCR (KASP) genotyping system with an optimised sample collection, preparation, and DNA extraction and quantitation process. A smaller volume of leaf-disc size plant samples was collected directly in 96-well plates for DNA extraction, using a slightly modified CTAB-based DArT DNA extraction protocol. DNA quality and quantity analyses were performed using a microplate reader, and the KASP genotyping and data analysis was performed in our laboratory. Results: Applying the optimized genotyping workflow expedited the QC and MAS experiments from over five weeks (when outsourcing) to two weeks and eliminated the shipping cost. Using a set of 28 KASP single nucleotide polymorphisms (SNPs) validated for maize, the QC experiment revealed the genetic identity of four maize varieties taken from five seed sources. Another set of 10 KASP SNPs was sufficient in verifying the parentage of 390 F 1 lines. The KASP-based MAS was successfully applied to a maize pro-vitamin A (PVA) breeding program and for introgressing the aflatoxin resistance gene into elite tropical maize lines. Conclusion: This improved workflow has helped accelerate maize improvement activities of IITA's Maize Improvement Program and facilitated DNA fingerprinting for tracking improved crop varieties. National Agricultural Research Systems (NARS) in developing countries can adopt this workflow to fast-track molecular marker-based genotyping for crop improvement. Molecular breeding KASP Genotyping workflow Marker-assisted selection Quality Control National Agricultural Research Systems (NARS) Developing countries CGIAR Research Program on Maize (MAIZE)This work was supported by the Bill & Melinda Gates Foundation and USAID [OPP1134248] under the project titled "Stress Tolerant Maize for Africa". This work was also supported by the Accelerating Genetic Gains in Maize and Wheat (AGG) Project and CGIAR Research Program on Maize (MAIZE). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Revised Amendments from Version 2 Version 3 has been updated based on the reviewers' comments on the previous version. We have included additional context to the Introduction section (P3) to articulate the use cases better and provided a genotyping cost comparison of the procedure as indicated by Reviewer 1. As pointed out by Reviewer 2, we have rephrased the indicated sentences in the Introduction and Methods section. The misplaced reference has been replaced and listed in the reference table. We have included additional context to suitably articulate the study objective towards the end of the Introduction section, which is "This study aims to develop a genotyping workflow optimized for cost-effective and fast turn-around time that can be deployed by less sophisticated and reasonably equipped laboratories in developing countries, to accelerate maize improvement research." We have furnished Table 1 with details of the exact number of genotypes and samples used for the experiments. We have also provided a new figure (Figure 5) and table (Table 5) to aptly articulate the KASP genotyping analysis for the hybrid verification experiment. ==== Body pmcIntroduction Agriculture is the mainstay of millions of low-income households in Sub-Saharan Africa ( SSA). However, productivity is way below the yield potential of significant crops due to several interacting factors contributing to the yield reduction. The paucity of nutritionally improved resilient crop varieties is a crucial constraint. This constraint can be mitigated by the rapid development of cultivars adapted to specific agroecology zones 1 . The current yield gain trend in major food crops has shown that relying on conventional breeding alone is insufficient to meet the food needs of an estimated nine billion people in 2025 2 . There is a need to accelerate genetic gain by deploying new breeding strategies 3, 4 . This need has led to the scientific community's massive investment in developing genomic resources and support systems, to provide valuable tools to accelerate breeding processes 5 . Various bottlenecks have hindered the substantial impact of molecular breeding for crop improvement, particularly in developing countries 6, 7 . The major limiting factors are a lack of infrastructure and capacity for genomics resources and poor information flow, resulting in reduced access to operational and decision support tools 8 . Private companies in developed countries usually own the proprietary rights to many emerging genomics resources and systems, making it difficult for public research sectors, non-profit research institutes, and small laboratories in developing countries to have direct access. These challenges are being curbed by various international initiatives such as the Excellence in Breeding (EiB) platform, which coordinates its activities with the Genomic and Open-source Breeding Informatics Initiative (GOBii), and High Through-Put Genotyping (HTPG). In addition, the Integrated Breeding Platform (IBP)-hosted Generation Challenge Program (GCP) and the Breeding Management System (BMS) 9 target the development and adoption of molecular breeding in developing countries. These and other consultative group-hosted initiatives and platforms galvanise worldwide partners drawn from public, private, and governmental institutions towards the common goal of increasing agricultural productivity through efficient tools, technologies, and data management systems 6 . Despite the availability of many low-cost genotyping platforms and resources, it is not easy to meet the genotyping needs of many users who work on different crops, different locations, and often fewer samples due to cost implications 7, 8 . The current available genotyping platforms have a minimum sample size requirement. For instance, the EiB facilitated genotyping at Intertek offers reduced cost if the user orders genotyping of 1536 samples; fewer samples are acceptable, but the price increases. Intertek's standard cost for routine KASP genotyping is $2.6 per sample per 10 SNPs, excluding shipping costs, compared to our in-house genotyping at $2.95. Even though large volume sizes can be consolidated and shipped for genotyping, there are times when breeders and partners may want to fingerprint a few dozen lines for identity or parentage analysis for quick decision making. In such cases, sending less than the minimum number of samples is not only more priced per datapoint but entails shipping cost and a turn-around time of 2–3 weeks. Using other markers, such as SSR, is more expensive and cumbersome. The use of genotyping systems such as KASP in-house alleviates all these issues. Also, the issue of inefficient courier services in this part of the world, which often results in reduced or damaged perishable specimens, can be circumvented if a reasonably affordable system is available locally. More so, we re-purposed standard laboratory instruments for the genotyping workflow. For instance, the qPCR machine, which is mostly used for expression analysis, was adapted to KASP genotyping with the installation of appropriate software for SNP calling. Likewise, the Fluostar plate reader was used for plate-level DNA quantification in lieu of single sample analysis by Spectrophotometer. For these reasons it is imperative to devise a sustainable strategy for routine, cost-effective, and easily accessible genotyping services to complement these international outsourcing initiatives by providing in-house or local (regional) genotyping platforms, where possible, to accelerate the genotyping workflow. One such regional initiative in Africa is the Integrated Genotyping Support Services (IGSS) genotyping facility at Biosciences eastern and central Africa/International Livestock Research Institute ( BeCA/ILRI), Kenya. This strategy will allow breeders to outsource to a regional genotyping service provider or set up a core facility in-house. One factor that influences breeders' choice of genotyping platform is the level of throughput. Other factors considered are the data turn-around time, ease of data analysis (available informatics), reproducibility, flexibility, and cost per datapoint or cost per sample 10, 11 . For high and ultra-high throughput markers, breeders outsource to array- and sequenced-based genotyping service providers. These platforms are suitable for discovery applications and approaches requiring hundreds to thousands of samples to be genotyped with tens to thousands of markers, such as genome-wide association studies (GWAS), gene mapping, and large-scale genomic selection 10, 12 They are also suitable for genotyping a few samples with many markers (multiplexing), such as genetic diversity analysis or background selection. While multiplex platforms provide higher throughput with lower reagent consumption, it limits scientists to using a multiplexed set of several thousand single nucleotide polymorphisms (SNPs) per assay 13 . They are also demanding in informatics resources and presently produce datasets with a significant percentage of missing data 13 . The high cost per sample and the initial assay development time of highly multiplexed platforms can be problematic for crop improvement applications, usually requiring low- to medium-density markers 11 . For these low- to mid-density genotyping approaches, a uniplex SNP genotyping platform is appropriate 14 . Uniplex genotyping assays are low-throughput genotyping systems that are ideally flexible regarding assay design, ease of running, and cost-effectiveness 15 . These systems provide plant breeders with the flexibility to mix and match different SNPs for a given sample set. They allow breeders to use a smaller subset of informative SNPs such as functional SNPs and trait-specific haplotypes, thereby eliminating the generation of unintended datapoints when using fixed-array SNPs. Even though a range of uniplex SNP genotyping assays exists, the most competitive uniplex systems that have been successfully applied in crop improvement research are TaqMan 16– 19 , competitive allele-specific PCR (KASP) 11, 20 , Amplifuor 21 , and rhAmP 22 assays. These uniplex genotyping systems vary in reaction chemistry, detection method, and reaction format. Uniplex systems can either be outsourced or installed in-house. In this study, we utilised the KASP assay, as it is one of the most used assays among plant breeders and biologists 15, 19 . KASP is an endpoint PCR-based SNP genotyping method from KBiosciences, now LGC Biosearch Technologies, UK. KASP uses fluorescently-labelled allele-specific primers for the bi-allelic discrimination of SNPs and insertion-deletion mutations (INDELs) 23 . KASP was developed to reduce cost, mainly from probe design, and improve genotyping efficiency, becoming a preferred alternative to TaqMan 11, 24 . The KASP genotyping system has been successfully applied in crops such as maize 11, 15, 25 , wheat 10, 20, 26 , rice 27 , soybean 28 , peanut 29 , amongst others. KASP has developed into a global benchmark technology for genotyping crop plants 11, 23, 29– 31 following the validation of KASP markers across crops of global importance (such as maize - 1250 markers, wheat - 1864 markers, and rice - 2015 markers) by the Generation Challenge Program of the Integrated Breeding Platform 9 . The International Maize and Wheat Improvement Center (CIMMYT) has successfully utilised the 1,250 maize KASP markers for various genetic applications, including quantitative trait loci (QTL) mapping, marker-assisted recurrent selection (MARS), allele mining, and QC analysis 11 . The Maize Improvement Program of the International Institute of Tropical Agriculture (IITA) has generated over 2,000 datapoints using KASP in-house for different genotype analyses, including QC and MAS. However, some bottlenecks in the genotyping workflow slow down the genotyping process, delaying crop improvement: (1) method of sample collection and processing, (2) level of DNA extraction and quantitation, and (3) DNA-based genotyping 8 . Gedil and Menkir (2019) provided a thorough review of the Maize Improvement Program's (MIP) molecular marker-based crop improvement activities. However, reports of research accelerating the entire genotyping process by minimizing these bottlenecks and providing a cost-effective genotyping workflow suitable for small-scale breeders and laboratories in developing countries are lacking. This study aims to develop a genotyping workflow optimized for cost-effective and fast turn-around time which can be deployed by less sophisticated and reasonably equipped laboratories in developing countries, to accelerate maize improvement research. Methods Plant materials The overall genotyping workflow was applied in some experiments representative of the genotyping activities common in small to medium breeding programs. Table 1 below describes the plant materials used in each experiment. The genetic identity experiment was performed using four well-adapted maize varieties originating from IITA but regenerated at four locations. For the hybrid verification experiment, 60 maize F 1 progenies originating from five bi-parental crosses were used. Lines KS23-3, KS23-5, and KS23-6 are resistant to maize lethal necrosis (MLN) disease, while IITATZI1653 and IITATZI1667 are IITA-adopted elite maize lines with high PVA content. Another 330 F 1 plants originating from four sets of bi-parental crosses involving Striga-susceptible (TZdEEI 102, TZdEEI 99, TZdEEI 4, and TZdEEI 13) and Striga-resistant (TZEEI 29, and TZEEI 79) parents were also screened to identify true hybrids. A total of 70 PVA-QPM enriched maize inbred lines were genotyped to select lines harbouring the favourable allele for the crtRB1 gene associated with PVA content in maize. In the fourth breeding cycle of the maize enrichment project using marker-assisted backcrossing to introgress resistance to aflatoxin accumulation in elite tropical maize lines, we genotyped a total of 159 BC 1S 2 maize lines. We applied a 15% selection intensity to identify lines harbouring the favourable alleles of the QTLs associated with resistance to aflatoxin accumulation. These plants were grown in maize fields at IITA Ibadan, Nigeria. Table 1. Plant materials used for the experimentation of the optimized genotyping workflow. S/N Experiments Genotypes (Parental maize lines: traits) Population Development (crosses) No of samples 1 Genetic identity ●   SAMMAZ 15 (IWDC2SynF2): Medium maturing, good seed quality, high yield potential, tolerance to Striga hermonthica. (Y-6.9t/ha) ●   SAMMAZ 16 (TZLComp1SynW-1): Late maturing, good seed quality, high yield, resistance to Striga hermonthica. (6.4t/ha) ●   SAMMAZ 27 (EV99DT-W-STR): Drought tolerant and Striga resistant. (5.5t/ha) ●   SAMMAZ 39 (PVA SYN8): Intermediate-level pro-vitamin A content (6.4µg/g), high yield potential. (6.8t/ha) Performed using four well-adapted maize varieties originating from IITA but regenerated at four locations. Maize seedlings were grown in pots for about two weeks until they reached the three-four- leaf stage in a screen house at the Bioscience Center of IITA Ibadan, Nigeria. 20 maize lines resulting from 4 genotypes by 5 locations. 2 Hybrid verification: ●   KS23-3, KS23-5, and KS23-6: Maize lethal necrosis (MLN) resistant maize lines ●   IITATZI1653 and IITATZI1667: Maize lines with high PVA content ●   TZdEEI 102, TZdEEI 99, TZdEEI 4, and TZdEEI 13: Striga susceptible maize inbred lines ●   TZEEI 29 and TZEEI 79: Striga resistant maize inbred lines. Set 1a: KS23-3 x IITATZI1653; Set 2a: KS23-5 x IITATZI1653; Set 3a: KS23-6 x IITATZI1653; Set 4a: KS23-3 x IITATZI1667; Set 5: KS23-5 x IITATZI1667; Set 1b: TZEEI 29 x TZdEEI 99; Set 2b: TZdEEI 4 x TZEEI 79; Set 3b: TZEEI 79 x TZdEEI 13; Set 4b: TZdEEI 102 x TZEEI 29 Seedlings for the F 1 plants were grown in a maize field at IITA Ibadan, Nigeria. ●   Set a: 60 F 1 maize lines originating from five crosses involving three KS23 (MLN-resistant) lines and two PVA enriched maize lines. ●   Set b: 330 F 1 maize lines originating from four bi- parental crosses involving two Striga resistant maize lines and four Striga susceptible lines. 3 Marker-assisted selection ●   PVA-QPM enriched maize inbred lines were genotyped to select lines harbouring the favourable allele for the crtRB1 gene associated with PVA content in maize. ●   Backcross (BC 1S 2 ) maize lines in the fourth breeding cycle of the maize enrichment project; using marker-assisted backcrossing to introgress resistance to aflatoxin accumulation in elite tropical maize lines Ten plant stands per row were planted for each inbred, and leaf tissues were collected from each row for DNA extraction by bulking leaves from all ten plant stands per row. For the aflatoxin population, we applied a 15% selection intensity to identify lines harbouring the favourable alleles of the QTLs associated with resistance to aflatoxin accumulation. All maize lines were grown at IITA’s maize field, Ibadan, Nigeria. ●  70 PVA-QPM maize lines ●  159 BC 1S 2 maize lines Legend: PVA = Pro-vitamin A; QTL = Quantitative trait loci. Source of plant materials: Maize Improvement Program, International Institute of Tropical Agriculture (IITA) Headquarters, Ibadan, Nigeria. Sample collection and preparation, and DNA extraction and quantitation A total of 16 to 20 leaf discs were collected from young leaves of each tagged plant, directly into Corning 96-well Polypropylene 1.2 ml cluster tubes with strip caps (Merck, Germany) using Haris Uni-core 4.0 mm puncher and cutting mat (Merck, Germany). Two 4.0 mm stainless steel grinding balls (SPEX SamplePrep) were placed in each tube. Plant tissues were preserved on ice for transport from the field to the laboratory. They were stored in a -80°C freezer before lyophilising for 48 hours using FreeZone Freeze Dryer (Labconco) following the manufacturer's manual. Lyophilised leaf tissues were ground into powder by shaking at 1,500 strokes per minute for 1.5 min using an automated high-throughput tissue homogeniser, Geno/Grinder 2010 (SPEX SamplePrep). Genomic DNA was extracted from ground leaf tissues using a cetyltrimethylammonium bromide (CTAB)-based DNA extraction method as described by Diversity Array Technology (DArT) 32 with minor modifications ( Table 2). Dry leaf tissues were used instead of fresh ones; we included a 30-minute incubation period during the alcohol precipitation step; the DNA pellet was resuspended in a nuclease-free water and RNaseA solution. The DNA quality and quantity were determined by spectrophotometry using the FLUOstar Omega Microplate Reader (BMG LABTECH) following the manufacturer's manual. Table 2. DArT DNA extraction protocol with minor modification. The chemicals and reagents used were as outlined in the Diversity Array Technology (DArT) Plant DNA extraction protocol (Accessed on June 2, 2020). Extraction procedure:      1.  Aliquot freshly prepared, well-mixed "fresh buffer solution" and preheat in a 65°C water bath.      2.  Grind sample leaf discs in 1.2 ml cluster tubes using a Geno/Grinder 2010 (Spex Sample Prep) to a fine powder      3.  Add 500 μl buffer solution to dissolve the powder completely      4.  Incubate at 65°C for 1 hr, with gentle shaking      5.  Cool down for 5 min and add 500 μl of chloroform: isoamyl alcohol (24:1) mixture      6.  Mix well by gentle inversion for 30 min, and spin for 20 min, at 10,000 x g, at room temperature      7.  Transfer about 400 μl of the water phase to a fresh 1.2 ml tube, add the same volume of ice-cold isopropanol and invert the tube approximately ten times, nucleic acids should become visible      8.  Incubate for 30 min at -20 °C, and spin for 30 min, at 10,000 x g, at room temperature      9.  Discard supernatant, and wash pellet with 400 μl 70 % EtOH      10.  Discard EtOH, dry pellet and dissolve in 100 µl of nuclease-free water-RNAseA solution in a 90:10 ratio. KASP genotyping and data analysis The isolated genomic DNA was diluted to a working concentration of 30 ng/µl and used as template DNA for the KASP genotyping reaction. A total of 28 KASP SNPs were used to determine the selected maize varieties' genetic identity, while 10 KASP SNPs were used to verify true hybrids among the F 1 maize lines. The SNPs ( Table 3) were taken from a maize QC SNP panel 9 recommended by CIMMYT 7, 33 and chosen for their high polymorphic information content (PIC) and uniform maize genome coverage. Trait-specific KASP markers ( Table 4) were used to screen BC 1S 2 lines carrying the favourable allele for resistance to aflatoxin accumulation and identify inbred lines with high PVA content. The KASP reaction was performed in 96- and 384-well plates. For the 96-well plate, a total reaction volume of 10 µl consisting of 5 µl template DNA and 5 µl of the prepared genotyping mix (2×KASP master mix and primer mix) was used. In contrast, for the 384-well plate, a total reaction volume of 5 µl consisting of 2.5 µl template DNA and 2.5 µl of the prepared genotyping mix was used. All reaction was performed following the KASP manual (accessed on June 24, 2020). The KASP assay and master mix were purchased from LGC Biosearch Technologies (LGC Group). The amplification reaction was run in-house (Bioscience Centre of IITA Ibadan, Nigeria) using the LightCycler 480 II PCR System (Roche Life Sciences, Germany) and GeneAmp PCR System 9700 (Applied Biosystems, USA). The description of the parameters for the LC480 II qPCR machine is outlined in the LC480 operator’s manual. To perform the KASP genotyping experiment on the LC480 II machine, we used the Endpoint Genotyping Analysis module within the LightCycler software, adjusting the parameters as outlined in the KASP genotyping protocol provided by LGC Biosearch Technologies. The Endpoint genotyping analysis module is based on the use of dual hydrolysis probes, which are designed for wild-type and mutant target DNA and are labelled with different dyes (FAM and HEX). However, when using a non-qPCR machine (such as the GeneAmp PCR System 9700) for amplification, a third colour probe (ROX) normalizes the fluorescence measurement. The LightCycler software within the LC480 II machine determines the sample genotypes automatically by measuring the intensity distribution of the two probes after a PCR amplification step. The relative dye intensities are then visualized in a scatter (cluster) plot that discriminates them as wild-type, heterozygous mutant, or homozygous mutant samples. The LightCycler software automatically groups similar samples and assigns genotypes based on the intensity distribution of the two dyes. The KASP amplification conditions included one cycle of KASP unique Taq activation at 94°C for 15 min, followed by 36 cycles of denaturation at 94°C for 20 s, and annealing and elongation at 60°C (dropping 0.6°C per cycle) for 1 min. Endpoint detection of the fluorescence signal was acquired for 1 min at 30°C when using the LightCycler 480 II real time-PCR System or read using the FLUOstar Omega Microplate reader (BMG Labtech, SA) when using the GeneAmp PCR System 9700. For fluorescence detection, the filter combination for the Excitation and Emission wavelength of both dyes was set at 465 – 533 (FAM) and 523 – 568 (HEX), respectively, when using LC480 II, and 485 - 520 (FAM), 544 - 590 (HEX) and 584 - 620 (ROX) when using FLUOstar Omega Microplate reader. The genotype calls were exported from the LightCycler software as fluorescent intensities of each sample in ".txt" file format and imported for analysis in the KlusterCaller analysis software (LGC Biosearch Technologies). The KlusterCaller software adjusted the cluster plot axes to enable the proper calling of genotypes. The genotype calls were grouped as homozygous for allele X (allele reported by FAM, X-axis), homozygous for allele Y (allele reported by HEX, Y-axis), heterozygous (alleles reported by FAM and HEX, between X- and Y-axis), or uncallable. The result from the KlusterCaller was exported in two file formats (".csv" and ".txt"). The ".csv" file was imported into the SNPviewer2 version 4.0.0 software (LGC Biosearch Technologies), where the cluster plot image was viewed and downloaded for publication. The genotype calls in the ".txt" file were used to calculate the genetic distance using the PowerMaker 3.25 statistical software 34 . Table 3. List of KASP single nucleotide polymorphisms (SNPs) used in the QC experiments. SNP ID Linkage group Position (cM) Allele X Allele Y Trait category Analysis Dataset ae1_7 5 79 A G QC GID & HV GCP/IBP-Maize PHM15331_16 10 28 A G QC GID GCP/IBP-Maize PHM2438_28 4 12 A G QC GID GCP/IBP-Maize PHM2770_19 10 36 A C QC GID GCP/IBP-Maize PHM3466_69 6 108 A G QC GID GCP/IBP-Maize PHM5181_10 9 26 C T QC GID & HV GCP/IBP-Maize PHM5502_31 3 58 A G QC GID & HV GCP/IBP-Maize PZA00413_20 3 60 A C QC GID & HV GCP/IBP-Maize PZA00726_10 4 55 A C QC GID GCP/IBP-Maize PZA01216_1 1 116 A G QC GID & HV GCP/IBP-Maize PZA01456_2 10 61 A G QC GID GCP/IBP-Maize PZA01477_3 4 81 C T QC GID GCP/IBP-Maize PZA01533_2 7 112 A G QC GID GCP/IBP-Maize PZA01885_2 2 115 A G QC GID & HV GCP/IBP-Maize PZA01919_2 10 44 C G QC GID & HV GCP/IBP-Maize PZA02090_1 3 15 A T QC GID & HV GCP/IBP-Maize PZA02164_16 5 70 A G QC GID & HV GCP/IBP-Maize PZA02269_3 1 149 C T QC GID & HV GCP/IBP-Maize PZA02358_1 4 31 A G QC GID GCP/IBP-Maize PZA02378_7 2 64 A G QC GID GCP/IBP-Maize PZA02741_1 1 91 C T QC GID GCP/IBP-Maize PZA02746_2 8 94 G T QC GID GCP/IBP-Maize PZA02779_1 4 108 A G QC GID & HV GCP/IBP-Maize PZA03135_1 8 57 A C QC GID & HV GCP/IBP-Maize PZA03363_1 7 49 A G QC GID & HV GCP/IBP-Maize PZA03605_1 10 75 A G QC GID GCP/IBP-Maize PZB01658_1 6 28 A T QC GID & HV GCP/IBP-Maize sh1_12 9 18 A G QC GID & HV GCP/IBP-Maize LEGEND: QC = Quality control; GID = Genetic Identity; HV = Hybrid verification; GCP/IBP = Generation Challenge Program/Integrated Breeding Platform. Source: Integrated Breeding Platform (Accessed June 26, 2020). Table 4. List of trait-specific KASP single nucleotide polymorphisms SNPs used in the MAS experiment. SNP ID Chromosome No. FAM allele HEX allele Trait category analysis Source S1_85016181 1 C G Aflatoxin MAS CIMMYT/IITA S3_14863214 3 G A Aflatoxin MAS CIMMYT/IITA S3_90027035 3 A G Aflatoxin MAS CIMMYT/IITA S3_90023939 3 T A Aflatoxin MAS CIMMYT/IITA S3_179639685 3 C G Aflatoxin MAS CIMMYT/IITA S3_14229695 3 T C Aflatoxin MAS CIMMYT/IITA S5_182519023 5 A G Aflatoxin MAS CIMMYT/IITA S5_63229636 5 C A Aflatoxin MAS CIMMYT/IITA S5_198883041 5 T A Aflatoxin MAS CIMMYT/IITA PHM12859_7 3 C T Aflatoxin MAS CIMMYT/IITA PZA02792_16 5 T C Aflatoxin MAS CIMMYT/IITA MZA4145_18 3 A G Aflatoxin MAS CIMMYT/IITA snpZM0015 10 A G PVA MAS CIMMYT LEGEND: MAS = Marker-assisted selection; PVA = Provitamin A; CIMMYT = International Maize and Wheat Improvement Center; IITA = International Institute of Tropical Agriculture. Source data The list of KASP SNPs for genotyping maize was obtained freely from the Integrated Breeding Platform website. The trait-specific KASP SNPs (Supplementary Table 1, Underlying data) and QC KASP SNPs (Supplementary Table 2, Underlying data) were purchased as KBDs (KASP-by-Design) from LGC Biosearch Technologies, UK, for use in our laboratory. Results Optimising in-house genotyping workflow Our laboratory's routine sampling procedure spans seven days, from plant sampling and preparation to DNA extraction and quantitation. We present an expedited workflow ( Figure 1) that ensures a good sample tracking system. Firstly, barcoding software, barcode readers, barcode labels, and barcode printers were introduced to facilitate sample tracking and data management. Waterproof/tear-proof tags and labels designed using BarTender barcoding software (Seagull Scientific) were printed using ZT230 Printer (Zebra, USA) and attached to plants before sample collection. Plate maps created in the BarTender software were linked to the sample location on the field and in the lab storage facility. Next, young plant leaf tissues were collected by punching leaf discs directly into the 96-well 1.2 mL polypropylene cluster tubes in wet-ice cooler bags, which reduced the sampling time and the time required for freeze-drying. Figure 1. Diagram showing improvement to minimize bottlenecks in the genotyping workflow. The sample DNA was extracted using the DArT DNA extraction protocol, slightly modified to maximise reagent and increase throughput, by using a reduced volume of reagents optimised to extract maize DNA from a smaller amount of leaf tissue (16–20 leaf discs, 4.0 mm). We also used freeze-dried leaf tissue, which allowed grinding using an automated high-throughput tissue homogeniser, Geno/Grinder 2010, with a 384-samples grinding capacity (4 × 96-sample plates) in two minutes. The UV absorbance protocol for the FLUOstar Omega microplate reader (BMG LABTECH) was used to measure the concentration and purity of the DNA samples. By using this method, the 637 DNA samples were quantified in less than 10 minutes. The DNA purity (A260/A280 ratio) ranged from 1.7 to 2.0, with an average concentration of 985 ng/µl. Following the optimized workflow, the total time from sampling and processing to DNA extraction and quantitation of the 637 leaf samples was reduced from seven to five days. In order to optimise and use the KASP system in-house, KASP assays and allele-calling software (KlusterCaller) were purchased from LGC, UK. The amplification parameters on the compatible PCR (GeneAmp 9700) and real-time PCR machines (Roche LightCycler 480 II) were optimised. Microtiter 96- and 384-well plates compatible with the different machines were acquired from Roche, Germany. We also optimised the FLUOstar Omega microplate reader for fluorescence measurement of amplified products following the manufacturer's manual. Then, we ran a KASP trial kit provided freely by LGC Biosearch to test for functionality with the different amplification equipment. Application of the optimised genotyping workflow Following the KASP set-up, we genotyped plant samples for QC and MAS in-house, with low-density markers. The QC genotyping ensured on-time identification of errors and mislabeling in inbred lines and false hybrids in F 1 maize breeding populations. Using the in-house KASP genotyping platform significantly reduced genotyping cost and time compared to outsourcing. Genetic identity. Using a subset of 28 maize QC KASP SNPs, we were able to identify the genetic origin of a set of twenty well-adapted maize varieties originating from IITA, which were regenerated at four other locations. Genetic identification was performed using the original maize varieties' molecular marker profile and the genetic distance approach. Seed sources having <5% genetic distance were considered the same. The genetic distance among the four original maize lines, and between lines from IITA and each of the four seed sources, was calculated using PowerMaker 3.25 statistical software. The genetic distance among the four designation lines from IITA ranged from 0.0563 to 0.1239, indicating that the lines were different. The genetic distance among the different seed sources of the same line designation was: 0.0105-0.0314 (SAMMAZ15), 0.0105–0.0418 (SAMMAZ16), 0.0105–0.0837 (SAMMAZ27), and 0.000–0.0563 (SAMMAZ39). The SNPviewer, a tool that enables viewing genotyping data as a cluster plot, was used to view and generate an image of the genotyping result. The SNPviewer image showed that designated lines from three out of the four seed sources grouped with lines from IITA ( Figure 2). The dendrogram image ( Figure 3) also showed a grouping of different seed sources of the same line designation except for SAMMAZ39-1, SAMMAZ16-3, and SAMMAZ27-4. This clustering pattern indicates that all seeds from the same line had a common origin. SAMMAZ27-4 appeared to be genetically distant from SAMMAZ27-IITA by 0.0837. However, it grouped with SAMMAZ15 ( Figure 3: blue circle), suggesting a possible mislabeling or mix-up of seeds during harvesting and storage. SAMMAZ16-2 and SAMMAZ39-1 grouped on a different tree limb ( Figure 3: red circle), indicating possible pollen contamination or seed mix-up during handling. Figure 2. SNPviewer screenshot showing clustering of IITA's maize lines with same lines from three out of four seed sources. ( a) Sammaz15-2, -3, and -4 grouped with IITA’s Sammaz15 (blue dots) using SNP PZA02746_2; ( b) Sammaz39-2, -3, and -4 grouped with IITA’s Sammaz39 (blue dots) using SNP PZB01658_1. For each SNP marker, blue dots represent homozygous genotypes, green dots represent heterozygote genotypes, and the black dots represent no-template controls (NTC) as indicated on the left side of each image. Figure 3. Neighbor-joining tree for four maize varieties taken from five seed sources based on genetic distance, performed with 1,000 bootstrap. Bootstrap values are indicated on the tree branches. The suffixes "-1", "-2", "-3", "-4", and "-IITA", after line name indicate seed source 1, 2, 3, 4, and IITA. Hybrid verification. In another QC experiment using our workflow, we screened two groups of F 1 plants for hybrid verification, including their parental inbred lines, with 10 KASP SNP markers. The parental inbred lines were screened with an initial 50 KASP SNP taken from a defined panel of maize QC KASP markers to identify polymorphic markers. Only 10 KASP markers polymorphic between the parental lines were used to screen the F 1 plants to verify their parentage. The KASP genotyping assay was useful in distinguishing between the parental genotypes and identifying the true hybrid lines. Cluster analysis of Group1 F 1s ( Figure 4) grouped the genotypes into three clusters. The heterozygous F 1 progenies were in the middle of the plot, and the homozygous parental inbred lines diverged from each other (along the X- and Y-axis of the plot) for all markers. The genotyping result ( Table 5) and the clustering pattern indicate that the F 1 progenies were true hybrids. Similar clustering was observed among F 1s in Group 2 except in Set 3b, where 38 F 1s grouped with parental genotypes. The homozygous F 1s could be due to contamination from foreign pollens during the crossing in the field or seed mix-up during storage or planting. Figure 4. SNPviewer screenshot showing the result of hybrids verification in two sets of F 1 Plants. ( a) Genotyping 12 F1 lines produced from a cross between KS23-5 and IITATZI1653, using SNP PZA03135_1. ( b) Genotyping of 12 F1 lines produced from a cross between KS23-5 and IITATZI1667, using SNP PZA02779_1. For each SNP marker, blue dots represent homozygous parental genotype reported by FAM, red dots represent homozygous parental genotype reported by HEX, green dots represent heterozygous hybrid genotypes, and the black dots represent no-template controls (NTC). Legend: FAM = Carboxyfluorescein; HEX = Hexachloro-fluorescein. Table 5. KASP genotyping result for the hybrid verification experiment. Subject ID: KS23-6 IITATZI1653 SCH-1 SCH-2 SCH-3 SCH-4 SCH-5 SCH-6 SCH-7 SCH-8 SCH-9 SCH-10 SCH-11 SCH-12 SNP ID ae1_7 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PHM5181_10 C:C T:T T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C PZA01216_1 G:G A:A G:A G:A G:A A:A G:A G:A G:A G:A G:A G:A G:A G:A PZA01885_2 G:G A:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA01919_2 C:C G:G G:C G:C G:C G:G G:C G:C G:C G:C G:C G:C G:C G:C PZA02090_1 T:T A:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A PZA02779_1 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA03135_1 A:A C:C C:A C:A C:A C:C C:A C:A C:A C:A C:A C:A C:A C:A PZA03363_1 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZB01658_1 T:T A:A T:A T:A T:A A:A T:A T:A T:A T:A T:A T:A T:A T:A True Hybrid? Parent 1 Parent 2 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Subject ID: KS23-3 IITATZI1667 SCH-1 SCH-2 SCH-3 SCH-4 SCH-5 SCH-6 SCH-7 SCH-8 SCH-9 SCH-10 SCH-11 SCH-12 SNP ID ae1_7 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA00413_20 C:C A:A C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A PZA01885_2 G:G A:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA01919_2 C:C G:G G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C PZA02269_3 C:C T:T T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C PZA02779_1 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA03135_1 A:A C:C C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A PZA03363_1 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZB01658_1 T:T A:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A sh1_12 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A True Hybrid? Parent 1 Parent 2 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Subject ID: KS23-5 IITATZI1667 SCH-1 SCH-2 SCH-3 SCH-4 SCH-5 SCH-6 SCH-7 SCH-8 SCH-9 SCH-10 SCH-11 SCH-12 SNP ID PZA00413_20 C:A A:A A:A A:A C:A C:A C:A A:A C:A C:A A:A C:A C:A C:A PZA01885_2 G:G A:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA01919_2 C:C G:G G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C PZA02164_16 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA02269_3 C:C T:T T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C PZA02779_1 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA03135_1 A:A C:C C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A PZA03363_1 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZB01658_1 T:T A:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A sh1_12 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A True Hybrid? Parent 1 Parent 2 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Subject ID: KS23-3 IITATZI1653 SCH-1 SCH-2 SCH-3 SCH-4 SCH-5 SCH-6 SCH-7 SCH-8 SCH-9 SCH-10 SCH-11 SCH-12 SNP ID ae1_7 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PHM5502_31 G:G G:G G:G G:G G:G G:G G:G G:G G:G G:G G:G G:G G:G G:G PZA01885_2 G:G A:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA01919_2 C:C G:G G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C PZA02269_3 C:C T:T T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C PZA02779_1 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA03135_1 A:A C:C C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A PZA03363_1 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZB01658_1 T:T A:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A sh1_12 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A True Hybrid? Parent 1 Parent 2 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Subject ID: KS23-5 IITATZI1653 SCH-1 SCH-2 SCH-3 SCH-4 SCH-5 SCH-6 SCH-7 SCH-8 SCH-9 SCH-10 SCH-11 SCH-12 SNP ID PZA00413_20 C:A C:C C:A C:C C:A C:C C:A C:A C:C C:C C:C C:C C:C C:A PZA01885_2 G:G A:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA01919_2 C:C G:G G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C G:C PZA02164_16 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA02269_3 C:C T:T T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C T:C PZA02779_1 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZA03135_1 A:A C:C C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A C:A PZA03363_1 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A PZB01658_1 T:T A:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A T:A sh1_12 A:A G:G G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A G:A True Hybrid? Parent 1 Parent 2 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Legend: SCH = Single cross hybrid, and the suffixes ‘-1 to -12’ represent the number of F1s genotyped for each cross. Nonetheless, the KASP genotyping assay suffers some genotyping errors, especially during the automatic calling of genotypes. For instance, one F1 line (SCH-4) developed from the bi-parental cross, KS23-6 and IITATZI1653, appeared to cluster with the parent 2 (IITATZI1653) when genotyped with marker PZB01658_1 ( Figure 5). The datapoint representing IITATZI1653 ( Figure 5, information in the yellow square) was plotted higher up, away from the X-axis, which brought it closer to the datapoint representing SCH-4 plotted slightly away from the other F1s in the middle. Because genotype calls are generated based on the relative position of datapoints on the plot, SCH-4 was automatically called as the nearby parental genotype, A:A, which was an error seeing that line SCH-4 was heterozygous (true hybrid) for the rest of the markers. The upward positioning of line IITATZI1653 away from the X-axis could be possibly due to trace contamination of line IITATZI1653 sample DNA with line KS23-6 sample DNA during sample preparation. A monomorphic marker is seen in the genotyping of F1 lines developed from the bi-parental crosses KS23-3 x IITATZI1653 using marker PHM5502_31. Figure 5. SNPviewer screenshot showing the result of hybrid verification of F 1 Plants. Genotyping 12 F 1 lines produced from a cross between KS23-6 and IITATZI1653, using SNP PZB01658_1.The blue dots represent homozygous parent 2 (IITATZI1653) genotype reported by FAM, red dots represent homozygous parent 1 (KS23-6) genotype reported by HEX, green dots represent heterozygous (F 1s) genotypes, and the black dots represent no-template controls (NTC). Legend: SCH-4 = Single cross hybrid (F 1) sample 4; FAM = Carboxyfluorescein; HEX = Hexachloro-fluorescein. Marker-assisted backcrossing. We performed multiple field selections annually by applying our workflow in MAS projects, which accelerated the maize breeding process. For instance, in the MABC project, a set of trait-specific KASP SNPs was used to select 24 BC 1S 2 maize lines potentially introgressed with resistance to aflatoxin accumulation after four selection cycles in less than two years. Potentially introgressed lines are undergoing field evaluation under artificial infestation for resistance to aflatoxin accumulation. The result of the MAS of high PVA lines, on the other hand, identified nine out of 70 inbred maize lines harbouring favourable alleles of the crtRB1 gene, which is associated with high PVA content in maize. Discussion There are different methods of plant tissue sampling, including collecting samples in silica gel 35 , NaCl/CTAB 36 , alcohol 37 , blotter paper, gel pack, dry ice, and liquid nitrogen 38 . These methods provide reasonably good quality and quantity of DNA for molecular marker genotyping. However, deciding which method to use is based on the number of samples and distance from the field to the laboratory 38 . We routinely use wet ice in Styrofoam boxes and cooler bags. It is cost-effective and suitable for close-proximity sample collection, and leaf samples are preserved by freeze-drying 39 before DNA extraction. We collected fresh leaf tissues directly into 96-well extraction tubes rather than the traditional jute or tea bags, which means our procedure provides high throughput sampling. This sampling process also ensured that sample DNA was not degraded by prolonged exposure of leaf tissues to moisture as it occurs in post-freeze drying cutting of leaf tissues stored in jute and tea bags. Our protocol aimed to extract high-quality DNA suitable for KASP genotyping from a smaller amount of leaf tissues. The reduced sample volume lowered the cost of reagents and the time for DNA extraction. The automated grinding in 96-well plates increased throughput and minimised the time required for manual grinding. Thus, this method would benefit MAS breeding programs that often screen thousands of plant samples each season 40 . A similar high-throughput result was achieved by Anderson et al. (2018) 41 . They optimised the DNA extraction method by Whitlock et al. (2008) 42 , used a 96-well plate for extraction and achieved a consistent yield across the plate with a low failure rate. Three steps of the original DArT DNA extraction method were slightly modified to achieve our aim. The first modification was made in the sample grinding step, where we used dried leaf tissues instead of fresh ones;—using dried samples enabled high-throughput grinding using a Geno/Grinder, reducing the time used in manual grinding with liquid nitrogen. The second modification was at the alcohol precipitation step: the sample tubes were incubated at -20°C for 30 minutes after adding the ice-cold isopropanol, instead of only mixing by inversion. This incubation is necessary for slow and complete DNA precipitation. The third modification was reconstituting the DNA pellet: we dissolved the DNA in a solution of nuclease-free water and RNaseA instead of using a Tris-EDTA (TE) buffer to prevent the chelating effect of EDTA on Mg 2+ during PCR 43, 44 . The success of the KASP genotyping experiment is dependent on the quality and quantity of genomic DNA. Usually, a final minimum DNA concentration of 5 ng/µl is required for maize, to generate clear and consistent allele calls using the KASP assay 45 . Our slightly modified DNA extraction method provided good quality DNA, suitable for KASP genotyping. Jain et al. (2013) extracted suitable quality DNA from honey that was amplifiable by PCR, using an optimised DArT DNA extraction protocol. Some commonly used DNA quality and quantity analysis methods include agarose-gel electrophoresis, fluorescence, and Ultraviolet (UV) absorbance-based measurement 38 . Fluorescence-based measurement using DNA-binding dyes such as PicoGreen is fast, sensitive, and dsDNA-specific; however, it comes with the DNA-binding reagent's added cost 46, 47 . Agarose gel electrophoresis is laborious and carries the risk of exposure to hazardous chemicals like ethidium bromide 47 . The UV absorbance measurement is the most common DNA quantitation method. It is based on DNA absorbing UV light at a specific wavelength; DNA concentration is calculated by measuring the absorbance at 260nm and using the relationship A260 of 1.0 equals 50 µg/ml pure dsDNA 46 . DNA purity is estimated based on two UV absorbance ratios: A260/A280 ≥1.7 and A230/A260 ≥ 1.5 for pure DNA 46 . Our workflow optimized the nucleic acid quantitation method to a high throughput using a microplate reader and 96- and 384-well plates. The FLUOstar microplate reader uses ultrafast UV/Vis spectrometers for absorbance measurements, measuring 96 samples (96-well plate) to 384 samples (384-well plate) simultaneously within one second per well. It combines speed and the acquisition of complete absorbance spectra (220 to 1000 nm), making it ideal for nucleic acid quantification 48 . Although outsourcing KASP offers a lower cost per data point, this lower genotyping cost is usually driven by a high volume of samples, impracticable for most MAS projects genotyping smaller sample volumes with select markers 49 . Our in-house genotyping system provides reduced cost, mainly from logistics, and faster data turn-around times, ultimately accelerating the genotyping workflow. A few studies serve as the benchmark for QC analysis in maize using the KASP genotyping system. Semagn et al. (2012) suggested using a subset of 50 to 100 KASP markers for routine QC; Chen et al. (2016) used a smaller subset of markers (10 markers) to assess mislabeling of entries across a panel of CIMMYT Maize Lines (CMLs) achieving up to 99% detection probability. The latter also proposed using a rapid QC approach, with a smaller subset of markers, to ensure effective QC, lower genotyping costs, and shorten data turn-around time during seed production. Using a subset of markers, we were able to identify seed mix-up and labelling errors. For instance, the grouping of SAMMAZ27-4 with SAMMAZ15 ( Figure 3: blue circle) suggests a possible mislabeling or mix-up of seeds during harvesting and storage. Also, the grouping of SAMMAZ16-2 and SAMMAZ39-1 ( Figure 3: red circle) indicates possible pollen contamination or seed mix-up during handling. Similar errors due to seed mix-up and contamination were reported in Semagn et al. (2012), where 50 KASP SNPs were used to determine genetic identity among two to four seed sources of the same inbred line. Ertiro et al. (2015) also reported a high discrepancy in genetic purity and identity by the origin of seed sources irrespective of the genotyping platform used. They concluded that using a small subset of pre-selected high-quality markers was sufficient for performing QC analysis using low-marker density genotyping platforms like KASP. This study showed that the rapid QC method using 28 KASP SNPs efficiently distinguished the four maize varieties taken from five seed sources. Hybrid verification is often performed during seed production or population breeding to confirm that a particular hybrid is derived from the intended parental lines (free from contamination by foreign pollens). Reducing the data turn-around time is essential to ensure that an accurate hybrid is selected to be carried forward in breeding programs or dissemination to farmers in seed production 33 . A reduced turn-around time also saves the cost of inputs applied to undesired genotypes since they can be discarded as soon as they have been identified upon genotyping. Our expedited workflow was able to achieve this. The possibility of contamination by self-pollination or foreign pollen exists; as such, hybrid verification is necessary to enable a seed producer to check whether accurate crosses are made for the production of the hybrid; this increases the confidence of the end-users on the quality and integrity of seeds produced 33 . Our results showed that 10 KASP markers were sufficient in distinguishing between maize parental inbred lines and identified true hybrid lines, residual contaminations, and possible sampling errors. A small subset of KASP markers has also been used to verify hybrids in other plant species. Patterson et al. (2017) 50 achieved a highly accurate picture of Myriophyllum species distribution dynamics in North American lakes by genotyping 39 individuals from both parental watermilfoil and their hybrids, using a subset of three KASP markers. Osei et al. (2020) 51 used 38 KASP markers to screen tomato genotypes to identify true F 1 hybrids for the possible development of inbreds with long shelf life through marker-assisted backcrossing (MABC). Following our optimised workflow, we were able to identify high-PVA maize lines harbouring the favourable allele of the crtRB1 gene, which could serve as donor lines for the maize PVA breeding program. The KASP-based selection of aflatoxin-resistant maize lines promises to fast-track the development of tropical lines resistant to aflatoxin, which will contribute to genetic gain in maize production. Similar success was achieved by the Biotechnology Center of the University of California, Davis, USA, where KASP SNPs associated with Phytophthoria capsici resistance were used to identify and selectively breed pepper strains 52 . So far, we have generated over 2,000 data points using our in-house genotyping workflow. Applying our optimised workflow to the QC and MAS experiments outlined above reduced the volume of reagents and consumables used, shortened the data turn-around, and ultimately accelerated the crop improvement process. Conclusions This study describes for the first time an improvement of an entire conventional DNA-based genotyping workflow, including the benchmark KASP genotyping platform in-house in our facility to fast-track molecular marker-based selection for crop improvement. We acknowledge the initial capital investment to procure some of these instruments. However, it is not always necessary to equip each lab or breeding program. The use of shared facilities locally and regionally, and the re-purposing of existing equipment such as the PCR machine and the spectrophotometer, help overcome the high cost of essential instruments. The improved genotyping workflow promises to accelerate the marker-assisted selection process and push crop improvement activities to attain the yield potential over a shorter time period. The result of this work can be readily adopted by national institutions, public and small plant breeding laboratories in developing countries to accelerate molecular marker-based genotyping for crop improvement activities, including QC and MAS. The results will also be helpful to accelerate the QC activities of seed producers and facilitate cultivar identification and adoption-tracking studies. Data availability Underlying data Figshare: SNP data for "Developing and deploying an efficient genotyping workflow for accelerating maize improvement in developing countries.", https://doi.org/10.6084/m9.figshare.17157914 53 . This project contains the following underlying data: - Supplementary Table 1. List of trait-specific KASP SNPs used in the MAS experiment with sequence information - Supplementary Table 2. List of KASP SNPs used in the QC experiments with sequence information Data are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication). Acknowledgements The authors are grateful for the support provided by colleagues in MIP and Bioscience Center, IITA. The authors thank Oluyinka Ilesanmi for designing and printing the barcode labels for tagging plant materials on the field, and Dr Ryo Matsumoto for the images. 10.21956/gatesopenres.15059.r33596 Reviewer response for version 3 Sawers Ruairidh J H 1Referee 1 Plant Science, The Pennsylvania State University - University Park Campus, University Park, PA, USA 13 7 2023 Copyright: © 2023 Sawers RJH 2023 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendationapprove A nice survey of the options and applications for genotyping in breeding programs with useful details and experience of establishing an in-house KASP genotyping platform. The manuscript is strong on specific details that will be helpful to other researchers attempting similar work. I’ve included a few general comments/thoughts that might be useful to consider. Other reviewers have commented on the usefulness of a more complete economic breakdown. I’m not sure specific numbers (such as pricing given in the introduction) are necessarily that useful as they will no-doubt change. However, I would agree that more discussion of the relative costs of given approaches and different scales would be helpful. More generally, although services for the same technology are compared, less attention is given to different approaches. Indeed, given the context of extending access to molecular platforms, if would be informative to say more about the costs of using these technologies at all in comparison to conventional methods. For the specific KASP application, more could be said about the costs of primer design – especially if not using a crop well served with existing sequences – and synthesis. The case studies are informative but would benefit from providing more information on the markers (for example, map position) and the sample genotypes. More could be said about the selection of markers, and specifically more discussion of how many markers are actually needed for a given application (based on these empirical examples, as much as prior literature). Results for selected markers are presented visually. Are these “typical” examples? Can more of a summary be given as to how many markers “worked”, and how reproducible and reliable the results were? The results presented clearly separate genotypic classes (except for one highlighted individual). Was this always the case? Was calling of heterozygotes always robust? Was any additional confirmation performed? Would it be possible to estimate the rate of miscalling, either per marker or generally for the platform?    The cluster/calling of the KASP signal is most robust when each genotypic class is represented by multiple individuals. In Fig.2 the sample size is small, and the homozygous T/T class is not represented. In isolation such sampling may complicate “calibration” of the heterozygous calls. Similarly, in Fig. 4 only a single sample is typed for each of the two parental homozygotes. The examples presented look nice and clear, but was this always the case for all markers typed on these samples? It wasn’t entirely clear what was done in the dendrogram in Fig. 3. How many markers were used? How was this number/set determined? Were they spread throughout the genome? Do these lines show a level of heterozygosity? As above, was there any ambiguity/error in calling? It’s a small detail, but at times the use of the term “line” was a little confusing. It can help to keep “line” to refer to inbred (highly homozygous) stocks. While an F1 is typically a cross between two lines, don’t refer the F1 as an “F1 line” etc. The expectation with regard to heterozygosity – and the requirement to accurately make het calls – is directly relevant to selection and use of a genotyping platform. Is the rationale for developing the new method (or application) clearly explained? Yes Is the description of the method technically sound? Yes Are the conclusions about the method and its performance adequately supported by the findings presented in the article? Partly If any results are presented, are all the source data underlying the results available to ensure full reproducibility? Partly Are sufficient details provided to allow replication of the method development and its use by others? Partly Reviewer Expertise: Maize diversity, genetics and genomics; crop nutrition; mycorrhizal symbiosis I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. 10.21956/gatesopenres.15059.r33600 Reviewer response for version 3 Adhimoolam Karthikeyan 12Referee https://orcid.org/0000-0002-6270-5597 1 Horticulture, Jeju National University, Jeju-si, Jeju-do, South Korea 2 Tamil Nadu Agricultural University, Madurai, Tamil Nadu, India 30 6 2023 Copyright: © 2023 Adhimoolam K 2023 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendationapprove This study demonstrates the improved genotyping workflow for maize improvement in developing countries. I recommend the manuscript for indexing. However, I recommend drawing a better workflow figure instead of Figure 1. And present some data (i.e., Table) to support the cost-effectiveness. Also, I suggest the authors improve the language. Is the rationale for developing the new method (or application) clearly explained? Yes Is the description of the method technically sound? Yes Are the conclusions about the method and its performance adequately supported by the findings presented in the article? Yes If any results are presented, are all the source data underlying the results available to ensure full reproducibility? Yes Are sufficient details provided to allow replication of the method development and its use by others? Yes Reviewer Expertise: Plant molecular breeding and plant genomics I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. 10.21956/gatesopenres.15059.r32358 Reviewer response for version 3 Basnet Bhoja R. 1Referee https://orcid.org/0000-0002-1693-4807 1 International Maize and Wheat Improvement Center (CIMMYT), Texcoco, Mexico 16 9 2022 Copyright: © 2022 Basnet BR 2022 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendationapprove Dear authors, thank you for sincerely handling the reviews and putting your best effort into addressing those comments and concerns. I do not have further queries. I believe we let readers, the fellow researchers, and the broader scientific community judge the merit of this research. Is the rationale for developing the new method (or application) clearly explained? Partly Is the description of the method technically sound? Yes Are the conclusions about the method and its performance adequately supported by the findings presented in the article? Yes If any results are presented, are all the source data underlying the results available to ensure full reproducibility? Partly Are sufficient details provided to allow replication of the method development and its use by others? Partly Reviewer Expertise: Plant breeding and genetics, genomics, quantitative genetics, and breeding program optimization. I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. 10.21956/gatesopenres.14950.r32187 Reviewer response for version 2 Basnet Bhoja R. 1Referee https://orcid.org/0000-0002-1693-4807 1 International Maize and Wheat Improvement Center (CIMMYT), Texcoco, Mexico 8 7 2022 Copyright: © 2022 Basnet BR 2022 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendationapprove-with-reservations It should go through one more round of revision. Please address the following concerns: Introduction P3L3: "However, breeders often want to fingerprint a few dozen lines urgently for identity or parentage analysis for example". Please rephrase this sentence. Please provide well-articulated one or two objectives of this research (towards the end of the Introduction section). Introduction section P6L1: "Here we utilized the KASP assay" - revise as "In this study, we utilized...". Introduction P6L6: reference 27 does not provide any account of soybean - the paper is about maize. Please verify this information and correct it as needed. Methods P1L3: "Well-adopted" should be changed to "well-adapted". Method P1 Last sentence: The plants were grown, not raised. Table 1: This is a piece of good information. However, I ask you to provide the exact number of genotypes and the samples within each genotype for all the groups (please add additional columns as needed).  Use of BC1S2 does not seem to be reliable in this study unless you verify the selection with phenotypic data to estimate the sensitivity and specificity of the marker assessment. However, it doesn't seem to harm the manuscript either. One important analysis I would like to suggest to add to this study is HYBRID VERIFICATION. Please prepare a data table for each sample - identified within each genotype (F1), such as order the column as F1 cross name / no, sample #, Marker gen _P1, Marker gen_P2, Observed F1 gen, True Hyb (Yes or no), if not if the F1 gen is observed as maternal or paternal type, etc. Then please assess the true to hybrid types or % hybridity within each genotype (using samples within cross) and across all samples. Then also revise your results section with a detailed discussion on how this assay is helpful to discriminate true-to-type hybrids and also describe potential bias caused by the assay itself - genotyping error or so using data on samples within each genotype.  Did you sample multiple samples within each plant? If so, please revise the results section accordingly.  Is the rationale for developing the new method (or application) clearly explained? Partly Is the description of the method technically sound? Yes Are the conclusions about the method and its performance adequately supported by the findings presented in the article? Yes If any results are presented, are all the source data underlying the results available to ensure full reproducibility? Partly Are sufficient details provided to allow replication of the method development and its use by others? Partly Reviewer Expertise: Plant breeding and genetics, genomics, quantitative genetics, and breeding program optimization. I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above. Offornedo Queen International Institute of Tropical Agriculture, Nigeria 28 7 2022 The authors thank the reviewers for taking the time to review the manuscript and raise critical issues. We trust that addressing these issues will immensely improve the paper. Below are the responses to the concerns raised. Reviewer 2 comment: Introduction P3L3: "However, breeders often want to fingerprint a few dozen lines urgently for identity or parentage analysis, for example". Please rephrase this sentence. Author's response: The sentence has been rephrased: "Even though large volume sizes can be consolidated and shipped for genotyping, there are times when breeders and partners may want to fingerprint a few dozen lines for identity or parentage analysis for quick decision making." Reviewer 2 comment: Please provide well-articulated one or two objectives of this research (towards the end of the Introduction section). Author's  response: The objective of the research has been rephrased "This study aims to develop a genotyping workflow optimized for cost-effective and fast turn-around time that can be deployed by less sophisticated and reasonably equipped laboratories in developing countries, to accelerate maize improvement research." Reviewer 2 comment: Introduction section P6L1: "Here we utilized the KASP assay" - revise as "In this study, we utilized...". Introduction P6L6: reference 27 does not provide any account of soybean - the paper is about maize. Please verify this information and correct it as needed. Author's response: The phrase has been corrected. The misplaced reference has been replaced. A more suitable reference has been attached to the statement. " Shi, Z., Liu, S., Noe, J.  et al. SNP identification and marker assay development for high-throughput selection of soybean cyst nematode resistance.  BMC Genomics 16, 314 (2015). https://doi.org/10.1186/s12864-015-1531-3 " Reviewer 2 comment: Methods P1L3: "Well-adopted" should be changed to "well-adapted". Authors' response: The phrase has been modified accordingly. Reviewer 2 comment: Method P1 Last sentence: The plants were grown, not raised. Author's response: The sentence has been modified as requested. Reviewer 2 comment: Table 1: This is a piece of good information. However, I ask you to provide the exact number of genotypes and the samples within each genotype for all the groups (please add additional columns as needed). Author's response: Table 1 has been modified to accommodate the required information. Reviewer 2 comment: Use of BC1S2 does not seem to be reliable in this study unless you verify the selection with phenotypic data to estimate the sensitivity and specificity of the marker assessment. However, it doesn't seem to harm the manuscript either. Author's response: The project is still ongoing. Phenotyping at different locations is currently underway. Definitely, we will utilize the phenotype data to verify the markers accuracy. Reviewer 2 comment: One important analysis I would like to suggest to add to this study is HYBRID VERIFICATION. Please prepare a data table for each sample - identified within each genotype (F1), such as order the column as F1 cross name / no, sample #, Marker gen _P1, Marker gen_P2, Observed F1 gen, True Hyb (Yes or no), if not if the F1 gen is observed as maternal or paternal type, etc. Then please assess the true to hybrid types or % hybridity within each genotype (using samples within cross) and across all samples. Then also revise your results section with a detailed discussion on how this assay is helpful to discriminate true-to-type hybrids and also describe potential bias caused by the assay itself - genotyping error or so using data on samples within each genotype.  Author's response: The genotyping analysis for the hybrid verification experiment is presented in Figure 5 and Table 5, under the Result section. The result section has also been furnished with a detailed discussion on using the KASP assay for hybrid verification and the potential drawback of the technology, as shown below: Hybrid verification. In another QC experiment using our workflow, we screened two groups of F 1 plants for hybrid verification, including their parental inbred lines, with 10 KASP SNP markers. The parental inbred lines were screened with an initial 50 KASP SNP taken from a defined panel of maize QC KASP markers to identify polymorphic markers. Only 10 KASP markers, polymorphic between the parental lines, were used to screen the F 1 plants to verify their parentage. The KASP genotyping assay was useful in distinguishing between the parental genotypes and identifying the true hybrid lines. Cluster analysis of Group1 F 1s ( Figure 4) grouped the genotypes into three clusters. The heterozygous F 1 progenies were in the middle of the plot, and the homozygous parental inbred lines diverged from each other (along the X- and Y-axis of the plot) for all markers. The genotyping result (Table 5) and the clustering pattern indicate that the F 1 progenies were true hybrids. Similar clustering was observed among F 1s in Group 2 except in Set 3b, where 38 F 1s were grouped with parental genotypes. The homozygous F 1s could be due to contamination from foreign pollens during the crossing in the field or seed mix-up during storage or planting. Nonetheless, the KASP genotyping assay suffers some genotyping errors, especially during the automatic calling of genotypes. For instance, one F1 line (SCH-4) developed from the bi-parental cross, KS23-6 and IITATZI1653, appeared to cluster with the parent 2 (IITATZI1653) when genotyped with marker PZB01658_1 (Figure 5). The datapoint representing IITATZI1653 (Figure 5, information in the yellow square) was plotted higher up, away from the X-axis, which brought it closer to the datapoint representing SCH-4 plotted slightly away from the other F1s in the middle. Because genotype calls are generated based on the relative position of datapoints on the plot, SCH-4 was automatically called as the nearby parental genotype, A:A, which was an error seeing that line SCH-4 was heterozygous (true hybrid) for the rest of the markers. The upward positioning of line IITATZI1653 away from the X-axis could be possibly due to trace contamination of line IITATZI1653 sample DNA with line  KS23-6 sample DNA during sample preparation.  A monomorphic marker is seen in the genotyping of F1 lines developed from the bi-parental crosses KS23-3 x IITATZI1653 using marker PHM5502_31. Reviewer 2 comment: Did you sample multiple samples within each plant? If so, please revise the results section accordingly. Author's response: I am hoping that I got your question correct here. If you are referring to whether or not we sampled by bulking, the answer is no, except for the MAS experiment for selecting PVA enriched lines, where we bulked ten leaf tissues from 10 plant stands per row. Basnet Bhoja R. CIMMYT, Mexico 25 8 2022 1. Before making the final decision, I am still unsure how you controlled or separated genetic purity and genotyping error in the assay. The markers seem to predict the hybrids almost perfectly, with few exceptions. How was that possible? I am not trying to deny the fact, but being curious as it was not the case in wheat. 2.  My last question was about 'analyzing multiple samples from the same plant - without bulking.' Normal practice in QC for genetic purity and true-to-type hybrid verification is that multiple F1s samples are used (you have done it), and multiple samples within each plant are also used to control the genotyping or other handling errors that may arise during the genotyping workflow. It also gives confidence about the reproducibility of the same results for the same genetic materials. Offornedo Queen International Institute of Tropical Agriculture, Nigeria 6 9 2022 Reviewer 2 comment: Before making the final decision, I am still unsure how you controlled or separated genetic purity and genotyping error in the assay. The markers seem to predict the hybrids almost perfectly, with few exceptions. How was that possible? I am not trying to deny the fact, but being curious as it was not the case in wheat. Author's response: We are unsure what the case is when genotyping wheat using the KASP assay. However, the prediction of the maize hybrids could have been aided by the fact that maize is diploid, unlike wheat which is polyploid. The polyploid nature of wheat could make it challenging to distinguish between heterozygous and homozygous hybrid lines using KASP assays. Also, our genotyping is complemented by careful crossing and sample collection in the field and meticulous handling of the samples. The KASP assay genotyping prep is handled by select individuals and carried out in a dimmed light PCR workstation to minimize contamination and avoid activating the light-sensitive fluorophores in the KASP Mastermix. Furthermore, we first selected polymorphic markers that distinguished the parents, and only those were used to identify the hybrid lines. As such, any given F1 line is either a hybrid or not. Given the robustness and accuracy of the KASP assay, the use of multiple markers, and careful handling of the sample analysis, the chance of error is minimal. Reviewer 2 comment:  My last question was about 'analyzing multiple samples from the same plant - without bulking. 'Normal practice in QC for genetic purity and true-to-type hybrid verification is that multiple F1s samples are used (you have done it), and multiple samples within each plant are also used to control the genotyping or other handling errors that may arise during the genotyping workflow. It also gives confidence about the reproducibility of the same results for the same genetic materials." Author's Response: Thank you for clarifying your question. For this experiment, 12 different F1 plants per cross were sampled, although the number of samples per plant can vary depending on the breeder's request based on the number of seeds required for the subsequent experiment. We, however, did not do a duplicate analysis of the F1 plants. We acknowledge the importance of having technical replicates in an experiment; however, the accuracy of the KASP assay is well established (as cited in our manuscript). The specificity of the KASP assay means that even one validated marker can accurately distinguish between parents and offspring, and using up to 10 markers reduces the chance of genotyping error significantly. Therefore, using technical replicates may not be worth the cost; instead, we could increase the number of markers in case of any doubt or possible errors. 10.21956/gatesopenres.14950.r32149 Reviewer response for version 2 Platten John Damien 1Referee 1 International Rice Research Institute, Los Baños, Philippines 20 6 2022 Copyright: © 2022 Platten JD 2022 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendationapprove-with-reservations I thank the authors for taking the time to address the issues raised in the previous review. I do think there are still some outstanding issues that would substantially improve the case for the work presented: Articulation of what use cases are in mind needs to be better. Currently, the articulation is basically "low-throughput applications that may require fast turnaround time". This is true to a certain extent, but not the strongest case; it could easily be argued that any "low-throughput" request could either pay the extra up-front fees, as this will be cheaper than maintaining a lab just for this purpose, or samples aggregated with other larger jobs. In the end, this is essentially saying that because certain breeders were not organised in their workflow, we need to maintain an entire lab for them. I still don't see any costing of the procedures. This is sorely needed, even if this costing only includes consumables and not salaries of dedicated staff. It should be compared with service providers for jobs of the same # samples and # SNPs. Is the rationale for developing the new method (or application) clearly explained? Partly Is the description of the method technically sound? Partly Are the conclusions about the method and its performance adequately supported by the findings presented in the article? No If any results are presented, are all the source data underlying the results available to ensure full reproducibility? No source data required Are sufficient details provided to allow replication of the method development and its use by others? Partly Reviewer Expertise: Marker design, marker evaluation, marker QC metrics, marker-assisted introgression, molecular breeding I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above. Offornedo Queen International Institute of Tropical Agriculture, Nigeria 28 7 2022 The authors thank the reviewers for taking the time to review the manuscript and raise critical issues. We trust that addressing these issues will immensely improve the paper. Below are the responses to the concerns raised. Reviewer 1 comment: I thank the authors for taking the time to address the issues raised in the previous review. I do think there are still some outstanding issues that would substantially improve the case for the work presented: Articulation of what use cases are in mind needs to be better. Currently, the articulation is basically "low-throughput applications that may require fast turn-around time". This is true to a certain extent, but not the strongest case; it could easily be argued that any "low-throughput" request could either pay the extra up-front fees, as this will be cheaper than maintaining a lab just for this purpose, or samples aggregated with other larger jobs. In the end, this is essentially saying that because certain breeders were not organized in their workflow, we need to maintain an entire lab for them. Authors' response: Besides the fact that the logistics for sending a low sample volume is not cost-effective, we also indicated in our previous response that there is a benefit of turn-around time when we genotype in-house. We had also indicated that neither the lab nor the equipment is procured solely for this genotyping purpose. We reiterate that the lab is a shared facility serving multiple activities, as are the equipment; We have re-purposed the real-time PCR for KASP assay. There is no maintenance cost for this workflow. The only thing dedicated to the workflow is the Klustercaller software. We still outsource samples for the routine forward breeding application. However, there are applications where a quick genotyping of a small number of samples has to be done, for QC purposes, for breeders and partner seed companies. In summation, this study is focused on the scientific rigour rather than the business proposition of the technological workflow. Reviewer 1 comment: I still don't see any costing of the procedures. This is sorely needed, even if this costing only includes consumables and not salaries of dedicated staff. It should be compared with service providers for jobs of the same # samples and # SNPs. Authors' response: A statement detailing the cost comparison of the procedure has been included in the Introduction section (P3L6): "The standard cost for routine KASP genotyping by Intertek is $2.6 per sample per 10 SNPs, excluding shipping costs, compared to our in-house genotyping at $2.95." 10.21956/gatesopenres.14581.r31823 Reviewer response for version 1 Platten John Damien 1Referee 1 International Rice Research Institute, Los Baños, Philippines 18 3 2022 Copyright: © 2022 Platten JD 2022 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendationreject Is the rationale for developing the new method (or application) clearly explained? The value proposition for developing the ‘method’ is not well articulated. The authors mention several times the advantages of an in-house genotyping platform, and the general thrust of the paper is describing successful proof-of-concept application of some standard components of a SNP genotyping protocol.  However it is not especially clear if the authors are aiming to establish this service as a cost-effective alternative to outsourcing, to meet a specific need that outsourcing does not meet, or something else. Alongside this, it is not clear what the novelty of the new method is. The entire workflow represents an implementation of standard technologies (CTAB extraction, DNA quantification, KASP genotyping). None of these are new techniques, nor is their combination into a genotyping workflow. Is the description of the method technically sound? As with point number 1, the overall description is technically sound, but several key details are overlooked. The machinery used in the critical step of plate scanning (actual data acquisition) is described, but key parameters are missing (please substitute equivalent parameters depending on the model of machine): What settings are used for lamp energy? What filters are used for excitation and emission spectra? This should include part numbers, and technical details of their performance. How is the analysis (clustering) done? How are results aggregated and conclusions drawn? Are sufficient details provided to allow replication of the method development and its use by others? See the previous comments.  Some aspects are adequately described, but some others are sparse on critical technical details. If any results are presented, are all the source data underlying the results available to ensure full reproducibility? Largely not applicable. Are the conclusions about the method and its performance adequately supported by the findings presented in the article? This does not seem to be the case. In particular, benchmarking data on the capacity, technical performance, cost etc. are lacking. This makes it impossible to judge the merits of this in-house system compared to outsourcing options. Overall conclusion: The current manuscript shows ability to technically execute on a relatively small number of samples in a modest timeframe. However to be a substantial contribution in this space, more thought needs to be given to better articulate both the value proposition of the work, and provide some benchmarking data to back this up. For example if the overall purpose is to show the benefit of having an in-house genotyping platform as opposed to (or in addition to) outsourcing options, the following factors and results might be considered: What is the value of an in-house system? Turnaround time and flexibility are mentioned, which I agree with. However why is this particularly important, to justify the expense of setting up, maintaining and operating an in-house system? Are there logistical considerations that prevent the use of outsourced options? Is the in-house system functionally superior to outsourced options? Is there a particular part of the breeding process that does not lend itself to standard outsourced options – and if so, under what circumstances would it be advisable to use the in-house or outsourced options? See below comments on benchmarking. Full cost assessment of the in-house system, including salaries of technical staff, machine maintenance and depreciation. Some description of the staff involved (number of positions executing on various duties) would also be helpful. Also an assessment of technology life-cycles; genotyping platforms are evolving rapidly. I have seen many cases of expensive machines being purchased, only to sit idle as the technology has moved on even before they are delivered.  KASP is likely to be replaced in the next 5 years. How would the cost of staying up to date and current be factored in? Exploration of capacity. The authors mention completing 3 jobs (637 samples) in two weeks. This is plausible based on personal experience, though I have seen in-house systems with far higher throughput. However this is a far cry from handling 20,000 samples at peak operating times. This relates back to the first point. Also related to capacity, an exploration of current/anticipated peak demand for the system. Technical performance metrics: Average DNA quality. Call rate: what percentage of samples×markers (datapoints) do not amplify? Clarity: What proportion of datapoints are unscorable? Reproducibility: Amongst technical and biological replicates, what proportion of datapoints are scored incorrectly/inconsistently? Benchmarking against available outsourcing options. A comparison of parameters such as these: Turnaround time. Full cost per sample and per datapoint. Monthly capacity (samples and datapoints). In-house genotyping platforms can and do have merit and justification. However until these issues can be addressed, the manuscript in its current form offers no fundamental insights into how such a platform could add value to breeding over outsourcing options. If the authors can better explain why their hub is superior over other options, backed up with benchmarking data such as specified, this would greatly enhance its value. Is the rationale for developing the new method (or application) clearly explained? Partly Is the description of the method technically sound? Partly Are the conclusions about the method and its performance adequately supported by the findings presented in the article? No If any results are presented, are all the source data underlying the results available to ensure full reproducibility? No source data required Are sufficient details provided to allow replication of the method development and its use by others? Partly Reviewer Expertise: Marker design, marker evaluation, marker QC metrics, marker-assisted introgression, molecular breeding I confirm that I have read this submission and believe that I have an appropriate level of expertise to state that I do not consider it to be of an acceptable scientific standard, for reasons outlined above. Offornedo Queen International Institute of Tropical Agriculture, Nigeria 23 3 2022 The authors are very grateful to the reviewer who thoroughly read the manuscript and raised critical issues. We believe that addressing these issues will immensely improve the paper. Below are the responses to the issues raised. We noticed that our core message, which is complementing outsourcing in some situations, was not clearly articulated. We await additional reviewers' comments. We plan to revise the manuscript based on our response below and additional reviewers' comments. Is the rationale for developing the new method (or application) clearly explained? Reviewer's comment: The value proposition for developing the 'method' is not well articulated. The authors mention several times the advantages of an in-house genotyping platform, and the general thrust of the paper is describing successful proof-of-concept application of some standard components of a SNP genotyping protocol. However it is not especially clear if the authors are aiming to establish this service as a cost-effective alternative to outsourcing, to meet a specific need that outsourcing does not meet, or something else. Author's response:  The core message of our paper is to complement, not establish an alternative, to outsourcing (please see the third paragraph on page 3). We duly recognize the cost-effectiveness of the genotyping platform facilitated by CGIAR platforms such as HTPG/EiB. The need for developing such in-house workflow had been prompted by the following factors: The current available genotyping platforms have a minimum sample size requirement. For instance, the EiB facilitated genotyping at Intertek costs $2 / sample if the user orders genotyping of 1536 samples (a set of 16 plates; four plates are acceptable but price increases). Breeders often want to fingerprint a few dozen lines urgently for identity or parentage analysis. In such cases, sending less than the minimum number of samples is not only more priced per datapoint but entails shipping cost and a turn-around time of 2-3 weeks. Using other markers, such as SSR, is more expensive and cumbersome. The use of genotyping systems such as KASP alleviates all these issues. Logistical issues related to shipping by courier: In this part of the world, courier services are not very satisfactory and reliable, often resulting in damage to samples in transit or longer than normal delays, which may reduce the quality of perishable specimens. If a reasonably affordable system is available locally, it can circumvent such problems. The instruments used for this work are all standard instruments available in most molecular biology labs. Our workflow shows the re-purposing of these instruments for the genotyping workflow. For instance, the qPCR machine, which is mostly used for expression analysis, was adapted to KASP genotyping with the installation of appropriate software for SNP calling. Likewise, the Fluostar plate reader was used for plate-level DNA quantification in lieu of single sample analysis by Spectrophotometer. Reviewer's comment: Alongside this, it is not clear what the novelty of the new method is. The entire workflow represents an implementation of standard technologies (CTAB extraction, DNA quantification, KASP genotyping). None of these are new techniques, nor is their combination into a genotyping workflow. Author's response: This manuscript is about a workflow that combines carefully chosen and optimized best practices in lab techniques at different stages of genotyping to address pertinent problems faced by researchers in Sub-Saharan Africa (SSA). For users who want to genotype few samples quickly, some bottlenecks in the workflow has to be removed. Currently, the DNA extraction throughput has improved by isolating and quantifying DNA at a plate level (i.e., processing 96 samples simultaneously). Secondly, genotyping by other systems such as SSR is not cost-effective. Therefore, by implementing such a workflow, we could generate quality data quickly for application in the breeding pipeline. It should be noted that not many labs in developing countries are capable of using the KASP system in-house. Is the description of the method technically sound? Reviewer's comment: As with point number 1, the overall description is technically sound, but several key details are overlooked. The machinery used in the critical step of plate scanning (actual data acquisition) is described, but key parameters are missing (please substitute equivalent parameters depending on the model of machine): What settings are used for lamp energy? What filters are used for excitation and emission spectra? This should include part numbers,and technical details of their performance. How is the analysis (clustering) done? How are results aggregated and conclusions drawn? Author's response: The required information will be incorporated in the revised manuscript under the subsection "KASP genotyping and data analysis", as explained below: The description of the parameters for the LC480 II qPCR machine is outlined in the LC480 manual. To perform the KASP genotyping experiment on the LC480 II machine, we used the Endpoint Genotyping Analysis module within the LightCycler software, adjusting the parameters as outlined in the KASP genotyping protocol provided by LGC Biosearch Technologies. The Endpoint genotyping analysis module is based on the use of dual hydrolysis probes, which are designed for wild-type and mutant target DNA and are labelled with different dyes (FAM and HEX). However, when using a non-qPCR machine (such as the GeneAmp PCR System 9700) for amplification, a third colour probe (ROX) normalizes the fluorescence measurement. The LightCycler software within the LC480 II machine determines the sample genotypes automatically by measuring the intensity distribution of the two probes after a PCR amplification step. The relative dye intensities are then visualized in a scatter (cluster) plot that discriminates them as wild-type, heterozygous mutant, or homozygous mutant samples. The LightCycler software automatically groups similar samples and assigns genotypes based on the intensity distribution of the two dyes. The KASP amplification conditions included one cycle of KASP unique Taq activation at 94°C for 15 min, followed by 36 cycles of denaturation at 94°C for 20 s, and annealing and elongation at 60°C (dropping 0.6°C per cycle) for 1 min. Endpoint detection of the fluorescence signal was acquired for 1 min at 30°C when using the LightCycler 480 II real time-PCR System or read using the FLUOstar Omega Microplate reader (BMG Labtech, SA) when using the GeneAmp PCR System 9700. For fluorescence detection, the filter combination for the Excitation and Emission wavelength of both dyes was set at 465 – 533 (FAM) and 523 – 568 (HEX), respectively, when using LC480 II, and 485 - 520 (FAM), 544 - 590 (HEX) and 584 - 620 (ROX) when using FLUOstar Omega Microplate reader. The genotype calls were exported from the LightCycler software as fluorescent intensities of each sample in ".txt" file format and imported for analysis in the KlusterCaller analysis software (LGC Biosearch Technologies). The KlusterCaller software adjusted the cluster plot axes to enable the proper calling of genotypes. The genotype calls were grouped as homozygous for allele X (allele reported by FAM, X-axis), homozygous for allele Y (allele reported by HEX, Y-axis), heterozygous (alleles reported by FAM and HEX, between X- and Y-axis), or uncallable. The result from the KlusterCaller was exported in two file formats (".csv" and ".txt"). The ".csv" file was imported into the SNPviewer2 version 4.0.0 software (LGC Biosearch Technologies), where the cluster plot image was viewed and downloaded for publication. The genotype calls in the ".txt" file were used to calculate the genetic distance using the PowerMaker  3.25 statistical software. Are sufficient details provided to allow replication of the method development and its use by others? Reviewer's comment: See the previous comments. Some aspects are adequately described, but some others are sparse on critical technical details. If any results are presented, are all the source data underlying the results available to ensure full reproducibility? Reviewer's comment: Largely not applicable. Are the conclusions about the method and its performance adequately supported by the findings presented in the article? Reviewer's comment: This does not seem to be the case. In particular, benchmarking data on the capacity, technical performance, cost etc. are lacking. This makes it impossible to judge the merits of this in-house system compared to outsourcing options. Author's response: As mentioned above, the aim of the publication is not to replace outsourcing and we do not envisage competition with outsourcing. We are fully aware of the cost-effectiveness of the highly automated/robotic genotyping services accessible to us. We frequently use these services. The rationale for the workflow is to process fewer samples quickly. The time saved is invaluable. Loss of samples during shipments is a setback, which cannot be monetized easily. Overall conclusion: Reviewer's comment: The current manuscript shows ability to technically execute on a relatively small number of samples in a modest timeframe. However to be a substantial contribution in this space, more thought needs to be given to better articulate both the value proposition of the work, and provide some benchmarking data to back this up. For example if the overall purpose is to show the benefit of having an in-house genotyping platform as opposed to (or in addition to) outsourcing options, the following factors and results might be considered: Author's response: We do not intend to establish a rival genotyping service to replace outsourcing. As explained above, this workflow is what we are using for a while now for processing a smaller number of samples (smaller than the minimum sample required by service vendors). This gives the flexibility to assay multiple crops with multiple markers in a single plate in a matter of hours. We are a small group dedicated to maize genomics. All we want is to share our methods with partners in the same situation. As can be seen from the stats of the preprint, our manuscript is already making an impact with 20 downloads and 150 views. Reviewer's comment: What is the value of an in-house system? Turn-around time and flexibility are mentioned, which I agree with. However why is this particularly important, to justify the expense of setting up, maintaining and operating an in-house system? Are there logistical considerations that prevent the use of outsourced options? Is the in-house system functionally superior to outsourced options? Is there a particular part of the breeding process that does not lend itself to standard outsourced options – and if so, under what circumstances would it be advisable to use the in-house or outsourced options? See below comments on benchmarking. Author's response: Addressed above and below. Reviewer's comment: Full cost assessment of the in-house system, including salaries of technical staff, machine maintenance and depreciation. Some description of the staff involved (number of positions executing on various duties) would also be helpful. Author's response: Not applicable. Reviewer's comment: Also an assessment of technology life-cycles; genotyping platforms are evolving rapidly. I have seen many cases of expensive machines being purchased, only to sit idle as the technology has moved on even before they are delivered. KASP is likely to be replaced in the next 5 years. How would the cost of staying up to date and current be factored in? Author's response: We agree that the genotyping platforms are evolving rapidly. To make it clear, we have not purchased instruments solely for this technique. We have only re-purposed the existing machines. Both the qPCR machine and the plate reader have high demands for other uses. If we cease to use these machines for the KASP system, the normal utilization of the machines will continue.  Reviewer's comment: Exploration of capacity. The authors mention completing 3 jobs (637 samples) in two weeks. This is plausible based on personal experience, though I have seen in-house systems with far higher throughput. However this is a far cry from handling 20,000 samples at peak operating times. This relates back to the first point. Author's response: Our response here is related to the above explanation. When we have a large volume of samples, we use low-density and mid-density genotyping service providers. Reviewer's comment: Also related to capacity, an exploration of current/anticipated peak demand for the system. Technical performance metrics: Average DNA quality. Call rate: what percentage of samples × markers (datapoints) do not amplify? Clarity: What proportion of datapoints are unscorable? Reproducibility: Amongst technical and biological replicates, what proportion of datapoints are scored incorrectly/inconsistently? Benchmarking against available outsourcing options. A comparison of parameters such as these: Turn-around time. Full cost per sample and per datapoint. Monthly capacity (samples and datapoints). In-house genotyping platforms can and do have merit and justification. However until these issues can be addressed, the manuscript in its current form offers no fundamental insights into how such a platform could add value to breeding over outsourcing options. If the authors can better explain why their hub is superior over other options, backed up with benchmarking data such as specified, this would greatly enhance its value. Author's response: As mentioned above, the aim of the publication is not to replace or compete with outsourcing, rather complement it, particularly in cases of a small volume of samples that are not cost-effective for outsourcing. Is the rationale for developing the new method (or application) clearly explained? Reviewer's comment: Partly Is the description of the method technically sound? Reviewer's comment: Partly Are sufficient details provided to allow replication of the method development and its use by others? Reviewer's comment: Partly If any results are presented, are all the source data underlying the results available to ensure full reproducibility? Reviewer's comment: No source data required Are the conclusions about the method and its performance adequately supported by the findings presented in the article? Reviewer's comment: No Competing Interests: No competing interests were disclosed. Competing interests: No competing interests were disclosed. Competing interests: No competing interests were disclosed. Competing interests: No competing interests were disclosed. Competing interests: No competing interests were disclosed. Competing interests: No competing interests. Competing interests: No competing interests were disclosed. Competing interests: No competing interest Competing interests: No competing interests were disclosed. Competing interests: No competing interests. Competing interests: No competing interests were disclosed. Competing interests: No competing interests. ==== Refs 1 Delannay X McLaren G Ribaut JM : Fostering molecular breeding in developing countries. Mol Breeding. 2012;29 (4 ):857–873. 10.1007/s11032-011-9611-9 2 Ray DK Mueller ND West PC : Yield Trends Are Insufficient to Double Global Crop Production by 2050. PLoS One. 2013;8 (6 ):e66428. 10.1371/journal.pone.0066428 23840465 3 Xu Y Li P Zou C : Enhancing genetic gain in the era of molecular breeding. J Exp Bot. 2017;68 (11 ):2641–2666. 10.1093/jxb/erx135 28830098 4 Barabaschi D Tondelli A Desiderio F : Next generation breeding. Plant Sci. 2016;242 :3–13. 10.1016/j.plantsci.2015.07.010 26566820 5 Varshney RK Singh VK Hickey JM : Analytical and decision support tools for genomics-assisted breeding. Trends Plant Sci. 2016;21 (4 ):354–363. 10.1016/j.tplants.2015.10.018 26651919 6 Gedil M Tripathi L Ghislain M : Biotechnology Success Stories by the Consultative Group on International Agriculture Research (CGIAR) System. Science Policy Reports. 2014;95–114. 10.1007/978-3-319-04001-1_6 7 Ribaut JM de Vicente MC Delannay X : Molecular breeding in developing countries: challenges and perspectives. Curr Opin Plant Biol. 2010;13 (2 ):213–218. 10.1016/j.pbi.2009.12.011 20106715 8 Gedil M Menkir A : An Integrated Molecular and Conventional Breeding Scheme for Enhancing Genetic Gain in Maize in Africa. Front Plant Sci. 2019;10 :1430. 10.3389/fpls.2019.01430 31781144 9 Integrated Breeding Platform.Accessed on June 30 2020. Reference Source 10 Rasheed A Hao Y Xia X : Crop breeding chips and genotyping platforms: progress, challenges, and perspectives. Mol Plant. 2017;10 (8 ):1047–1064. 10.1016/j.molp.2017.06.008 28669791 11 Semagn K Olsen M Babu R : Single nucleotide polymorphism genotyping using Kompetitive Allele SpeciFic PCR (KASP): Overview of the Technology and its Application in Crop Improvement. Mol Breed. 2014;33 :1–14. 10.1007/s11032-013-9917-x 12 Semagn K Beyene Y Makumbi D : Quality control genotyping for assessment of genetic identity and purity in diverse tropical maize inbred lines. Theor Appl Genet. 2012;125 (7 ):1487–1501. 10.1007/s00122-012-1928-1 22801872 13 Beissinger TM Hirsch CN Sekhon RS : Marker Density and Read Depth for Genotyping Populations Using Genotyping-by-Sequencing. Genetics. 2013;193 (4 ):1073–1081. 10.1534/genetics.112.147710 23410831 14 Jain SA de Jesus FT de Marchioro GM : Extraction of DNA from honey and its amplification by PCR for botanical identification. Food Sci Tech (Campinas). 2013;33 (4 ):753–756. 10.1590/s0101-20612013000400022 15 Ertiro BT Ogugo V Worku M : Comparison of Kompetitive Allele Specific PCR (KASP) and genotyping by sequencing (GBS) for quality control analysis in maize. BMC Genomics. 2015;16 (1 ):908. 10.1186/s12864-015-2180-2 26545737 16 Giancola S McKhann HI Berard A : Utilization of the three high-throughput SNP genotyping methods, the GOOD assay, Amplifluor and TaqMan, in diploid and polyploid plants. Theor Appl Genet. 2006;112 (6 ):1115–24. 10.1007/s00122-006-0213-6 16453133 17 Helguera M Khan IA Kolmer J : PCR Assays for the Cluster of Rust Resistance Genes and Their Use to Develop Isogenic Hard Red Spring Wheat Lines. Crop Science. 2003;43 (5 ):1839. 10.2135/cropsci2003.1839 18 Stevanato P Broccanello C Biscarin F : High-throughput RAD-SNP genotyping for characterisation of sugar beet genotypes. Plant Mol Biol Rep. 2014;32 :691 –6. 10.1007/s11105-013-0685-x 19 Thomson MJ : High-Throughput SNP Genotyping to Accelerate Crop Improvement. Plant Breeding and Biotechnology. 2014;2 (3 ):195–212. 10.9787/pbb.2014.2.3.195 20 Wu J Liu S Wang Q : Rapid identification of an adult plant stripe rust resistance gene in hexaploid wheat by high-throughput SNP array genotyping of pooled extremes. Theor Appl Genet. 2018;131 (1 ):43–58. 10.1007/s00122-017-2984-3 28965125 21 Myakishev MV Khripin Y Hu S : High-Throughput SNP Genotyping by Allele-Specific PCR with Universal Energy-Transfer-Labeled Primers. Genome Res. 2001;11 (1 ):163–169. 10.1101/gr.157901 11156625 22 Broccanello C Chiodi C Funk A : Comparison of three PCR-based assays for SNP genotyping in plants. Plant Methods. 2018;14 :28. 10.1186/s13007-018-0295-6 29610576 23 He C Holme J Anthony J : SNP Genotyping: The KASP Assay. Crop Breeding. 2014;114 :75–86. 10.1007/978-1-4939-0446-4_7 24 Ayalew H Tsang PW Chu C : Comparison of TaqMan, KASP and rhAmp SNP genotyping platforms in hexaploid wheat. PLoS One. 2019;14 (5 ):e0217222. 10.1371/journal.pone.0217222 31116793 25 Chen J Zavala C Ortega N : The Development of Quality Control Genotyping Approaches: A Case Study Using Elite Maize Lines. PLoS One. 2016;11 (6 ):e0157236. 10.1371/journal.pone.0157236 27280295 26 Kaur B Mavi GS Gill MS : Utilisation of KASP technology for wheat improvement. Cereal Res Commun. 2020;48 :409–421. 10.1007/s42976-020-00057-6 27 Steele KA Quinton-Tulloch MJ Amgai RB : Accelerating public sector rice breeding with high-density KASP markers derived from whole genome sequencing of indica rice. Mol Breed. 2018;38 (4 ):38. 10.1007/s11032-018-0777-2 29563850 28 Shi Z Liu S Noe J : SNP identification and marker assay development for high-throughput selection of soybean cyst nematode resistance. BMC Genomics. 2015;16 (1 ):314. 10.1186/s12864-015-1531-3 25903750 29 Neelam K Brown-Guedira G Huang L : Development and validation of a breeder friendly KASPar marker for wheat leaf rust resistance locus Lr21. Mol Breeding. 2013;31 (1 ):233–237. 10.1007/s11032-012-9773-0 30 Khera P Upadhyaya HD Pandey MK : Single Nucleotide Polymorphism–based Genetic Diversity in the Reference Set of Peanut ( Arachis spp.) by Developing and Applying Cost-Effective Kompetitive Allele Specific Polymerase Chain Reaction Genotyping Assays. The Plant Genome. 2013;6 (3 ). 10.3835/plantgenome2013.06.0019 31 Pariasca-Tanaka J Lorieux M He C : Development of an SNP genotyping panel for detecting polymorphisms in Oryza glaberrima/O. sativa interspecific crosses. Euphytica. 2014;201 (1 ):67–78. 10.1007/s10681-014-1183-4 32 Diversity Array Technology Pty Ltd (DArT): Plant DNA extraction protocol for DArT. Accessed on June 24 2020. Reference Source 33 Gowda M Worku M Nair SK : Quality Assurance/ Quality Control (QA/QC) in Maize Breeding and Seed Production: Theory and Practice. CIMMYT: Nairobi.2017. Reference Source 34 Liu K Muse SV : PowerMarker: an integrated analysis environment for genetic marker analysis. Bioinformatics. 2005;21 (9 ): 2128–2129. 10.1093/bioinformatics/bti282 15705655 35 Chase MW Hills HH : Silica gel: an ideal material for Field preservation of leaf samples for DNA studies. Taxon. 1991;40 :215–220. Reference Source 36 Storchova H Hrdlickova R Chrtek J Jr : An improved method of DNA isolation from plants collected in the Field and conserved in saturated NaCl/CTAB solution. Taxon. 2000;49 (1 ):79–84. 10.2307/1223934 37 Flournoy LE Adams RP Pandy RN : Interim and archival preservation of plant specimens in alcohols for DNA studies. Biotechniques. 1996;20 (4 ):657–660. 8800686 38 Semagn K : Leaf Tissue Sampling and DNA Extraction Protocols. In: Besse P, Editor. Molecular Plant Taxonomy: Methods and Protocols. New York: Human Press;2014;1115 :53–67. 10.1007/978-1-62703-767-9_3 39 Till BJ Jankowicz-Cieslak J Huynh OA : Sample Collection and Storage. Low-Cost Methods for Molecular Characterization of Mutant Plants. 2015;9–11. 10.1007/978-3-319-16259-1_3 40 Gu WK Weeden NF Yu J : Large-scale, cost-effective screening of PCR products in marker-assisted selection applications. Theor Appl Genet. 1995;91 (3 ):465–470. 10.1007/BF00222974 24169836 41 Anderson CB Franzmayr BK Hong SW : Protocol: a versatile, inexpensive, high-throughput plant genomic DNA extraction method suitable for genotyping-by-sequencing. Plant Methods. 2018;14 (1 ):75. 10.1186/s13007-018-0336-1 30181764 42 Whitlock R Hipperson H Mannarelli M : A high-throughput protocol for extracting high-purity genomic DNA from plants and animals. Mol Ecol Resour. 2008;8 (4 ):736–41. 10.1111/j.1755-0998.2007.02074.x 21585881 43 Jim FH Novok T Garson JA : Differential susceptibility of PCR reactions to inhibitors: an important and unrecognised phenomenon. BMC Res Notes. 2008;1 (1 ):70. 10.1186/1756-0500-1-70 18755023 44 Lahiri DK Schnabel B : DNA isolation by a rapid method from human blood samples: Effects of MgCl 2, EDTA, storage time, and temperature on DNA yield and quality. Biochem Genet. 1993;31 (7–8 ):321–328. 10.1007/bf02401826 8274138 45 LGC Biosearch Technologies. Accessed on June 24, 2020. Reference Source 46 BMG LABTECH Floustar Omega Application Notes. Accessed on June 2 2020. 47 Kalpana R Stephen MC Milan M : Quantifying DNA concentrations using fluorometry: A comparison of fluorophores. Mol Vis. 2002;8 :416–21. 12432341 48 BMG LABTECH Floustar Omega Operating Manual. Accessed on June 2 2020. Reference Source 49 Kumpatla SP Buyyarapu R Abdurakhmonov IY : Genomics-assisted plant breeding in the 21st Century: Technological Advances and Progress.In I. Y. Abdurakhmonov (Ed.), Plant Breeding. IntechOpen. 2012. 10.5772/37458 50 Patterson EL Fleming MB Kessler KC : A KASP Genotyping Method to Identify Northern Watermilfoil, Eurasian Watermilfoil, and Their Interspecific Hybrids. Front Plant Sci. 2017;8 :752. 10.3389/fpls.2017.00752 28533795 51 Osei MK Danquah E Danquah A : Hybridity testing of tomato F 1 progenies derived from parents with varying fruit quality and shelf life using single nucleotide polymorphism (SNPs). Scientific African. 2020;8 :e00267. 10.1016/j.sciaf.2020.e00267 52 Majeed U Darwish E Rehman SU : Kompetitive Allele-Specific PCR (KASP): A Singleplex Genotyping Platform and Its Application. Journal of Agricultural Science. 2018;11 (1 ):11. 10.5539/jas.v11n1p11 53 Offornedo Q Menkir A Babalola D : SNP data for "Developing and deploying an efficient genotyping workflow for accelerating maize improvement in developing countries". figshare. Dataset. 2021. 10.6084/m9.figshare.17157914.v2

PMC010xxxxxx/PMC10314698.txt

==== Front J Immunother Cancer J Immunother Cancer jitc jitc Journal for Immunotherapy of Cancer 2051-1426 BMJ Publishing Group BMA House, Tavistock Square, London, WC1H 9JR 37399355 jitc-2023-006699 10.1136/jitc-2023-006699 Immune Cell Therapies and Immune Cell Engineering 1506 2436 Original researchLimited efficacy of APRIL CAR in patients with multiple myeloma indicate challenges in the use of natural ligands for CAR T-cell therapy http://orcid.org/0000-0002-6092-8949 Lee Lydia 1 Lim Wen Chean 2 Galas-Filipowicz Daria 1 Fung Kent 1 Taylor Julia 2 Patel Dominic 3 Akbar Zulaikha 2 Alvarez Mediavilla Elena 1 Wawrzyniecka Patrycja 4 Shome Debarati 4 http://orcid.org/0000-0001-7471-7715 Reijmers Rogier M 4 Gregg Trillian 4 Wood Leigh 5 Day William 2 Cerec Virginie 2 Ferrari Mathieu 2 Thomas Simon 2 Cordoba Shaun 2 Onuoha Shimobi 2 Khokhar Nushmia 2 Peddareddigari Vijay 2 Al-Hajj Muhammad 2 Cavet Jim 6 Zweegman Sonja 7 Rodriguez-Justo Manuel 3 Yong Kwee 1 Pule Martin 12 Popat Rakesh 5 1 Research Department of Haematology, UCL Cancer Institute, London, UK 2 Autolus Ltd, London, UK 3 Department of Pathology, UCL Cancer Institute, London, UK 4 Lumicks, Amsterdam, The Netherlands 5 Department of Haematology, University College London Hospitals NHS Foundation Trust, London, UK 6 The Christie NHS Foundation Trust, Manchester, UK 7 Vrije Univ Amsterdam, Amsterdam, The Netherlands Correspondence to Dr Martin Pule; m.pule@ucl.ac.uk 2023 30 6 2023 11 6 e00669926 5 2023 © Author(s) (or their employer(s)) 2023. Re-use permitted under CC BY. Published by BMJ. 2023 https://creativecommons.org/licenses/by/4.0/ This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See https://creativecommons.org/licenses/by/4.0/. Background We used a proliferating ligand (APRIL) to construct a ligand-based third generation chimeric antigen receptor (CAR) able to target two myeloma antigens, B-cell maturation antigen (BCMA) and transmembrane activator and CAML interactor. Methods The APRIL CAR was evaluated in a Phase 1 clinical trial (NCT03287804, AUTO2) in patients with relapsed, refractory multiple myeloma. Eleven patients received 13 doses, the first 15×106 CARs, and subsequent patients received 75,225,600 and 900×106 CARs in a 3+3 escalation design. Results The APRIL CAR was well tolerated. Five (45.5%) patients developed Grade 1 cytokine release syndrome and there was no neurotoxicity. However, responses were only observed in 45.5% patients (1×very good partial response, 3×partial response, 1×minimal response). Exploring the mechanistic basis for poor responses, we then compared the APRIL CAR to two other BCMA CARs in a series of in vitro assays, observing reduced interleukin-2 secretion and lack of sustained tumor control by APRIL CAR regardless of transduction method or co-stimulatory domain. There was also impaired interferon signaling of APRIL CAR and no evidence of autoactivation. Thus focusing on APRIL itself, we confirmed similar affinity to BCMA and protein stability in comparison to BCMA CAR binders but reduced binding by cell-expressed APRIL to soluble BCMA and reduced avidity to tumor cells. This indicated either suboptimal folding or stability of membrane-bound APRIL attenuating CAR activation. Conclusions The APRIL CAR was well tolerated, but the clinical responses observed in AUTO2 were disappointing. Subsequently, when comparing the APRIL CAR to other BCMA CARs, we observed in vitro functional deficiencies due to reduced target binding by cell-expressed ligand. clinical trials as topic immunotherapy immunotherapy, adoptive T-lymphocytes Autolus Therapeutics n/a http://dx.doi.org/10.13039/501100000265 Medical Research Council MR/S001883/1 special-featureunlocked ==== Body pmcWHAT IS ALREADY KNOWN ON THIS TOPIC Dual targeting chimeric antigen receptor (CAR) constructs may address the challenges of low target tumor expression and the possibility of antigen negative escape in multiple myeloma. The APRIL CAR is a ligand-based, dual targeting CAR able to target two myeloma cell antigens, B-cell maturation antigen (BCMA) and transmembrane activator and CAML interactor. WHAT THIS STUDY ADDS Clinical responses from this Phase 1 trial in relapsed refractory multiple myeloma were disappointing prompting a series of reverse translation experiments of the APRIL CAR in direct comparison to other BCMA CAR constructs. In this unique exploration for the reason underpinning suboptimal clinical responses, we find many similarities in the in vitro activity of the APRIL CAR in direct comparison to two other BCMA CARs except for reduced interleukin-2 secretion and lack of sustained tumor control. We ultimately attribute poor efficacy to the APRIL binder itself which binds target poorly when expressed on the surface of T-cells. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY On a practical level, we establish simple in vitro experimentation which may be informative as regards clinical performance. Further the investigation of the APRIL CAR continues preclinically and in clinical trials and there will only be greater exploration in the field of dual targeting CAR constructs. This manuscript focuses exclusively on the APRIL CAR and throws caution to the pursuit of APRIL, and possibly other ligand-based CAR constructs in the future development of effective, dual-targeting CARs that will benefit patients. Introduction Despite advances, multiple myeloma (MM) remains an incurable and common cancer characterized by sequential relapses requiring retreatment. Patients inevitably develop resistance to multiple therapies at which point their prognosis is poor. Notably, CD19 chimeric antigen receptor (CAR) T-cell therapies can achieve durable complete responses (CR) in a proportion of patients with relapsed, refractory B-cell malignancies.1 2 Although development of CAR T-cell therapy in MM is still early compared with that in B-cell malignancies, several studies show high remission rates and durable responses.3 CAR T-cell therapy in MM have mainly targeted the B-cell maturation antigen (BCMA), a member of the tumor necrosis factor receptor superfamily (TNFRS), which is selectively expressed on mature B cells and plasma cells (PC) as well as tumor cells from the majority of patients with MM.4 When CAR T-cells against BCMA were first considered a decade ago, some limitations were anticipated: first, BCMA expression on the surface of myeloma cells is significantly less than the number of CD19 molecules expressed on the surface of acute lymphoblastic leukemia blasts.1 5 This might lead to incomplete signaling and limited expansion and persistence. Further, target downregulation is frequently observed in patients treated with CD19 and CD22 targeting therapies6 7; the possibility of BCMA modulation was also anticipated and reported in the earliest use of a BCMA CAR in patients.8 Hence to increase the level of targetable tumor antigen and address the potential for antigen negative tumor escape, we developed the APRIL CAR for the treatment of MM. In comparison to conventional CAR constructs which typically employ antibody-based binders specific for tumor antigens, the APRIL CAR was based on a proliferating ligand (APRIL)—the natural ligand for BCMA. APRIL also recognizes transmembrane activator and CAML interactor (TACI), also a member of the TNFRS and expressed on B cells and PC.5 Given that both BCMA and TACI were recognized, the total targetable antigen density detected by APRIL CAR was increased and targeting two antigens should also reduce antigen escape. In a preclinical study, we observed the cytotoxicity of APRIL CAR T-cells against cell lines expressing physiological levels of BCMA and TACI as well as primary tumor cells, maintained target kill in the presence of soluble BCMA, TACI or APRIL, and finally, rapid clearance of tumor in an in vivo myeloma model.5 On the basis of these data, we designed a Phase 1 clinical trial of APRIL CAR T-cells in patients with refractory MM (NCT03287804, AUTO2). We observed low toxicity, low engraftment but response rates were low. In contrast, contemporaneous studies targeting BCMA with standard CAR designs showed high response rates, although with modest CAR persistence. Notably, BCMA loss has been shown to be an infrequent occurrence.9 10 Subsequently, we went from the bedside back to the bench and compared the in vitro characteristics of the APRIL CAR with those of other BCMA CARs with established clinical efficacy to identify key differences which might explain poor performance in patients. In this paper, we describe results of the AUTO2 clinical study and this subsequent exploration. Materials and methods Participants and study design This open-label, dose escalation Phase 1 study was conducted in University College London Hospital, London, UK; The Christie NHS Foundation Trust, Manchester, UK; Amsterdam UMC, Cancer Center Amsterdam, Amsterdam, The Netherlands and The Freeman Hospital, Newcastle Hospitals NHS Foundation Trust, UK. Eligibility criteria included an age of 18 years or older; an Eastern Cooperative Oncology Group performance-status score of 0 or 1; measurable disease, defined by a concentration of monoclonal protein in serum of at least 5 g/L or in urine of at least 200 mg/24 hours, serum-free light chains (involved free light chain concentration of ≥100 mg/L with abnormal ratio); at least three previous lines of therapy, including a proteasome inhibitor (PI), an immunomodulatory drug (IMID), an alkylator or CD38 monoclonal antibody (MoAB) or disease refractory to both PIs and IMID; peripheral lymphocyte count of >0.5×109/L, creatinine clearance (Cr/Cl) >30 mL/min as well as adequate hepatic and cardiac function. Patients with central nervous system (CNS) disease, prior allogeneic stem cell transplant, were excluded. Tumor expression of BCMA and TACI was not an exclusion factor. Patients were administered lymphodepletion with fludarabine (30 mg/m2/day) and cyclophosphamide (300 mg/m2/day) on days –6, –5, and –4, followed by an infusion of APRIL CAR on day 0. Dose escalation initially followed an accelerated dose titration design, in which a single patient was dosed at 15×106 CAR T-cells, followed by further CAR doses (75,225,600 and 900×106 CAR T-cells) in a 3+3 escalation design. After completion of the 24-month follow-up period or following AUTO2 treatment and early withdrawal, all patients are followed until death or for up to 15 years from treatment administration. The study was conducted in accordance with the Declaration of Helsinki and International Conference on Harmonization guidelines for Good Clinical Practice and all applicable national and local laws and regulations for clinical research at each center. Written informed consent was obtained from each patient. End points and assessments Primary outcome measures were incidence of adverse events graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events V.4.03, abnormal laboratory test results, and dose-limiting toxicities as defined below. Secondary outcome measures include disease-specific response criteria (according to the International Myeloma Working Group Uniform Response Criteria for Multiple Myeloma from the day of CAR infusion) and measurement of tumor BCMA expression by immunohistochemistry (IHC) performed as previously described.1 The exploratory end point progression-free survival (PFS) was defined as the time from CAR infusion to the date of either the first observation of progressive disease or death of any cause. Retroviral vector and CAR T-cell manufacture APRIL CAR has been previously described.5 Briefly, APRIL CAR was constructed by fusing APRIL, with a deleted proteoglycan binding domain, to the human IgG1 hinge and then to the CD28 transmembrane domain and the endodomains of CD28, OX40 and CD3-ζ. APRIL CAR was coexpressed in the γ-retroviral vector SFG5 with the sort-suicide gene RQR811 using a foot-and-mouth like 2A peptide sequence from thosea asigna virus.12 RQR8/APRIL CAR encoding γ-retroviral vector was generated by transfecting 293 T-cells with the SFG plasmid, and plasmids encoding the RD114 envelope and MoMLV gagpol. Supernatant was purified by anion exchange chromatography and filtration. CAR T-cells were generated from autologous peripheral blood mononuclear cells harvested by leukapheresis. Leukapheresate was stimulated with transact (Miltenyi), transduced with the γ-retroviral vector on retronectin (Takara), then transferred to the Miltenyi prodigy and expanded in interleukin (IL)-7 and IL-15 for 7–10 days following which cell product was cryopreserved in dimethyl sulfoxide(DMSO). Transduction efficiency was determined by fluorescence activated cell sorting (FACS) staining of T-cells for RQR8 marker expression. FACS analysis Bone marrow mononuclear cells (BM MNCs) were isolated by Ficoll Paque. Vials of stored BM MNCs or manufactured CAR products were defrosted and stained with antibodies as specified in supplementary data before analysis with a Fortessa (BD) and FlowJo (V.10.6). Statistical analyses Unless otherwise stated, data are expressed as mean±SE, and analyses were performed in GraphPad Prism, V.9 as specified in the body of this manuscript. P value<0.05 was deemed statistically significant. Data supporting the findings of this study are available on request from the corresponding author. Further trial details and methods available in online supplemental data. 10.1136/jitc-2023-006699.supp1 Supplementary data Results Study participant and disease characteristics We tested autologous APRIL CAR T-cells in a Phase 1 dose-escalation study of relapsed refractory MM. Twelve subjects were enrolled (table 1, online supplemental table 1). Successful harvest and manufacture of target dose was achieved for all patients but one patient withdrew prior to treatment due to disease progression. 10.1136/jitc-2023-006699.supp2 Supplementary data Table 1 Summary of AUTO2 patient demographics Total treated 11 Sex  Male 8 (72.7%)  Female 3 (27.4%)  Age (median/range) 61(45–69) Isotype  IgG 9 (81.8%)  LC 2 (18.2%) ISS at presentation  I 6 (54.5%)  II 1 (9.1%)  III 3 (27.3%)  Unknown 1 (9.1%) Cytogenetics  High risk 4 (36.4%)  Standard risk 2 (18.2%)  Unknown 5 (45.5%)  Years since diagnosis (median/range) 6 (1–11)  Extramedullary disease 3 (27.3%) Previous therapy  Lines (medium/range) 5 (3–6)  Previous ASCT 6 (54.5%)  Anti CD38 exposed 6 (54.5%)  Progressed on last line 5 (45.5%)  Refractory to PI or IMID 11 (100%)  Refractory to anti CD38 5 (45.5%)  Double refractory (PI and IMID) 9 (81.8%)  Triple refractory (PI, IMID, CD38) 3 (27.3%) Refractory, progressed on or within 60 days of receiving these agents. High risk cytogenetics defined as t(4;14), t(14;16), t(14;20), del(17p), 1q gain, 1p loss. ASCT, autologous stem cell transplant; IMID, immunomodulatory imide; ISS, International Staging System; LC, Light Chain; PI, proteasome inhibitor. Of the 11 patients treated, the median age was 61 (range 41–69). Three (27.3%) of patients had International Staging System (ISS) stage III disease at diagnosis and one (11.1%) at screening, four (36.4%) had high risk cytogenetics (defined as t(4;14), t(4;16), t(4;20), del(17p) (≥50% of total nucleated cells), 1q gain, 1p loss) and three (27.3%) had extramedullary (EM) disease. Patients had received a median of 5 prior therapy lines (range, 3–6). All patients had received a PI and IMID to which nine (81.8%) were double refractory. Over half (54.5%) had received daratumumab, three patients (27.3%) were refractory to PI, IMID and daratumumab and two (18.2%) were penta-refractory (bortezomib, carfilzomib, lenalidomide, pomalidomide and a CD38 MoAB). None of the patients in the AUTO2 cohort had received a BCMA targeting therapeutic agent prior to enrollment. Six (54.5%) of the patients had received a previous autologous stem cell transplant (table 1, online supplemental table 1). Four patients received bridging therapy between leukapheresis and CAR T-cell infusion (online supplemental table 2) all bridged patients had stable or progressive disease between initial screening and start of lymphodepletion and still had measurable disease. Baseline renal function ranged from a Cr/Cl of 55–133 mL/min (median 85). Table 2 Summary of adverse events Any grade Grade 3 Grade 4 Any adverse event 11 (100%) 11 (100%) 11 (100%) Hematology  Neutropenia 11 (100%) 0 11 (100%)  Anemia 8 (72.7%) 8 (72.7%) 0  Thrombocytopenia 4 (36.4%) 0 3 (27.3%)  Lymphopenia 1 (9.1%) 0 1 (9.1%) Gastrointestinal  Dysgeusia 2 (18.2%) 0 0  Mucositis 3 (27.3%) 0 0  Nausea 7 (63.6%) 0 0  Vomiting 2 (18.2%) 1 (9.1%) 0  Diarrhea 5 (45.5%) 0 0  Constipation 5 (45.5%) 0 0  Abnormal liver function tests 2 (18.2%) 0 0 Respiratory  Dyspnea 5 (45.5%) 1 (9.1%) 0  Cough 3 (27.3%) 0 0 Cardiovascular  Hypotension 1 (9.1%) 0 0  Peripheral edema 4 (36.4%) 0 0  MI 1 (9.1%) 1 (9.1%) 0 Skin  Rash 2 (18.2%) 0 0  Pruritus 1 (9.1%) 0 0 Neurology  Dizziness 1 (9.1%) 0 0  Parasthesia 2 (18.2%) 0 0  Headache 5 (45.5%) 1 (9.1%) 0 Infections (any) 9 (81.8%) 5 (45.5%) 0 Other  Fatigue 9 (81.8%) 0 0  Fevers 6 (54.5%) 2 (18.2%) 0  Chills 4 (36.4%) 0 0  Body or joint pain 7 (63.6%) 2 (18.2%) 0  Low calcium 1 (9.1%) 1 (9.1%) 0  Low phosphate 1 (9.1%) 1 (9.1%) 0 CRS 5 (45.5%) 0 0  Macrophage activation syndrome 1 (9.1%) 0 0 ICANS 0 0 0 CRS, cytokine release syndrome; ICANS, immune effector cell associated neurotoxicity syndrome. Baseline tumor expression of BCMA and TACI were assessed by FACS3 and IHC (online supplemental table 3). Treated patients had a median surface expression level of BCMA and TACI of 596 (430–780) and 381 (0–1819) antigens bounds per cell (ABC), respectively, from BM tumor cells at study entry and antigen expression was maintained following CAR T-cell infusion as measured from 1 month and at disease progression (online supplemental table 3). Toxicity and serum cytokines Shown are adverse events not designated as symptoms of cytokine release syndrome (CRS) that occurred in the first 60 days of a dose of APRIL CAR T-cells. CRS was graded according to the criteria in Lee et al. 11 Five (45.5%) patients developed CRS, all of which were mild (Grade 111). Three patients developed CRS within the first 2 weeks (Patients 003, 011 and 012 on days 11, 9 and 0, respectively) and one patient developed CRS late (Patient 001, day 26). A further patient developed CRS early with fevers on day of infusion and subsequently macrophage activation syndrome at day 29 which responded to tociluzimab. Three patients received a single dose of tociluzimab and steroids were not administered post CAR T-cells. There were no instances of immune effector cell associated neurotoxicity syndrome (ICANS) (table 2). A rise in interferon-γ (IFN-γ) was only detected in the four patients receiving the highest doses of CAR T-cells (Patients 009, 010, 011, 012) to a maximum of 2984 pg/mL (online supplemental figure 1). IL-6 rose to 413 pg/mL (median 64 pg/mL). All patients experienced Grade 4 neutropenia and 72% Grade 3 anemia. Duration of cytopenias could be prolonged. Grade 3 or 4 events were observed beyond 30 days in the following numbers of patients: anemia 5; thrombocytopenia 3; neutropenia 7. And beyond 90 days in the following: thrombocytopenia 2; neutropenia 3. Patients were supported with transfusions and granulocyte colony stimulating factor at their physicians discretion. However, excluding hematological toxicity, eight patients experienced Grade 3 or higher toxicity of any cause of which five patients experienced Grade 3 infections and there were no AUTO2 related deaths (online supplemental table 4). One patient receiving the highest dose of CAR T-cells experienced a myocardial infarction on day of CAR infusion that was thought possibly attributed to APRIL CAR infusion due to the temporal nature of the adverse event in relation to treatment. Thus, we did observe cytopenias and infections associated with lymphodepletion and, overall, APRIL CAR was well tolerated with a low incidence of CRS and no reported cases of ICANs. Product characterization and CAR T-cell persistence The median transduction efficiency (TE) was 23% (range 5.81–40.2%, online supplemental table 2) and the CD4:CD8 ratio in the T-cell product varied (median 2.8, range 0.3–12.9, online supplemental figure 2A) with a predominance of CD4. The majority of CD4 cells were effector memory (EM) (median 74.8%, range 41.3–90.0%, online supplemental figure 2B) while the CD8 CAR T-cells had a smaller proportion of EM cells (median 50.1, range 16.4–76%) and more terminally differentiated EM cells re-expressing CD45RA (TEMRA) cells (median 27.1%, range 3.4–61%). There was a low proportion of central memory (CM) CD4 and CD8 APRIL CAR T-cells (median 11.9 and 9.4%, respectively). Circulating APRIL CAR was detectable following 12/13 doses administered (figure 1A) within the first 10 days (median 5, range 1–10) and peaked early (median 12 days, range 8–24 days) with a median Tmax of 8673.7 copies/µg DNA (range 26–96,598 copies/µg DNA). APRIL CAR T-cell expansion or tracking to tumor niches was not associated with dose (figure 1A, online supplemental figure 3A–C). APRIL CAR was typically undetectable by 30 days, had a median persistence of 21 days and detectable up to 3 months in one patient at the highest dose. In this small cohort, APRIL CAR expansion did not correlate with disease response with no clinical responses observed with APRIL CAR expansion (Patients 001, 011 and 012) and vice versa (Patient 005). Figure 1 CAR expansion and best response. (A) CAR T-cell expansion as assessed by PCR of peripheral blood. (B) Best response to APRIL CAR infusion (as of February 2022) according to dose (15×106 to 900×106) of chimeric antigen receptor–positive (CAR+) T-cells. Two patients were retreated with a higher dose of 225×106 cells at time points indicated with a red star. All responses were confirmed and assessed according to the International Myeloma Working Group Uniform Response Criteria for Multiple Myeloma. MR, minimal response; PR, partial response; SD, stable disease; PD, progressive disease; VGPR, very good partial response. At 1-month post treatment, APRIL CAR T-cells were detected by flow cytometry in BM in five patients (online supplemental figure 3B–D). Compared with therapeutic product, there was a preferential expansion of CD8 CAR T-cells, increase in mature memory phenotypes as well as increased expression of immunomodulatory proteins (online supplemental figure 3F–G). There was an increased expression of Programmed cell death protein 1 (PD-1) (p<0.05 for CD4 and CD8 CAR T-cells by paired t-test) and a trend for increased T-cell immunoglobulin and mucin-domain containing-3 (TIM3) expressing CD4 and CD8 T-cells (p=0.06, p=0.07, respectively, by paired t-test). Further, APRIL CAR T-cells expressed more TIM3 and PD-1 and less Ki67 compared with non-CAR T-cells from patient BM samples (online supplemental figure 3H). Disease response The objective response rate was low and short lived (figure 1B). Five (45.5%) patients responded. One patient achieved a very good partial response (VGPR) which included regression of EM disease, 3 a partial response (PR) and 1 a minimal response. Of these patients, median PFS was 5 months (range 2–8). As of February 1, 2022, median time to progression was 3 months following APRIL CAR, 2 patients were still alive and the median overall survival for the 11 patients was 375 days following the first APRIL CAR T-cell dose. Patient 001 has remained in stable disease 55 months after receiving CAR T-cells. Retreatment with APRIL CAR Expecting a higher therapeutic dose, two patients who received the lowest doses of CAR T-cells were retreated with a higher dose of APRIL CAR (225×106). Patient 1 had a significant CAR T-cell expansion following a dose of 15×106 cells (nearly 1×105 copies/µg genomic DNA (gDNA)) but stable disease. Following the second infusion 11 months later, there was significantly reduced CAR T-cell expansion (to 4×102 copies/µg gDNA) without disease response. Patient 5 had progressive disease at 1 month following the initial dose of 75×106 cells and received four doses of weekly daratumumab from month 2. There was a PR following a second CAR T-cell dose at month 3 and low level CAR expansion (10–100 copies/µg) following both CAR doses. We previously described activity of APRIL CAR T-cells against TACI and BCMA expressing targets in vitro and in vivo including in response to low-density target cells. Despite this, the APRIL CAR demonstrated inferior performance to other BCMA CARs.13 14 We undertook a series of reverse translation experiments to understand the reasons for this. APRIL CAR T-cells are deficient in IL-2 secretion and activity on repeated challenge We compared the function of APRIL CAR to what has become standard BCMA targeting CARs. The sequences for the standard single chain variable fragment (scFv) used in bb212115–17 and two VHH domains in LCAR-B38M18 19 were derived from patents, and the latter joined by a (G4S)3 linker. Both BCMA binders were then cloned into second generation (41bbζ) backbones and henceforth referred to as bb2121 and LCAR-B38M CARs, respectively (figure 2A). Figure 2 Functional assessment of APRIL CAR variants, bb2121 and LCAR-B38M CAR in vitro. CAR transduced peripheral blood mononuclear cells from normal donors (n=3) were co-cultured with non-transduced SUPT1 targets (SUPT1 NT) or targets engineered to express low levels of BCMA (estimated 636 molecules per cell SUPT1 BCMA) at an effector to target ratio (E:T) of 1:4. (A) Diagram summarizing various CAR constructs assessed functionally. (B) Target kill as a percentage of targets in media alone. (C) IFN-γ and IL-2 release as assessed by ELISA of culture supernatant at 24 hours. (D) After 4 days of co-culture, CAR T-cells co-cultured alone, with SUPT1 NT or SUPT1 BCMAlo targets were enumerated. (E) Next CARs were also co-cultured with MM1s cells at an E:T ratio of 1:4 and further live MM1s cells added to wells of culture plate twice a week from 4 days after initial co-culture set-up (=ReStim 1). Assessment of viable cells in culture plate occurred 3 days after number of restimulations indicated and percentage tumor of total live cells after sequential stimulations displayed on graph. Data shown is from single experiment, with each data point representing mean of duplicate or triplicate. In B–D, effector alone, SUPT1 NT and SUPT1 BCMA are represented by gray circles, black squares and orange triangles, respectively. Statistical tests by one-way analysis of variance with focus on performance of AUTO2 and bb2121 or LCAR-B38M. ***p<0.001. BMCA, B-cell maturation antigen; CAR, chimeric antigen receptor; IFN, interferon; IL, interleukin; LD, lentivirus; NT, non-transduced; RD, retrovirus. The APRIL CAR varies from most other BCMA CARs not only by its ligand-based binder, but also by having a third generation (CD28-OX40-CD3ζ) endodomain, and use of γ-retroviral vector for manufacture. Hence, to also understand if these differences impacted on relative performance, along with AUTO2 (the therapeutic: γ-retroviral transduction, third generation CD28-OX40-CD3ζ endodomain), we also compared APRIL CAR in second generation formats (CD28ζ, 41BBζ endodomains) and manufactured in lentiviral vectors. CAR T-cells were cultured alone, with unmodified SUPT1 or SUPT1 targets expressing low levels of BCMA (636 ABC). We observed equivalent target kill, IFN-γ and cell proliferation at 5 days on co-culture with antigen expressing target of the γ-retroviral, third generation APRIL CAR compared with BCMA CARs. These parameters were not improved by APRIL CAR transduction method or co-stimulatory endodomain (figure 2B–D). However, one notable difference of the APRIL CAR was the low levels of IL-2 release of all APRIL CAR formats compared with bb2121 and LCAR-B38M on co-culture with BCMA expressing targets (figure 2C). IL-2 secretion by APRIL constructs remained comparatively deficient on co-culture with targets expressing increased levels of BCMA (online supplemental figure 4) or on altering the extracellular linker of the APRIL CAR construct (online supplemental figure 5). IL-2 secretion on co-culture with BCMA expressing targets could be increased by increasing TE and therefore increasing CAR expression on T-cells (online supplemental figure 6). Second, despite equivalent target specific kill on co-culture, and in contrast to bb2121 and LCAR-B38M, notably, all APRIL CARs failed to control tumor growth on repeated in vitro stimulation with MM1s cells (figure 2E). Figure 5 Characteristics of AUTO2, bb2121 and LCAR-B38M BCMA binders. (A) Binding kinetics of BCMA binders as assessed by surface plasmon resonance. IgG2a Fc conjugated binders were immobilized on a CM5 biacore chip before binding assessment with soluble BCMA. (i) Summary table. Binding for (ii) WT APRIL, (iii) C11D5.3, (iv) VHH binders used in LCAR-B38M (Nanjing_GSI5022_Dual_VHH). (B) Protein stability of BCMA binders as determined by nano differential scanning fluorimetry (nanoDSF) (i) Thermal unfolding curves, (ii) graph plotting onset of unfolding as well as melting temperatures of binder used in AUTO2 (APRIL-HNG), bb2121(C11D5.3) and LCAR-B38M (two VHH in series/Nanjing GSI5022 dual, and each VHH in isolation/VHH1 or VHH2). (C) Binding of CAR expressing cells to soluble ligand were assessed by transducing peripheral blood mononuclear cells from three healthy donors with bicistronic constructs coexpressing RQR8 marker gene (in format RQR8_2A_CAR) before incubation with (i) BCMA Fc and secondarily stained with APC conjugated anti Fc antibody (eg, FACS plots from a single donor shown). (ii) Alternatively, CARs were incubated with different concentration with APC conjugated BCMA. Controlling for set MFI of RQR8 expression, graph showing APC MFI of transduced T-cells. (D) Cell avidity between CAR transduced T-cells from four healthy donors (mean of minimum of two replicates shown) and myeloma cell line H929 assayed by acoustic force microfluidic microscopy. (i) Avidity curves represent mean±SEM from separate donors. (ii) Cell binding avidity from h at 1000 pN. Individual donors represented with different shapes. Multiple paired t-tests and Holm-Sidak correction. *p<0.05, **p<0.01, ***p<0.001. APC, Allophycocyanin; BMCA, B-cell maturation antigen; CAR, chimeric antigen receptor; FACS, fluorescence activated cell corting; Fc, fragment crystallizable; MFI, mean fluorescence intensity; NT, non-transduced. These observations of reduced IL-2 secretion on co-culture and failure of prolonged tumor control were seen in all APRIL CAR formats regardless of co-stimulation domains or γ-retroviral versus lentiviral transduction. Reduced cytokine signaling on stimulation of APRIL CAR by RNA sequencing We then looked for differences in T-cell activation by interrogating the transcriptomic profiles obtained by bulk RNA sequencing of non-transduced (NT) T-cells from three donors, and T-cells transduced with APRIL CAR, bb2121 or LCAR-B38M following in vitro activation by plate bound BCMAFc for 24 hours. There was evidence of antigen-dependent T-cell activation in all three BCMA targeting CARs compared with NT T-cell with increased T-cell receptor (TCR) signaling, cytokine signaling and cell cycling (online supplemental figure 7). Comparing the three BCMA targeting CAR constructs, we did observe a small number of transcripts significantly expressed at a lower frequency in APRIL CAR compared with bb2121 and LCAR-B38M (n=86 with >log2 fold change in expression and adjusted p<0.05, figure 3A, B). Interestingly, at 24 hours there was no difference in IRF4, which is thought to correlate to strength of TCR activation,20 or the cytokines IFN-γ or IL-2. However, there was evidence of reduced cytokine signaling. Many of the genes significantly upregulated in bb2121/LCAR-B38M are involved in type 1 (eg, IRF7, MX1, OAS genes) and type 2 IFN (eg, CXCL10) signaling. It is not possible to assess if cytokine signaling was isolated to CARs but this data presumably indicates greater response to cytokines by all T-cells in response to cytokine release by antigen-activated CARs. Figure 3 RNA sequencing of ligand activated CAR T cells from different donors. (A) Volcano plot of differentially expressed genes in APRIL versus bb2121 and LCAR-B38M (to the left, genes upregulated in bb2121/LCAR-B38M). Genes highlighted have greater than twofold difference in expression and adjusted p<0.05 by DESeq2. (B) Heatmap of selected differentially expressed genes (adjusted p<0.05 by DESeq2). CAR, chimeric antigen receptor. APRIL CAR does not cause autoactivation APRIL naturally trimerizes21 and it has been shown previously that membrane-bound APRIL are present in oligomeric forms.22 We sought to look for evidence of autoactivation of the APRIL CAR but found no evidence for target independent CAR activation compared with bb2121 and LCAR-B38M. There was no significant increase in cytokine release or phenotypic markers of activation (CD35, CD27, CD127, 41BB, PD-1, TIM3, Lymphocyte Activation Gene 3 or LAG3) in CAR expressing T-cells cultured in the absence of target antigen for 7 days (figure 4). Furthermore, autoactivation was also not seen in the APRIL CAR in varying formats (summarized in figure 2A) varying in co-stimulation domains or γ-retroviral versus lentiviral transduction (online supplemental figure 8). Figure 4 Assessment of autoactivation of AUTO2, bb2121 and LCAR-B38M. (A) Cytokine release from non-transduced (NT) PBMNCs or PBMNCs transduced with CAR constructs (retrovirus/RD transduced APRIL CAR, or LD/lentivirus transduced bb2121 or LCAR-B38M, four donors) and then co-cultured alone. IFN-γ and IL-2 from supernatant was quantified by ELISA. (B) Cells were then phenotyped by FACS at baseline (gray filled histograms) and after 7 days following initial activation and transduction with CAR constructs (AUTO2 in red, bb2121 in blue and LCAR-B38M in orange). Isotype control depicted in hatched black line. (C) Graph showing MFI of labeled proteins relative to isotype control. For each marker there was no significant difference between protein expression between AUTO2 and bb2121 or LCAR-B38M by multiple paired t-tests and Holm-Sidak correction. CAR, chimeric antigen receptor; FACS, Fluorescence activated cell sorting; GZMB, Granzyme B; IFN, interferon; IL, interleukin; LAG3: Lymphocyte Activation Gene 3; MFI: Mean Fluorescence Intensity; PBMNCs, peripheral blood mononuclear cells; PD-1, Programmed cell death protein 1; TIM3, T-cell immunoglobulin and mucin-domain containing-3. Binding characteristics of APRIL is similar to other BCMA binders but there is deficient target binding by CAR expressing T-cells We then looked for differences in tumor binding by APRIL. By surface plasmon resonance, we noted similar binding affinities of APRIL compared with BCMA binders used in bb2121 and LCAR-B38M (figure 5A). We then assessed protein stability of BCMA binders used in the APRIL, bb2121 (C11D5.3), LCAR-B38M (VHH binders in series or alone) CARs and found thermal unfolding consistently above 50°C indicative of stable proteins (figure 5B). We hypothesized that despite the stability of recombinant APRIL, APRIL CAR may be less stable or suboptimally presented when expressed on the membrane surface. T-cells were transduced with constructs coexpressing RQR8 with bb2121, LCAR-B38M and APRIL CAR. T-cells were then stained with increasing concentrations of labeled soluble recombinant BCMA and analyzed by flow cytometry without removing unbound BCMA so that cell fluorescence would assess relative availability of surface CAR, unconfounded by difference in binding kinetics. APRIL CAR consistently bound less soluble ligand when incubated with various concentrations of fluorophore labeled BCMA compared with bb2121 and LCAR-B38M (figure 5C, online supplemental figure 9). Appreciating the limitations of these platforms to assess the complex interaction between cells, we next sought to quantify the collective interactions of multiple receptor/ligand complexes and co-receptors which make up the immunological synapse.23 Using a platform of acoustic force technology to quantify avidity between effector and target cells (z-Movi), by demonstrated reduced avidity between APRIL CAR expressing cells and the human myeloma cell line H929 (figure 5D) compared with bb2121 and LCAR-B38M. Related to this, there was a trend for reduced phosphorylation of ZAP70 and LAT by phosphoflow supporting reduced TCR activation of the APRIL CAR in comparison to LCAR-B38M and bb2121 (online supplemental figure 10). Thus despite the similar affinity of the APRIL protein to BCMA and the stability of the protein, we demonstrate deficient target binding by cell expressed APRIL as the cause of deficient CAR activation. Discussion In 2017, we described a CAR based on the ligand APRIL for treatment of MM. Our motivation was to overcome what we anticipated would be limitations of scFv-based BCMA CARs: namely low antigen density of BCMA resulting in suboptimal signaling, and tumor escape through loss of BCMA. Rather than using an antibody-based binder, we used the natural ligand APRIL. Since APRIL CAR could recognize both BCMA and the related PC lineage antigen TACI, the total targetable antigen density was increased, and co-targeting two antigens should prevent single antigen negative escape. Initial functional tests of the APRIL CAR demonstrated antigen directed activation in vitro and in vivo, kill of low antigen expressing targets and efficacy against primary myeloma cells.5 We evaluated autologous APRIL CAR T-cells in patients with relapsed or refractory MM in a Phase 1, dose-escalation study (AUTO2). Eleven patients were treated, with median 5 prior lines therapy, 27% were of ISS III and 36% had high-risk cytogenetics. While the APRIL CAR was well tolerated, responses were only observed in 45.5% of patients and to a best response of a VGPR. In contrast to these results, two Phase 2 studies have become the reference points for BCMA CARs testing ide-cel and cilta-cel.13 14 A total of over 200 patients were treated in these studies and were more heavily pretreated with a median of 6 prior lines. While the frequency of severe CRS and ICANS was also low, the overall response rate, rate of CR and PFS was 73%, 33% and 8.8 months with ide-cel14 and 98%, 67% and not reached at 27.7 months with cilta-cel.13 24 Suboptimal APRIL CAR activity was also evidenced by low cytokine release and poor CAR expansion in patients. In AUTO2, increases in IFN-γ were only observed in four patients (range in trial of 0–2984 pg/mL) and peak IL-6 reached a median of 63 pg/mL (0–413) in the AUTO2 cohort. In comparison, in an early study with cilta-cel, increases in IL-6 were seen in most patients (13/17), reaching nearly 1×105 pg/mL.25 An early study by Brudno et al describing a BCMA CAR with a CD28ζ endodomain described median fold change of >100 baseline of both IFN-γ and IL-6.26 Further, circulating APRIL CAR reached a median Tmax of 8673.7 copies/µg DNA (range 26–96,598) and was detectable for a median of 21 days (max 3 months). In comparison, CARs peaked and were detectable for a median of over 1×105 copies/µg of DNA (max 5×106) and a median of 6 months with ide-cel27 and a mean of over 1×104 copies/µg (max 2×106) for a median 4 months with cilta-cel,13 25 respectively. APRIL CAR T-cells were manufactured using unmanipulated autologous cryopreserved leukapheresis as starting material, transduction with γ-retroviral vector and expansion in IL-7/IL-15 for 7–10 days. In comparison, manufacture of ide-cel27 and cilta-cel25 use lentiviral transduction and do not include IL-7/IL-15 which is thought to optimize memory phenotype in the manufactured CAR product.28 Looking for a cause for our poor clinical responses, we observed a median TE of 23% (range 5.81–40.2%) and a low proportion of naive and CM phenotypes (median combined percentage 18% of CARs). However, low TEs were also described with an early study of cilta-cel (median 22%),25 without impacting efficacy. From CD19 CAR T-cell studies, the proportion of naïve memory phenotypes has been shown to correlate with longer CAR persistence and improved patient outcomes.29 In their trial of BCMA-targeting CD28ζ CAR, Cohen et al describe between 20% and 30% of naive and stem cell memory T-cell populations9 and this figure was under 10% with ide-cel.30 Looking for a further reason for the low clinical responses, we also noted that BCMA expression was not a criteria for trial entry in the AUTO2 trial in common to other BCMA CAR trials.13 24 Thus the high proportion of mature memory phenotypes, low TEs observed and no requirement for tumor BCMA expression for trial entry were not unique to AUTO2 and could not sufficiently explain the low response rates in the AUTO2 trial. We previously described antigen specific cytotoxicity in vitro by the APRIL CAR of targets expressing physiological levels of antigen and primary cells as well as antigen dependent IFN-γ release. We also demonstrated efficient in vivo clearance of tumor in a xenograft model. Given the effectiveness of the APRIL CAR in this murine model, we sought to explain the low rate of clinical responses observed in AUTO2 using a series of in vitro experiments, both making direct comparison to other BCMA CARs, bb2121 and LCAR-B38M, and also seeking more experimental readouts than attempted before. In this study, we demonstrate similar kill and IFN-γ secretion compared with bb2121 and LCAR-B38M. However, we also observed little or no IL-2 release by the APRIL CAR in response to low-density antigen expressing targets and reduced capacity for serial killing on repeated stimulation. These assays are likely to reflect the efficiency of CAR activation by target. Deficiencies in IL-2 secretion have been described following suboptimal T-cell activation by reduced recruitment of downstream proteins involved in TCR signaling such as LAT31 and Lck.32 Suboptimal T-cell signaling is associated with T-cell anergy33 which may in part explain reduced capacity for persistent tumor control. Additionally, deficiencies in cytokine signaling was a prominent finding on transcriptomic analysis of APRIL CAR compared with bb2121 and LCAR-B38M where we observed significantly reduced transcripts of genes involved in type 1 (eg, IRF7, MX1, OAS genes) and type 2 IFN (eg, CXCL10) signaling in stimulated APRIL CAR transduced T-cells. It is also noteworthy that APRIL CAR IL-2 secretion increased with TE. Thus, the low TE achieved in the AUTO2 trial may have exacerbated deficiencies in cytokine production as well as T-cell signaling. These described functional deficiencies remained despite controlling for vector and endodomain, therefore, these deficiencies were likely due to use of APRIL as an antigen recognition domain. APRIL naturally trimerizes and may form higher-order concatemers.22 However functional testing revealed no increase in tonic signaling compared with bb2121 or LCAR-B38M. Next, we compared binding kinetics and protein stability of antigen recognition domains and found APRIL had similar binding affinity to BCMA compared with the antibody binders used in bb2121 and LCAR-B38M. Protein stability of the binders for the three CAR constructs was also observed at physiological temperatures. However, APRIL CAR expressing T-cells consistently bound less soluble BCMA than bb2121 or LCAR-B38M expressing T-cells. A significant limitation of commonly used techniques for affinity readouts is that these do not accurately ascertain the complex interaction which exist between cells. In contrast, assessment of avidity accounts for the collective interactions of multiple receptor/ligand complexes and co-receptors which make up the immunological synapse between cells and can thus lead to more accurate predictions of T-cell functionality.23 We confirmed deficient target binding by APRIL CAR and reduced avidity between APRIL CAR expressing cells and the human myeloma cell line H929 compared with bb2121 and LCAR-B38M. Collectively, this data suggests that despite equivalent affinity of APRIL and BCMA at a protein level by Biacore, there is suboptimal target binding by cell expressed APRIL. This, in turn, indicates suboptimal or unstable presentation of APRIL in a CAR format, thereby resulting in poor APRIL CAR function and limited clinical responses observed in the AUTO2 trial. APRIL also binds TACI, a tumor necrosis factor receptor that has a role in the maturation of B cells, with increased expression on maturing B-cell stages compared with BCMA.34 35 TACI was initially thought to be a T-cell antigen36 but now accepted to be expressed primarily on B cells21 with a recent report suggesting expression in suppressive T-cells.37 Importantly, TACI is expressed on MM cells.5 We found BCMA and TACI to be coexpressed on tumor for the majority (78%) of patients, but at generally lower levels than BCMA (median BCMA 1061 ABC, median TACI 333 ABC). AUTO2 is the first study which directly targets TACI; with the caveat of poor expansion and persistence, we did not find any toxicity attributable to TACI targeting. Several CARs based on natural ligands have been described including some which have been tested in human subjects.38 39 Our data focuses exclusively on the APRIL CAR, however, our experience of the AUTO2 trial and subsequent comparisons to two other antibody-based BCMA CAR constructs do suggest that the physiological requirements of a natural ligand interaction may well differ from those optimal for CAR activation. Engineering strategies can be employed to influence the specificity or binding dynamics of natural ligands. For example, a single amino acid mutation has enhanced selective binding of IL-13 CARs to IL-13Rα2,38 40 we truncated the terminal amino acid of APRIL to minimize proteoglycan binding5 and more recently the enforced trimerization of APRIL has been demonstrated to improve target binding and CAR activity.22 Our data may indicate that improving the stability of the APRIL binder in a CAR format may improve function of the APRIL CAR. It is noteworthy that with the increased use of BCMA CARs in patients, BCMA loss has been shown to be less frequent than initially anticipated9 10 indicating that potential benefit from dual targeting CARs will be primarily dependent on improved tumor targeting rather than preventing antigen negative escape. Thus, despite elegant engineering solutions available to optimize ligand based CARs, ligands are frequently limited to a single and possibly suboptimal presentation on a T-cell and may not meet the complex requirements for dual antigen binders. In comparison, human scFv or single-domain antibodies are now readily available and provide a choice of multiple potential binders that are amenable to manipulation for iterative optimization to meet specific clinical requirements. In the future, CAR T discovery may thus be better served using antibody derived antigen binding domains. In summary, clinical testing of an APRIL based CAR proved disappointing compared with other studies testing antibody-based CARs. Our experience suggests that in comparison to the versatility and well understood stability of antibody derived binders, natural ligands may not be ideal antigen binding domains for CARs. Simple in vitro experimentation, may be informative as regards clinical performance. Namely, direct comparison to CAR constructs with proven clinical efficacy if available, use of targets expressing physiological levels of antigen, assessment of IL-2 release and prolonged tumor control on repeated stimulation. Further, beyond an assessment of affinity, we also indicate the value to ascertaining interaction of cellular interactions including avidity assessment. Finally, this work represents the first attempts at co-targeting of TACI, although future development may be better achieved with BCMA/TACI bicistronic CAR cassettes.41–43 10.1136/jitc-2023-006699.supp3 Supplementary data 10.1136/jitc-2023-006699.supp4 Supplementary data 10.1136/jitc-2023-006699.supp5 Supplementary data The authors also acknowledge Tobias Menne who recruited patients to the AUTO2 trial. Data availability statement Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information. Ethics statements Patient consent for publication Not applicable. Ethics approval This study involves human participants and was approved by IRAS PROJECT ID 208047, REC APPROVAL: 16/LO/2008. Twitter: @CavetJim Correction notice: This article has been corrected since it was first published to correct author name Kwee Yong. Contributors: MP and RP are supported by the UK National Institute of Health research UCLH Biomedical research center. LSHL is supported by the UK medical Research Council. Preclinical development of the APRIL CAR was supported by Bloodwise and Kay Kendell Leukemia Fund. VP, NK, KY and RP designed the clinical study. RP, JC and SZ recruited patients. VC and LW were involved in trial management. WD and MA-H were responsible for translational data. DG-F, DP, MR-J and LSHL processed and analysed clinical samples. WCL, KF, JT, ZA, EAM, MF, ST, SC and SO generated the research data. MP and LSHL wrote the manuscript. MP is guarantor for this manuscript. Funding: LSHL is funded by the MRC (MR/S001883/1), AUTO2 trial was funded and sponsored by Autolus Therapeutics Competing interests: LSHL, KY and MP are inventors on patents relevant to this paper filed by UCL and are entitled to share of royalties there from. KY and MP owns equity in Autolus Therapeutics. WCL, JT, ZA, WD, VC, MF, ST, SC, SO, NK, VP, MA-H and MP are either current present or former employees of Autolus Therapeutics. LSHL and KY receive research funding from Autolus Therapeutics. Provenance and peer review: Not commissioned; externally peer reviewed. Supplemental material: This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise. ==== Refs References 1 Maude SL , Laetsch TW , Buechner J , et al . Tisagenlecleucel in children and young adults with B-cell Lymphoblastic leukemia. N Engl J Med 2018;378 :439–48. 10.1056/NEJMoa1709866 29385370 2 Neelapu SS , Locke FL , Bartlett NL , et al . Axicabtagene Ciloleucel CAR T-cell therapy in refractory large B-cell lymphoma. N Engl J Med 2017;377 :2531–44. 10.1056/NEJMoa1707447 29226797 3 Manier S , Ingegnere T , Escure G , et al . Current state and next-generation CAR-T cells in multiple myeloma. Blood Rev 2022;54 :S0268-960X(22)00003-0. 10.1016/j.blre.2022.100929 4 Lee L , Bounds D , Paterson J , et al . Evaluation of B cell maturation antigen as a target for antibody drug conjugate mediated cytotoxicity in multiple myeloma. Br J Haematol 2016;174 :911–22. 10.1111/bjh.14145 27313079 5 Lee L , Draper B , Chaplin N , et al . An APRIL-based Chimeric antigen receptor for dual targeting of BCMA and TACI in multiple myeloma. Blood 2018;131 :746–58. 10.1182/blood-2017-05-781351 29284597 6 Neelapu SS , Rossi JM , Jacobson CA , et al . Cd19-loss with preservation of other B cell lineage features in patients with large B cell lymphoma who Relapsed post-Axi-Cel. Blood 2019;134 :203. 10.1182/blood-2019-126218 7 Fry TJ , Shah NN , Orentas RJ , et al . Cd22-targeted CAR T cells induce remission in B-ALL that is naive or resistant to Cd19-targeted CAR Immunotherapy. Nat Med 2018;24 :20–8. 10.1038/nm.4441 29155426 8 Ali SA , Shi V , Maric I , et al . T cells expressing an anti-B-cell maturation antigen Chimeric antigen receptor cause remissions of multiple myeloma. Blood 2016;128 :1688–700. 10.1182/blood-2016-04-711903 27412889 9 Cohen AD , Garfall AL , Stadtmauer EA , et al . B cell maturation antigen-specific CAR T cells are clinically active in multiple myeloma. J Clin Invest 2019;129 :2210–21. 10.1172/JCI126397 30896447 10 Samur MK , Fulciniti M , Aktas Samur A , et al . Biallelic loss of BCMA as a resistance mechanism to CAR T cell therapy in a patient with multiple myeloma. Nat Commun 2021;12 :868. 10.1038/s41467-021-21177-5 33558511 11 Lee DW , Gardner R , Porter DL , et al . Current concepts in the diagnosis and management of cytokine release syndrome. Blood 2014;124 :188–95. 10.1182/blood-2014-05-552729 24876563 12 Donnelly MLL , Hughes LE , Luke G , et al . The 'cleavage' activities of foot-and-mouth disease virus 2A site-directed Mutants and naturally occurring '2A-like' sequences. J Gen Virol 2001;82 :1027–41. 10.1099/0022-1317-82-5-1027 11297677 13 Berdeja JG , Madduri D , Usmani SZ , et al . Ciltacabtagene Autoleucel, a B-cell maturation antigen-directed Chimeric antigen receptor T-cell therapy in patients with Relapsed or refractory multiple myeloma (CARTITUDE-1): a phase 1B/2 open-label study. Lancet 2021;398 :314–24. 10.1016/S0140-6736(21)00933-8 34175021 14 Munshi NC , Anderson LD , Shah N , et al . Idecabtagene Vicleucel in Relapsed and refractory multiple myeloma. N Engl J Med 2021;384 :705–16. 10.1056/NEJMoa2024850 33626253 15 Carpenter RO , Evbuomwan MO , Pittaluga S , et al . B-cell maturation antigen is a promising target for adoptive T-cell therapy of multiple myeloma. Clin Cancer Res 2013;19 :2048–60. 10.1158/1078-0432.CCR-12-2422 23344265 16 Chekmasova AA , Horton HM , Garrett TE , et al . A novel and highly potent CAR T cell drug product for treatment of BCMA-expressing hematological Malignances. Blood 2015;126 :3094. 10.1182/blood.V126.23.3094.3094 17 Morgan R , Friedman K . Bcma Chimeric antigen receptors. Wo2016094304A2. 2016. Available: https://patents.google.com/patent/WO2016094304A2/en?oq=WO+2016%2f094304+A2+ 18 Fan X , Zhuang Q , Wang P , et al . Chimeric antigen receptors targeting Bcma and methods of use thereof. Wo2018028647A1. 2018. Available: https://patents.google.com/patent/WO2018028647A1/en?oq=WO2018028647A1 19 Ubani EZ . Bcma-targeted car-t cell therapy of multiple myeloma. 2021. Available: https://patents.google.com/patent/US20210128618A1/en?oq=US20210128618A1 20 Man K , Miasari M , Shi W , et al . The transcription factor Irf4 is essential for TCR affinity-mediated metabolic programming and Clonal expansion of T cells. Nat Immunol 2013;14 :1155–65. 10.1038/ni.2710 24056747 21 Bossen C , Schneider P . APRIL and their receptors: structure, function and signaling. Semin Immunol 2006;18 :263–75. 10.1016/j.smim.2006.04.006 16914324 22 Schmidts A , Ormhøj M , Choi BD , et al . Rational design of a Trimeric APRIL-based CAR-binding domain enables efficient targeting of multiple myeloma. Blood Adv 2019;3 :3248–60. 10.1182/bloodadvances.2019000703 31698455 23 Greenman R , Pizem Y , Haus-Cohen M , et al . Shaping functional avidity of CAR T cells: affinity, avidity, and antigen density that regulate response. Mol Cancer Ther 2021;20 :872–84. 10.1158/1535-7163.MCT-19-1109 33649106 24 Martin T , Usmani SZ , Berdeja JG , et al . Ciltacabtagene Autoleucel, an anti-B-cell maturation antigen Chimeric antigen receptor T-cell therapy, for Relapsed/refractory multiple myeloma: CARTITUDE-1 2-year follow-up. J Clin Oncol 2023;41 :1265–74. 10.1200/JCO.22.00842 35658469 25 Xu J , Chen L-J , Yang S-S , et al . Exploratory trial of a biepitopic CAR T-targeting B cell maturation antigen in relapsed/refractory multiple myeloma. Proc Natl Acad Sci USA 2019;116 :9543–51. 10.1073/pnas.1819745116 30988175 26 Brudno JN , Maric I , Hartman SD , et al . T cells genetically modified to express an anti-B-cell maturation antigen Chimeric antigen receptor cause remissions of poor-prognosis Relapsed multiple myeloma. JCO 2018;36 :2267–80. 10.1200/JCO.2018.77.8084 27 Raje N , Berdeja J , Lin Y , et al . Anti-BCMA CAR T-cell therapy Bb2121 in Relapsed or refractory multiple myeloma. N Engl J Med 2019;380 :1726–37. 10.1056/NEJMoa1817226 31042825 28 Gargett T , Brown MP . Different cytokine and stimulation conditions influence the expansion and immune phenotype of third-generation Chimeric antigen receptor T cells specific for tumor antigen Gd2. Cytotherapy 2015;17 :487–95. 10.1016/j.jcyt.2014.12.002 25573334 29 Fraietta JA , Lacey SF , Orlando EJ , et al . Determinants of response and resistance to Cd19 Chimeric antigen receptor (CAR) T cell therapy of chronic lymphocytic leukemia. Nat Med 2018;24 :563–71. 10.1038/s41591-018-0010-1 29713085 30 Alsina M , Shah N , Raje NS , et al . Updated results from the phase I CRB-402 study of anti-Bcma CAR-T cell therapy Bb21217 in patients with Relapsed and refractory multiple myeloma: correlation of expansion and duration of response with T cell phenotypes. Blood 2020;136 :25–6. 10.1182/blood-2020-140410 31 Dong R , Libby KA , Blaeschke F , et al . Rewired signaling network in T cells expressing the Chimeric antigen receptor (CAR). EMBO J 2020;39 :e104730. 10.15252/embj.2020104730 32643825 32 Tan TCJ , Knight J , Sbarrato T , et al . Suboptimal T-cell receptor signaling compromises protein translation, ribosome biogenesis, and proliferation of mouse CD8 T cells. Proc Natl Acad Sci USA 2017;114 :E6117–26. 10.1073/pnas.1700939114 28696283 33 Sadegh-Nasseri S , Dalai SK , Korb Ferris LC , et al . Suboptimal engagement of the T-cell receptor by a variety of peptide-MHC ligands triggers T-cell Anergy. Immunology 2010;129 :1–7. 10.1111/j.1365-2567.2009.03206.x 20002785 34 Ng LG , Sutherland APR , Newton R , et al . B cell-activating factor belonging to the TNF family (BAFF)-R is the principal BAFF receptor facilitating BAFF Costimulation of circulating T and B cells. J Immunol 2004;173 :807–17. 10.4049/jimmunol.173.2.807 15240667 35 Avery DT , Kalled SL , Ellyard JI , et al . BAFF selectively enhances the survival of Plasmablasts generated from human memory B cells. J Clin Invest 2003;112 :286–97. 10.1172/JCI18025 12865416 36 Wang H , Marsters SA , Baker T , et al . TACI-ligand interactions are required for T cell activation and collagen-induced arthritis in mice. Nat Immunol 2001;2 :632–7. 10.1038/89782 11429548 37 Tai Y-T , Lin L , Xing L , et al . APRIL signaling via TACI mediates immunosuppression by T regulatory cells in multiple myeloma: therapeutic implications. Leukemia 2019;33 :426–38. 10.1038/s41375-018-0242-6 30135465 38 Brown CE , Alizadeh D , Starr R , et al . Regression of glioblastoma after Chimeric antigen receptor T-cell therapy. N Engl J Med 2016;375 :2561–9. 10.1056/NEJMoa1610497 28029927 39 Branella GM , Spencer HT . Natural Receptor- and ligand-based Chimeric antigen receptors: strategies using natural ligands and receptors for targeted cell killing. Cells 2021;11 :21. 10.3390/cells11010021 35011583 40 Debinski W , Gibo DM , Obiri NI , et al . Novel anti-brain tumor Cytotoxins specific for cancer cells. Nat Biotechnol 1998;16 :449–53. 10.1038/nbt0598-449 9592393 41 Spiegel JY , Patel S , Muffly L , et al . CAR T cells with dual targeting of Cd19 and Cd22 in adult patients with recurrent or refractory B cell malignancies: a phase 1 trial. Nat Med 2021;27 :1419–31. 10.1038/s41591-021-01436-0 34312556 42 Cordoba S , Onuoha S , Thomas S , et al . CAR T cells with dual targeting of Cd19 and Cd22 in pediatric and young adult patients with Relapsed or refractory B cell acute Lymphoblastic leukemia: a phase 1 trial. Nat Med 2021;27 :1797–805. 10.1038/s41591-021-01497-1 34642489 43 Ramakrishnan A . Phase 1 Alexander study of Auto3, the first Cd19/22 dual targeting CAR T cell therapy, with Pembrolizumab in patients with Relapsed/refractory (R/R) DLBCL. 2020. Available: https://ash.confex.com/ash/2020/webprogram/Paper133960.html

PMC010xxxxxx/PMC10324118.txt

==== Front J Nematol J Nematol jofnem jofnem Journal of Nematology 0022-300X 2640-396X Sciendo 37426723 jofnem-2023-0018 10.2478/jofnem-2023-0018 Research Paper Plant-Parasitic Nematodes Associated with Cannabis sativa in Florida Desaeger J. jad@ufl.edu Coburn J. Freeman J. Brym Z. University of Florida, Department of Entomology and Nematology, Gulf Coast Research and Education Center, Wimauma, FL 33598 University of Florida, Horticultural Sciences Department, North Florida Research and Education Center, Quincy, FL 32351 University of Florida, Department of Agronomy, Tropical Research and Education Center, Homestead, FL 33031. This paper was edited by Ralf J Sommer. 2 2023 6 7 2023 55 1 2023001829 7 2022 © 2023 J. Desaeger et al., published by Sciendo 2023 J. Desaeger et al., published by Sciendo https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. Abstract The subtropical climate of Florida allows for a wide range of crops to be grown. With the classification of hemp (Cannabis sativa L., <0.3% delta-9-tetrahydrocannabinol) as an agricultural commodity, hemp has become a potential alternative crop in Florida. Hemp cultivars of different geographies (Europe, China, and North America), and uses (fiber, oil and CBD), were evaluated in three field experiments. The field experiments evaluated a total of 26 cultivars and were conducted for two consecutive seasons at three different locations (soil types) in North (sandy loam), Central (fine sand), and South Florida (gravelly loam). Nematode soil populations were measured at the end of each season. A diverse population of plant-parasitic nematodes was found, with reniform nematodes (RN, Rotylenchulus reniformis) the dominant species in North and South Florida (up to 27.5 nematodes/cc soil), and RKN (Meloidogne javanica) the main species in central Florida (up to 4.7 nematodes/cc soil). Other nematodes that were commonly found in south Florida (and to a lesser extent north Florida) were spiral (Helicotylenchus spp.), stunt (Tylenchorhynchus spp.) and ring nematodes (Criconemoids), while in central Florida, stubby root (Nanidorus minor) and sting nematodes (Belonolaimus longicaduatus) were found. No significant difference among hemp cultivars was noted at any of the locations. RKN were found in all three regions and soils, while RN were only found in North and South Florida. This is the first report on plant-parasitic nematodes associated with hemp in Florida fields. Natural nematode populations varied greatly, depending on where in Florida hemp was grown. Growers who wish to include hemp in their crop rotation need to be aware of potential pest pressure from nematodes. More research is needed to determine to what extent nematodes, especially RKN and RN, can reduce hemp growth and yield. Keywords Cannabis sativa root-knot nematodes reniform nematodes soil type ==== Body pmcThe subtropical climate of Florida allows for a wide range of crops to be grown, often as an expanded or extreme geographical range. With the removal of hemp (Cannabis sativa L., <0.3% delta-9-tetrahydrocannabinol) from the controlled substances list (2018 Farm Bill and 2019 Florida Statute, SB1020 Fla. Stat. § 581.217), hemp now has the potential to be an alternative crop in Florida. There are multiple types of hemp crops harvested for fiber, grain, and cannabidiol, or CBD (Small and Marcus, 2002; Williams, 2019). Different parts of the hemp plant are harvested for each of these specific production targets. Fiber hemp varieties are known to have tall, preferably slender stalks and can grow anywhere from 1 to 5 m in height. Clothing, rope, housing materials, compost, and paper are a few items that can be made from just the stalk of the fiber hemp plant (Cherney and Small, 2016). Fiber hemp products can be used as potentially eco-friendly alternatives or replacements for construction materials and plastics. Grain hemp cultivars are plants that produce seed that can be used for food items, animal feed, and cosmetics (Cherney and Small, 2016). These plants do not grow as tall as fiber plants, usually averaging 1 to 2 m in height. Dual-use cultivars tend to produce many seeds like grain hemp cultivars, but they also have a stalk that is suitable for use in fiber hemp production. Producers harvesting hemp grain would also harvest the remaining stems for fiber, and this is the most common example of a dual-use hemp crop. Hemp that is cultivated and bred in order to process flower parts for cannabinoid extracts are usually shorter and have a more brachiate or bushy appearance as compared to fiber and grain. CBD, or cannabidiol, is the non-psychoactive compound commonly extracted from hemp flower and closely resembles THC. CBD can be obtained from the plant by extracting the oil from the flower or by burning or vaporizing the dried flower material. CBD products include topicals, edibles, and smokable products, and is used medicinally to curb pain and muscle spasms (National Academies of Sciences, Engineering, and Medicine, 2017). Cannabinoids are highly concentrated in the trichomes of the bracts and nearby leaves of unfertilized female flowers, much lower in the root and plant tissue, and are at even lower concentrations in hemp pollen and seeds. CBG, or cannabigerol, is another non-psychoactive cannabinoid found in cannabis. Just like CBD hemp, CBG hemp is bred and cultivated for production of the unpollinated female flower. To support the future viability and sustainability of hemp and considering the importance and prevalence of plant-parasitic nematodes in Florida, it is critical to collect information on the interaction of hemp and nematodes. Plant-parasitic nematodes (PPN) are extremely prevalent in Florida, and many different species are found that can potentially cause damage to many of the crops grown. The most important PPN in Florida in terms of damage potential are root-knot nematodes (Meloidogyne spp.), sting nematodes (Belonolaimus longicaudatus) and reniform nematodes (Rotylenchulus reniformis). All these nematodes have very wide geographic and crop host ranges. Root-knot nematodes are one of the main overall limiting factors to crop production in the world (Sasser, 1980). They are extremely prevalent in Florida, because of the subtropical climate and often sandy soils, with multiple species and races, and a wide host range that includes grasses, vegetables, and fruits. They are the primary nematode problem in Florida vegetables, causing significant losses in tomato, cucurbit, and other vegetables. In Florida strawberries, M. hapla, the northern root-knot nematode, has started to become more important as well (Desaeger, 2018). Sting nematodes are also an important soil pest in strawberry, and can cause considerable damage in cantaloupe, potato, sugarcane, forage and turf grasses, citrus and corn, especially in the sandy soils (Smart and Nguyen, 1991; Crow and Brammer, 2001; Noling, 2016). Reniform nematodes also have a very wide host range, including many tropical fruits, ornamentals as well as cotton and soybean (McSorley et al., 1983; Kinlock and Sprenkel, 1994). They are especially common in the Rockdale soils of the southernmost counties in Florida, as well as in the slightly heavier soils of the northwestern Panhandle region (Wang, 2004). The trials were done during 2019 and 2020 at three different field locations in Florida, (1) the North Florida Research and Education Center (NFREC) in Quincy (North Florida); (2) the Gulf Coast Research and Education Center (GCREC) in Wimauma (West-Central Florida); and (3) the Tropical Research and Education Center (TREC) in Homestead (South Florida) (Table 1). To support the future viability of hemp and considering the importance of PPN in Florida, it is critical to collect information on PPN associated with hemp in Florida. Table 1. Experimental site description. Field site County Elevation Soil type Coordinates GCRECa – Central FL Hillsborough 31 m Myakka fine sand 27.76° N, 82.23° W TRECb – South FL Miami—Dade 1 m Shallow Krome gravelly loam 25.47° N, 80.50° W NFRECc – North FL Gadsden 63 m Orangeburg loamy fine sand 30.58° N, 84.59° W a GCREC, Gulf Coast Research and Education Center, Wimauma, FL. b TREC, Tropical Research and Education Center, Homestead, FL. c NFREC = North Florida Research and Education Center, Quincy, FL. Materials and Methods Nematode sampling and extraction methodology Nematode soil samples were taken at the end of each crop/experiment, and roots were visually examined for nematode damage (no preplant samples were collected). Nematode soil samples were taken as four cores per cultivar plot with a locally made cone sampler (3.0 cm-diam. × 25 cm-deep). The four cores were composited for each plot. Soil samples were sealed in plastic bags and were stored at 4°C and nematodes extracted within two weeks. Nematodes were extracted at the GCREC nematology lab by taking a 200 cm3 soil sample from each composite sample. A modified Baermann method using a salad spinner (Henrik Preutz, IKEA® USA) was used (Viglierchio and Schmitt, 1983). Plant-parasitic nematodes (PPN) were identified to genus or species level and by feeding group. Morphological identification was performed using a compound Zeiss AXIO Scope A1 microscope (Carl Zeiss, Göttingen, Germany). Plant-parasitic nematodes (PPN) were identified to genus, and depending on the morphological complexity of the species, the nematode was also identified into species level by using the pictorial key proposed by Mai and Mullin (1996). When root galls were observed on roots, Meloidogyne species identification was done by extracting DNA from female nematodes found inside root galls, and mitochondrial haplotype-based identification was done using primers, PCR and RFLP conditions as described by Pagan et al. (2014) and Baidoo et al. (2016). Non-plant-parasitic nematodes (NPPN) nematodes were also morphologically identified into trophic group (bacterivores, fungivores, omnivores). Location 1: Central Florida (GCREC = Gulf Coast Research and Education Center) Hemp cultivar trials at the Gulf Coast Research and Education Center farm in Wimauma, FL were conducted in October 2019 (season 1, fall) and February 2020 (season 2, spring) in the same location. The soil type in this area is classified as Myakka fine sand (95% sand, <1% organic matter). The experimental area was adjacent to an experimental hops (Humulus lupulus) field that had artificial lighting installed by poles with LED lights (GreenPower LED flowering DR/W, Philips) suspended 5.5 m above the ground, staggered 6 m apart (Agehara et al., 2020). Lights provided supplemental light (6 hrs/day) during the first two months of growth which is necessary to extend Florida’s natural day length (max. 14 hrs at GCREC) and create long day conditions required to produce meaningful vegetative growth of hops. Both hemp and hops belong to the Cannabaceae plant family, are dioecious and day-length-sensitive (DLS), with flowering being triggered by short days. The same supplemental lighting regime for hops was applied to hemp, meaning the lights started automatically illuminating an hour before sunset and then turned off at midnight for the first two months after planting. Also, the same fertigation and irrigation schedule was followed in the hemp trials as was done for hops (Desaeger et al., 2022). The experimental area measured 78 m long and 18 m wide and consisted of three rows 78 m long x 1 m wide. Rows were 6 m apart, with bahia grass in between, and were covered with landscape fabric with two lines of drip tape underneath. The drip tape used was Toro BlueLine® with 12″ emitter spacing (0.26 gph, 0.62″ internal diameter, 0.045″ wall thickness; Agehara et al., 2020). Eight selected hemp cultivars (Table 2) were planted in a randomized block design consisting of four plants of the same cultivar per plot, and each plot was replicated six times (two replicates in each row), for a total of 192 hemp plants. Hemp seedlings were grown in a growth room for two months prior to transplanting in the field. Seeds were sown into 128-cell seedling trays filled with PRO-MIX HP growth medium (Premier Horticulture Inc., Quakertown, PA) on August 16, 2019, and December 29, 2019. Seedlings were grown under supplemental lighting (16-h light and 8-h dark) to maintain vegetative growth. Irrigation was supplied as needed using overhead irrigation. Uniform seedlings of each variety were transplanted to the field on 16 October 2019, and 23 February 2020. Plants measured on average 15–20 cm at the time of transplanting. Watering and fertilizing were administered via in-bed drip tape in both areas. Watering consisted of four one-hour cycles every five hours each day and a soluble fertilizer (N-P2O5-K2O: 5-2-8) was applied with irrigation twice a week throughout the season, based on an accumulated rate of 103 kg N ha−1. Table 2. Hemp cultivars with their specific use, country of origin and year(s) planted at each location. Cultivars Type/Use Origin Year Planted GCRECa,b and TRECc Yuma-2 Fiber/Dual China 2019 Bama Fiber/Dual China 2019 Puma-3 Fiber/Dual China 2019, 2020 Tygra Fiber/Dual Poland 2019 Carmagnola Selezionata Fiber/Dual Italy 2019 Eletta Campana Fiber/Dual Italy 2019 Cherry Blossom x T1 CBD US 2019 Cherry Blossom CBD US 2019 TREC Berry Blossom CBD US 2019 Canda Fiber/Dual Canada 2019 Si-1 Fiber/Dual China 2019 Fibranova Fiber/Dual Italy 2019 Han-NE Fiber/Dual China 2019, 2020 Carmagnola Fiber/Dual Italy 2020 NBS CBD US 2020 Wife CBD US 2020 Maverick CBD US 2020 NFRECd KG 9201 CBD US 2019 KG 9202 CBD US 2019 Cherry Wine CBD US 2019, 2020 Cherry Blossom x T1 CBD US 2019 Cherry Blossom CBD US 2019, 2020 Berry Blossom CBD US 2020 Queens Dream CBD US 2020 Cinderella Story CBD US 2020 Cloud Berry CBD US 2020 Cherry Blonde CBD US 2020 Hot Blonde CBD US 2020 a GCREC, Gulf Coast Research and Education Center. b Planting at GCREC was done in fall 2019 and spring 2020, and at TREC and NREC in summer 2019 and summer 2020. c TREC, Tropical Research and Education Center. d NFREC, North Florida Research and Education Center. Both trials were terminated after three to four months, depending on the cultivar. Location 2: South Florida (TREC = Tropical Research and Education Center) Hemp cultivar trials at TREC in Homestead, FL (25.4687° N, 80.5007° W) were conducted in summer 2019 and summer 2020 on two different fields that were adjacent to each other. The soil type is a shallow Krome, a gravelly loam soil series (loamy-skeletal, carbonatic, hyperthermic, Lithic Rendoll) containing 58% sand, 19% silt, 15% clay, and 8% gravel (Li and Zhang, 2002). Its plowed surface layer, largely crushed bedrock, is 15–20 cm deep with 34% to 76% of limestone fragments (≥ 2 mm in diameter; Bryan and Lance, 1991). A total of 17 different hemp cultivars (fiber, dual, and CBD) were planted at this location: 13 in 2019 and six in 2020 (Table 2). All cultivars were direct-seeded in the field, and planting densities differed according to their usage. All fiber/dual cultivars were sown by hand in eight rows measuring 1.8 m long and were spaced 30 cm apart within a 5.58 m2 (1.83 m × 3.05 m) rectangular plot. For dual cultivars, 900 seeds were planted per plot with an intended planting density of 161 plants/m2, while 1500 seeds were planted for fiber cultivars to achieve 269 plants/m2. CBD cultivars were planted to establish a density of 11 plants/m2 by sowing 60 seeds in the experimental plot and after germination thinning to 11 plants. All plots were fertilized once before planting at 112, 56, and 300 kg/ha of N, P, and K, respectively, using a 6-3-13 slow-release granular fertilizer. Due to early indications of N deficiency, there was an additional N application in mid-July, which was top-dressed using 46-0-0 conventional urea at a rate of 56 kg/ha. Irrigation was provided as needed by overhead sprinklers. Location 3: North Florida (NFREC = North Florida Research and Education Center) Hemp cultivar trials at NFREC in Quincy, FL were conducted in summer 2019 and summer 2020. Soil type at NFREC is an Orangeburg loamy fine sand. The experimental design was a randomized complete block design with four replications. Three DLS cultivars, including Cherry Blossom (CBL), Cherry × T1 (CT1), and Cherry Wine (CW), and two day-length-neutral (DLN) cultivars, including Kayagene 9201 (KG9201) and Kayagene 9202 (KG9202) were evaluated in 2019. In 2020, cultivars evaluated were CBL, CW, Berry Blossom, Cinderella Story, Cherry Blonde, Cloud Berry, Hot Blonde, and Queens Dream (Table 2). Seeds were sown in the greenhouse using the same methodology as for Location 1, on 14 June 2019 and 2 May 2 2020. Uniform seedlings of each cultivar were transplanted to the field on 3 July 2019 and 12 June 2020. Agronomic practices were similar between 2019 and 2020. The field setup was similar as for vegetable production in the area, which is plastic-mulch raised beds that were 20-cm high and 30-cm wide with a single drip tape underneath the plastic mulch. The spacing between rows and between plants within a row was 1.8 and 1.5 m, respectively. Therefore, the plant density was ~3600 plants per hectare. Irrigation was supplied daily through the drip tape. Fertilizer (N-P2O5-K2O: 10-10-10) was applied at a rate of 112 kg ha−1 immediately prior to transplanting and disked into soils. A soluble fertilizer (N-P2O5-K2O: 4-0-8) was applied with irrigation as needed throughout the season based on an accumulated rate of 56 kg N ha−1. Results At the GCREC, by the end of the first season, stubby root nematodes (Nanidorus minor) and NPPN (bacterivorous, fungivorous, and omnivorous) were found (Table 3). Nematode populations in the soil were higher in the second season. By the end of season two, stubby root nematodes, root-knot nematodes (Meloidogyne javanica), as well as sting nematodes (Belonolaimus longicaudatus) were found. No significant difference was noted between cultivars for either root-knot, sting or stubby root nematodes (Table 3). No root galls were noted in the first season, and few small root galls were seen on two cv. Cherry Blossom plants in season 2. Table 3. Nematode soil populations (no./200 cc soil) at harvest of eight different hemp cultivars for two consecutive seasons, GCREC, Fall 2019 (trial 1) and Spring 2020 (trial 2)a. Sept. – Dec. 2019 (Fall) Bacterivore Fungivore Omnivore Stubby root Root-knot Sting Yuma-2 49 31 12 31 a 0 0 Bama 117 18 7 21 ab 0 0 Puma-3 120 32 12 32 a 0 1 Tygra 70 30 7 6 b 0 0 Carmagnola Selezionata 121 30 9 12 ab 0 1 Eletta Campana 89 19 7 11 ab 0 0 Cherry Blossom x T1 121 32 5 23 ab 0 1 Cherry Blossom 72 25 7 15 ab 0 1 P value 0.24 0.75 0.31 0.02 - - Feb.-May 2020 (Spring) Bacterivore Fungivore Omnivore Stubby root Root-knot Sting Yuma-2 - - - 18 abc 16 15 Bama - - - 62 ab 437 3 Puma-3 - - - 62 ab 610 13 Tygra - - - 11 bc 289 14 Carmagnola Selezionata - - - 11 bc 9 16 Eletta Campana - - - 8 c 117 9 Cherry Blossom x T1 - - - 65 a 100 8 Cherry Blossom - - - 24 abc 315 32 P value - - - 0.002 0.16 0.17 a Numbers listed describe the number of a type of nematode found in 200 cc of soil collected from the root area of the hemp cultivars listed. There are multiple types/uses of hemp here. P values were calculated on Log (x+1) transformed data according to Tukey’s HSD where P ≤ 0.05; bacterivores, fungivores, omnivores were not counted in the second season. Nematode soil populations in the hemp trials at TREC consisted of several PPN: reniform nematodes (Rotylenchulus reniformis), stunt nematodes (Tylenchorhynchus spp.), spiral nematodes (Helicotylenchus spp.), ring nematodes (Criconemella spp. or Mesocriconema spp.), and root-knot nematodes (Meloidogyne spp.), as well as NPPN (Table 4). The highest reniform nematode populations were found in the first trial on the fiber cultivars Puma and Carmagnola Selezionata (P < 0.004). The second trial was done on an adjacent field with mostly CBD-type hemp cultivars. In this trial, reniform, ring and root-knot nematodes were found (Table 4). No significant differences for any of the nematodes were observed between the tested cultivars in season 2, and no root galls were observed on any of the roots. Table 4. Nematode soil populations (no./200 cc soil) at harvest of different hemp cultivars for two consecutive seasons, summer 2019 and summer 2020; TREC soil survey resultsa. Summer 2019 Bacterivore Fungivore Omnivore Reniform Stunt Spiral Ring RKN Puma-3 116 42 0 5,564 a 112 176 12 0 Bama 70 74 0 1,987 ab 156 221 26 0 Berry Blossom 28 52 0 574 abc 32 92 2 0 Tygra 22 26 2 762 abc 8 26 14 0 Carmagnola Selezionata 52 100 8 4,872 a 80 84 12 0 Helena 68 48 2 648 abc 24 48 0 0 Canda 80 116 8 480 abc 40 54 6 0 Si-1 54 21 0 319 c 178 286 0 0 Fibranova 83 27 0 333 bc 15 373 0 0 Yuma-2 39 29 2 636 abc 119 524 1 0 Han NE 25 12 1 195 c 11 266 0 0 Cherry Blossom x T1 42 25 0 989 abc 38 119 0 0 Eletta Campana 36 52 1 844 abc 30 216 14 0 P value 0.06 0.04 0.05 0.004 0.20 0.19 0.19 - Summer 2020 Bacterivore Fungivore Omnivore Reniform Stunt Spiral Ring RKN Puma-3 141 31 1 78 79 148 73 41 Carmagnola 173 43 1 101 81 109 65 74 Han NE 180 42 4 214 216 409 17 62 NBS 73 19 2 22 54 256 59 20 Wife 215 50 2 69 108 686 447 57 Maverick 111 32 1 32 53 289 35 18 P value 0.21 0.22 0.82 0.15 0.11 0.06 0.25 0.57 a Numbers listed describe the number of a type of nematode found in 200 cc of soil collected from the root area of the hemp cultivars listed. There are multiple types/uses of hemp here. P values were calculated on Log (x+1) transformed data according to Tukey’s HSD where P ≤ 0.05; RKN = root-knot. At NFREC, nematode soil populations consisted of the same PPN as at TREC, although at lower numbers (Table 5). Reniform nematodes (Rotylenchulus reniformis) were the most common, and low populations of stunt nematodes (Tylenchorhynchus spp.), spiral nematodes (Helicotylenchus spp.), ring nematodes (Criconemella spp. or Mesocriconema spp.), and root-knot nematodes (Meloidogyne incognita) were found (Table 5). Only CBD cultivars were planted at this location, and no significant differences were noted among cultivars. Few root galls were observed on cv. Cherry Blossom in the second season. Table 5. Nematode soil populations (no./200 cc soil) at harvest of different hemp cultivars for two consecutive seasons, summer 2019 and summer 2020; NFREC soil survey resultsa. Summer 2019 Bacterivore Fungivore Omnivore Reniform Stunt Spiral Ring RKNb KG 9201 139 40 0 41 1 0 0 3 Cherry Wine 158 47 0 92 4 2 0 5 KG 9202 312 74 1 127 3 0 1 4 Cherry Blossom x T1 307 80 0 239 1 6 0 65 Cherry Blossom 304 68 1 191 2 3 0 16 P value 0.42 0.30 0.54 0.43 0.71 0.36 0.44 0.33 Summer 2020 Bacterivore Fungivore Omnivore Reniform Stunt Spiral Ring RKN Berry Blossom 309 52 0 67 0 2 0 9 Queens Dream 212 32 1 40 2 0 0 3 Cinderella Story 119 61 0 103 1 2 0 38 Cherry Blossom 441 86 0 328 2 4 0 4 Cloud Berry 320 75 2 101 3 0 2 33 Cherry Blonde 167 50 0 53 1 1 0 41 Hot Blonde 319 115 0 453 1 9 0 28 Cherry Wine 224 64 1 90 2 1 0 83 P value 0.95 0.96 0.19 0.83 0.76 0.34 0.62 0.17 a Numbers listed describe the number of a type of nematode found in 200 cc of soil collected from the root area of the hemp cultivars listed. There are only CBD hemp cultivars here. P values were calculated on Log (x+1) transformed data according to Tukey’s HSD where P ≤ 0.05. b RKN, root-knot. Bacterivorous, fungivorous, and omnivorous nematodes (NPPN) were prevalent at all three locations and showed no difference between cultivars or seasons. Discussion Seven different PPN were found on hemp in Florida. Most common were root-knot nematodes, which were found at all three locations, and reniform nematodes, which were found in North (NFREC) and South Florida (TREC). Root-knot nematodes were the dominant nematode in Central Florida (GCREC) after the second season, despite the fact that none were found in the same field at the end of the first season. This shows our limited ability to quantify nematode populations and how difficult it can be to detect the presence of root-knot in a newly planted field. Probably, the deep sand soils that are common in Central Florida harbor significant amounts of root-knot inoculum in its deeper layers (Noling, 2019). It also indicates that in the presence of a good host, root-knot nematodes can build up rapidly in the subtropical climate of Florida. Recent reports from the US and China have shown that hemp can be a good host to various root-knot nematodes, including M. incognita, M. hapla, and M. enterolobii (Song et al., 2017; Kotcon et al., 2018; Ren et al., 2020; Lawaju et al., 2021; Bernard et al., 2022; Coburn et al., 2022). No noticeable root damage symptoms were seen on any of the cultivars, except at GCREC and NFREC on a few plants from cv. Cherry Blossom that showed small but visible root galls. Stubby root nematodes (Nanidorus minor) were the only PPN found in the first season at the GCREC, with slightly higher numbers in the second season. Sting nematodes (Belonolaimus longicaudatus), like root-knot, only showed up in the second season. Both stubby root and sting are ectoparasitic nematodes and very common in the sandy soils of Florida and were only found at the GCREC. They can cause considerable damage to a wide range of crops, such as potato, strawberry, vegetables, corn, citrus, and others (Crow and Brammer, 2001; Noling, 2016). Reniform nematodes were only found at TREC and NFREC and were most common on the fiber cultivars Puma and Carmagnola at TREC in 2019. Reniform nematodes are semi-endoparasitic nematodes that live and feed somewhat immersed within plant roots, with the tail end of the female body protruding into the soil (Wang, 2004). There are around 300 plant species that are known hosts to reniform nematodes. In the southern part of Florida where TREC is located, reniform nematode is a pest to many tropical fruits (McSorley et al., 1983). Additionally, in the northern part of Florida where NFREC is located, reniform nematode is known to be a major nematode pest in cotton (Wang, 2004). Finding this nematode in both locations can be cause for concern due to the types of crops grown in those areas of Florida. Cultivar differences were limited, and none of the 26 cultivars that were planted across the different locations showed a consistent response. More controlled experiments are needed to verify potential cultivar and genetic differences with regard to their host potential to root-knot and other PPN. This will allow us to tell which cultivars are best suited for which location. At TREC, hemp cv. Wife, which was previously reported to be resistant to the RKN species M. incognita (Bernard et al., 2022), had root-knot nematode counts similar to other hemp cultivars. Possibly, this was due to root-knot nematodes in the field trial reproducing on weeds in these plots (which were abundant but not identified), or the root-knot nematode in the field may have been a different species, race, or population. We did not identify the root-knot nematode species at TREC, but previous studies at this location identified the RKN species at this field as M. incognita (Zhang et al., 2022). The above results show that many different PPN can be associated with hemp in Florida, and that populations will vary by region. In North Florida, root-knot and reniform nematodes were the most common. In South Florida, in addition to reniform and root-knot, also spiral, stunt, and ring nematodes were abundant. In Central Florida, root-knot, sting, and stubby root nematodes were found. These regional differences in nematode populations are likely correlated with the different soil types. Sting and stubby root nematodes seem to prefer the almost pure sandy soils in Central Florida, while reniform and spiral nematodes are more common in North and South Florida, where soils are less sandy, having a higher loam and silt content. Root-knot nematodes are of most concern due to their statewide distribution and high damage potential to many of the crops grown in Florida. A literature review by Bernard et al. (2022) on nematode interactions with hemp also reported that Meloidogyne was the most reported genus. No differences were noted in nematode populations for the different cultivars, but more research is needed to study this in more detail and under more controlled conditions. In addition, there is also a need to study to what extent especially root-knot and reniform nematodes can impact hemp production in the state. Hemp production is still very new in Florida and for it to become profitable much more research will be needed. The lack of an earlier soil sample date that would enable comparisons between transplanting and harvest is a weakness in this study and cannot be corrected. Nevertheless, this study provides a “first look” and gives new information on which PPN are associated with different hemp types and cultivars in three different regions of Florida. It is a first step, and it is hoped that this research can form a basis for more focused nematode research in terms of nematode management. Such information would be very valuable to current and future hemp growers in Florida. Acknowledgments The authors acknowledge financial support from Green Roads and Florida Hemp Endowment and technical support provided by David Moreira, Bonnie Xie, Justin Carter, Dustin Jacobs from the University of Florida, and the Gulf Coast Research and Education Center, Wimauma, Florida. ==== Refs Literature Cited Agehara S. 2020 “Using Supplemental Lighting to Control Flowering of Hops in Florida.” EDIS 2020 2 HS1365 10.32473/edis-hs1365-2020 Baidoo R. Joseph T. M. Gu M. Brito J. A. McSorley R. Stamps R. H. Crow W. T. 2016 Mitochondrial haplotype-based identification of root-knot nematodes (Meloidogyne spp.) on cut foliage crops in Florida Journal of Nematology 3 48 193 202 Bernard E. C. Chaffin A. G. Gwinn K. D. 2022 Review of nematode interactions with hemp (Cannabis sativa) Journal of Nematology 54 1 e2022 e2 10.21307/jofnem-2022-002 Bryan H. H. Lance C. J. 1991 Compost trials on vegetables and tropical crops Biocycle 27 3 36 37 Compost trials on vegetables and tropical crops - Univ. of Florida (exlibrisgroup.com). Cherney J. Small E. 2016 Industrial hemp in North America: Production, politics and potential Agronomy (Basel) 6 4 58 10.3390/addrsagronomy6040058 Coburn J. D. Moreira D. Freeman J. Gu M. Bui H. X. Desaeger J. A. 2022 First report of Meloidogyne incognita infecting Cannabis sativa in Florida, USA Nematropica https://journals.flvc.org/nematropica/article/view/132829. Crow W. T. Brammer A. S. 2001 Sting nematode Belonolaimus longicaudatus. ENY-239 Entomology and Nematology Department, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida Gainesville, FL https://edis.ifas.ufl.edu/pdffiles/IN/IN39500.pdf. Desaeger J. 2018 Meloidogyne hapla, the northern root-knot nematode, in Florida strawberries and associated double-cropped vegetables. ENY-070 Entomology and Nematology Department, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida Gainesville, FL https://journals.flvc.org/edis/article/view/106583/117949. Desaeger J. Coburn J. Agehara S. 2022 Using Supplemental Lighting to Grow Hemp in Florida. Institute of Food and Agricultural Sciences The University of Florida Gainesville, Florida https://edis.ifas.ufl.edu/hs1452 Kinlock R. A. Sprenkel R. K. 1994 Plant-parasitic nematodes associated with cotton in Florida Journal of Nematology 26 4 Suppl 749 752 https://pubmed.ncbi.nlm.nih.gov/19279958/ 19279958 Kotcon J. B. Wheeler K. Cline R. Carter S. L. 2018 Susceptibility and yield loss relationships of Meloidogyne hapla and M. incognita infecting Cannabis sativa. [Abstract] Journal of Nematology 50 4 644 https://sciendo.com/pl/article/10.21307/jofnem-2018-057 Lawaju B. R. Groover W. Kelton J. Conner K. Sikora E. Lawrence K. S. 2021 First report of Meloidogyne incognita infecting Cannabis sativa in Alabama Journal of Nematology 53 1 1 3 10.21307/jofnem-2021-052 34179819 Li Y. Zhang M. 2002 Effects of Urea and Nitric Acid on Water and Medium Quality and on Response of Anthurium HortTechnology 12 1 131 134 https://doi.org/10.21273/HORTTECH.12.1.131 Mai W. F. Mullin P. G. Lyon H. H. Loeffler K. 1996 Plant-Parasitic Nematodes: A Pictorial Key to Genera 5th ed. Cornell University Press http://www.jstor.org/stable/10.7591/j.ctv5rdz0t McSorley R. Parrado J. L. Conover R. A. 1983 Population buildup and effects of the reniform nematode on papaya in southern Florida Proceedings Annual Meeting Florida State Horticultural Society 96 198 200 National Academies of Sciences, Engineering, and Medicine 2017 The health effects of cannabis and cannabinoids: The current state of evidence and recommendations for research Washington, DC National Academies Press https://doi.org/10.17226/24625 Noling J. W. 2016 Nematode management in strawberries. Document ENY-031 Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida Gainesville, FL http://edis.ifas.ufl.edu/pdffiles/NG/NG03100.pdf. Noling J. W. 2019 Nematode management in cucurbits (cucumber, melons, squash). Document ENY-025 Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida Gainesille, FL https://edis.ifas.ufl.edu/publication/NG025. Pagan C. Coyne D. Carneiro R. Kariuki G. Luambana N. Affokpon A. Williamson V. M. 2015 Mitochondrial haplotype-based identification of ethanol-preserved root-knot nematodes from Africa Phytopathology 105 350 357 25271352 Ren Z. Chen X. Luan M. Guo B. Song Z. First report of Meloidogyne enterolobii on industrial hemp (Cannabis sativa) in China 2020 Plant disease 105 1 10.1094/PDIS-07-20-1451-PDN Sasser J. N. Root-knot nematodes: A global menace to crop production 1980 Plant disease 64 36 41 10.1094/PD-64-36 Song Z. Q. Cheng F. X. Zhang D. Y. Liu Y. Chen X. W. 2017 First report of Meloidogyne javanica infecting hemp (Cannabis sativa) in China Plant disease 101 5 842 10.1094/PDIS-10-16-1537-PDN Small E. Marcus D. 2002 Hemp: A new crop with new uses for North America 284 326 Janick J. Whipkey A. Trends in new crops and new uses Alexandria, VA ASHS Press Smart G. C. Nguyen K. B. 1991 Sting and awl nematodes 627 668 Nickle W. R. Manual of agricultural nematology New York Marcel Dekker Viglierchio D. R. Schmitt R. V. 1983 On the methodology of nematode extraction from field samples: Baermann funnel modifications Journal of Nematology 3 15 438 444 Wang K.-H. 2004 Reniform Nematode, Rotylenchulus Reniformis Linford & Oliveira (Nematoda: Tylenchida: Tylenchoidea: Hoplolaimidae: Rotylenchulinae): EENY210/IN367, 5/2001 2004 EDIS 2004 9 20230018 https://doi.org/10.32473/edis-in367-2001. Williams D. W. 2019 Industrial hemp as modern commodity crop Madison, WI American Society of Agronomy, Crop Science Society of America, Soil Science Society of America https://acsess.onlinelibrary.wiley.com/doi/book/10.2134/industrialhemp Zhang W. Sun W. Wang Y. Liu H. Zhang S. Dong B. Ji Z. Qiao K. Management of Meloidogyne incognita on cucumber with a new nonfumigant nematicide fluopimomide 2022 Plant disease 106 1 151 155 10.1094/PDIS-05-21-0943-RE 34515507

PMC010xxxxxx/PMC10324121.txt

==== Front J Nematol J Nematol jofnem jofnem Journal of Nematology 0022-300X 2640-396X Sciendo 37426724 jofnem-2023-0028 10.2478/jofnem-2023-0028 Research Paper Belonolaimus longicaudatus management using metam potassium and fluensulfone in potato Grabau Zane J. zgrabau@ufl.edu Liu Chang Navia Gine Pablo A. Entomology and Nematology Department, University of Florida, 1881 Natural Area Drive, Gainesville, FL 32611, United States 1728 Shoreside Circle, Wellington, FL 33414, United States This paper was edited by Shaun D Berry. 2 2023 6 7 2023 55 1 2023002823 3 2023 © 2023 Zane J. Grabau et al., published by Sciendo 2023 Zane J. Grabau et al., published by Sciendo https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. Abstract Belonolaimus longicaudatus (sting nematode) is an important pest in Florida potato production and is managed primarily by fumigation using 1,3-dichloropropene (1,3-D). Other effective nematicides are needed for more flexibility in managing this pest. The objective of this study was to evaluate fluensulfone, metam potassium, and mixtures of the two products, relative to 1,3-D and untreated control, for efficacy at managing sting nematode, and for non-target effects on free-living nematodes in potato. To test this objective, a small-plot field experiment was conducted in northeast Florida in 2020 and repeated in 2021. Metam potassium fumigation (390 kg a.i./treated ha)—with or without fluensulfone—managed sting nematode soil abundances but was phytotoxic to potato. Strategies that mitigate metam potassium phytotoxicity, such as reduced application rates, are needed before efficacy of metam potassium in this system can be determined. As a preplant soil spray, fluensulfone alone (403 g a.i./treated ha) did not manage sting nematode abundances and had an inconsistent effect on yield. Fumigation with 1,3-D (88.3 kg a.i./treated ha) was the only treatment that consistently managed sting nematode and increased potato yield. Nematicides did not consistently affect free-living nematodes. Keywords 1,3-Dichloropropene Belonolaimus longicaudatus fluensulfone fumigant metam potassium nematicide nematode community potato Solanum tuberosum sting nematode ==== Body pmcPotato (Solanum tuberosum) production is an important industry in the United States, with the 2021 potato crop worth 4.17 billion dollars (USDA-NASS, 2022). Potato production is also a valued industry in Florida, with 8,500 hectares of production, and a total crop value of 95 million dollars in 2021 (USDA-NASS, 2022). While this is a relatively small proportion of U.S. production (377,000 hectares), it is a substantial portion of spring potato production, and Florida is the leading potato-producing state in the Southeast (USDA-NASS, 2022). Florida potato production is exclusively short season (approximately 100 days) types grown for the fresh market and chipping industry (USDA-NASS, 2022). Sting nematode (Belonolaimus longicaudatus) is widespread in Florida potato production—particularly in the important production area of northeast Florida—and highly damaging (Crow et al., 2000b). Crop rotation consisting of continuous winter potato followed by summer sorghum–sudan grass (Sorghum x drummondii) cover crop is standard in northeast Florida potato production, and sting nematode increases on both of these crops (Crow et al., 2000a, 2001). This rotation, along with ubiquitous sandy soils, helps make the sting nematode prevalent in Florida potato production. Management of sting nematode—and other plant-parasitic nematodes—on potato in Florida relies heavily on nematicide application. There are a limited number of nematicides available, so additional effective nematicides are always needed to reduce dependency on current options. For sting nematode management in potato, fumigation with 1,3-dichloropropene (1,3-D) is the most common tactic, and growers may also apply chloropicrin in the fumigant mixture because it improves soil-borne disease management (Csinos et al., 2000; Gilreath et al., 2004). More recently, the non-fumigant nematicide fluensulfone was introduced and has shown efficacy at managing sting nematode in Florida potato (Grabau et al., 2019), although fluensulfone is not widely adopted in commercial production. Another nematode management option is fumigation with metam salts (metam potassium or metam sodium), which have been available for many years and are used by some Florida potato growers. Metam is a broad-spectrum fumigant with activity against fungal pathogens and weeds as well as nematodes (Csinos et al., 2000; Gilreath et al., 2004; Desaeger et al., 2017). However, despite being used commercially, there is need for formal evaluation of metam efficacy for managing sting nematodes in potato, due to a lack of published reports. The only reports on efficacy of metam salts against sting nematode are on strawberry (Fragaria x ananassa) and tomato (Solanum lycopersicum) in Florida (Santos et al., 2006; Gilreath et al., 2008; Watson and Desaeger, 2019). Results have been mixed in those systems with metam salts variously being effective (Santos et al., 2006), ineffective (Watson and Desaeger, 2019), or effective only in combination with chloropicrin (Gilreath et al., 2008). Metam products have been evaluated more broadly against root-knot nematodes in the Southeast. Metam can be effective for root-knot nematode management (Csinos et al., 2000; Desaeger and Csinos, 2006), but is considered less consistent for nematode management than 1,3-D in the sandy soils common in the Southeast (Desaeger et al., 2017; Watson and Desaeger, 2019). A potential tactic for improving nematode management by metam potassium (MP) is combining MP with an effective nematicide such as fluensulfone (FL). This may also increase grower adoption relative to either product alone, as a successful mixture would have efficacy against a wider spectrum of pests. Fumigants are commonly combined or applied in tandem for this purpose (Csinos et al., 2002; Desaeger et al., 2017). Applying a non-fumigant nematicide at or after planting to supplement preplant fumigation has also been tested, albeit with varying levels of improved efficacy relative to fumigation alone (Desaeger et al., 2017; Watson and Desaeger, 2019; Grabau et al., 2021). Mixing a fumigant with a non-fumigant nematicide is more novel, but mixtures of fluensulfone and MP effectively managed sting nematode and southern root-knot nematodes in initial tests in Southeast vegetable production (Navia Gine and Hajihassani, 2021). As emphasis on sustainability and the entire soil community—including beneficial organisms—has increased, so too has the importance of considering non-target effects of pesticides, including nematicides. In particular, nematicide application can have unintended negative impacts on free-living nematodes (Waldo et al., 2019; Grabau et al., 2020). Free-living nematodes feed on fungi, bacteria, or other nematodes and can provide benefits to the soil ecosystem such as nutrient cycling (Holajjer et al., 2016; Trap et al., 2016) and pest regulation (Khan and Kim, 2005). Therefore, evaluating free-living nematodes provides important additional information about how nematicides affect the broader agroecosystem. Based on these needs, the objectives of this study were to evaluate the efficacy of MP and a mixture of MP and FL on (1) management of plant-parasitic nematodes (2) yield response, and (3) non-target impacts on free-living nematodes in Florida potato production. Materials and Methods Site and Experimental Design To investigate these objectives, a field experiment was conducted at the University of Florida Hastings Agricultural Education Center near Hastings, Florida (29.692, −81.441). The soil type was Ellzey fine sand (sandy, silicaceous, hyperthermic Arenic Ochraqualf) with 95% sand, 2% silt, 3% clay; and <1% organic matter. Production was on bareground, raised beds irrigated by a subsurface tile system, which is common in the region. Cropping history was continuous spring potato followed by summer sorghum–sudan grass cover crop. The experiment was conducted twice, with the first trial occurring in early 2020 and the second trial occurring in early 2021, with some establishment activities occurring in the preceding year for each trial. Experimental units were 4-row beds 19.8 m long. Center-to-center bed spacing was 101 cm. The experiment was a randomized complete block design with six replicates and a single factor: nematicide treatment. There were five treatments, including 1) untreated control (UTC), 2) MP alone (MP-A), 3) MP mixed with fluensulfone (MP+FL), 4) fluensulfone alone (FL-A), and 5) 1,3-D. Treatment rates and products are described in Table 1. Table 1. Chemistries and rates for nematicide treatments. Treatment abbreviation Product Active ingredient Liters product/bedded haa kg a.i./bedded ha Method 1. UTC Untreated control None - - - 2. MP-A Metam KLRb Metam potassium 562 390 Fumigation 3. MP+FL Metam KLR Metam potassium 562 390 Fumigation (tank mix) NIMITZc Fluensulfone 7.0 3.36 4. FL-A NIMITZ Fluensulfone 7.0 3.36 Ground spray 5. 1,3-D Telone IId 1,3-dichloropropene 75 88.3 Fumigation a Bedded acreage was 75% of broadcast acreage. In this trial, the bedded acreage was considered the acreage to which treatments were applied. Further details of treatment applications are described in the materials and methods. b Metam KLR (Taminco, Allentown, PA) c NIMITZ (ADAMA, Research Triangle Park, NC) d Telone II (Dow Agrosciences LLC, Zionsville, IN) Fumigation application methods Metam potassium (alone or as a mixture) and 1,3-D were applied by preplant fumigation using shank injection, but fumigation rig setup varied slightly for each product according to standard practices in the region. For 1,3-D, the fumigation rig had 1 shank per bed (101 cm spacing between shanks). The MP-A and MP+FL treatments were applied using 3 shanks per bed that were spaced 13 cm apart. More shanks per bed were used for MP-A and MP+FL to ensure adequate bed coverage due to relatively low volatility of MP (Gertzl et al., 1977). For MP+FL, FL and MP were combined in the containing tank of the fumigation rig in the correct proportions before application. Fumigation was conducted 25 or 26 days before planting each season (Table 2). In all cases, nematicides were applied to flattened beds that had been tilled with a rotary chopper and subsoiler after summer sorghum–sudan grass production. Fumigants were delivered approximately 15 cm deep in flattened beds and immediately followed by a second tractor equipped with angled disc implements that formed a hilled bed of loose soil in fumigant-treated plots. The fumigation rig that delivered 1,3-D was also equipped with angled discs directly on the rig. After bed formation, final fumigant placement was at least 30 cm below the tops of beds. Bed formation is the standard practice for containing fumigants in the soil—as required on fumigant labels—in potato production in the area. Table 2. Schedule for data collection and trial establishment. Numbers in parentheses are days before (negative) or after (positive) planting. Event Year 1 (2019–2020) Year 2 (2020–2021) Preplant soil sampling 18 Dec (−38) 14 Dec (−30) Nematicide application 2 Jan (−25) 18 Dec (−26) Beds tilled and re-formed 13 Jan (−14) 5 Jan (−8) Planting 27 Jan (0) 13 Jan (0) Stand count 1 26 Feb (30) 18 Feb (36) Stand count 2 4 Mar (37) 26 Feb (44) Vigor rating 1 26 Feb (30) 18 Feb (36) Vigor rating 2 4 Mar (37) 26 Feb (44) Vigor rating 3 18 Mar (52) 3 Mar (50) Vigor rating 4 – 12 Mar (59) Midseason soil sampling 4 Mar (37) 26 Feb (44) Harvest soil sampling 8 May (103) 14 Apr (92) Harvest 11 May (106) 21 Apr (99) Fluensulfone application For the standalone treatment, FL was sprayed on top of the soil as a broadcast application using a commercial tractor and boom sprayer. Immediately following application, fluensulfone was incorporated to 15 cm by tilling with a rotary chopper. Fluensulfone was applied on the same day as fumigant treatments, either 25 or 26 days before planting (Table 2). On the same day, following application of all treatments, individual fluensulfone and untreated control plots were angle-disced to form beds. This ensured that all plots were bedded the same number of times. Then, the entire trial was angle-disced to form uniform beds. Crop production One to two weeks before planting (Table 2), preplant fertilizer was applied, followed by incorporation with a rotary chopper after which beds were re-formed. In each trial, “Red LaSota”, a red fresh market potato common in the area, was grown. Seed potatoes were planted in January (Table 2) using a mechanical planter. The planter was equipped with a tillage implement that opened a furrow in the beds. Seed pieces were planted at approximately 21 cm spacing in the seed furrow. After the trial was planted, a tank-mix of standard fungicide (azoxystrobin at 191 g a.i./ha) and insecticide (fipronil at 98 g a.i./ha) treatments were sprayed in furrow. The furrow was then closed, and the beds were formed again. Aside from nematicide treatments, each year, the whole trial was managed uniformly with full herbicide, fungicide, and insecticide regimes as standard for the area. Environmental conditions following nematicide application Environmental conditions can affect nematicide efficacy, and soil moisture and temperature requirements are listed on fumigant labels. Soil moisture was not quantified formally at the site but was acceptable at fumigation for both the 2020 and 2021 trials based on the USDA-established feel test (USDA-NRCS, 2005). Early season soil temperature, air temperature, and rainfall are described in Table 3 based on information from the Florida Automated Weather Network (https://fawn.ifas.ufl.edu/) station at the study site. Briefly, the 2020 trial was very dry between fumigation and planting, whereas the 2021 trial received more regular rainfall during that period (Table 3). In both years soil temperatures were acceptable for fumigation, with higher temperatures in 2020 than 2021. Table 3. Environmental conditions during the weeks before and after nematicide application for trials in 2019–2020 and 2020–2021 Time Perioda 2019–2020 2020–2021 Temperature (°C) Temperature (°C) Rainfallb Soilc Air Rainfall Soil Air 2 WBA 7.54 16.9 16.0 0.53 16.3 12.0 1 WBA 1.02 18.6 18.3 0.69 16.7 15.2 Day of application 0 16.8 16.5 0 14.9 7.3 1 WAA 0.53 16.4 14.0 2.54 15.5 13.3 2 WAA 0.00 18.4 21.1 0.25 14.3 13.9 3 WAA 0.00 17.0 13.1 0.23 15.9 14.2 4 WAAd 0.03 16.1 13.2 0.20 14.2 11.2 5 WAA 1.42 16.3 15.4 0.13 13.3 11.3 6 WAA 0.08 17.3 17.5 2.51 16.1 15.8 7 WAA 0.86 18.5 18.1 0.51 13.4 11.1 8 WAA 1.93 17.1 14.0 6.73 16.7 18.1 9 WAA 0.20 16.9 16.4 5.66 18.2 17.7 10 WAA 0.08 17.3 17.5 2.51 16.1 15.8 a WBA and WAA are weeks before nematicide application and weeks after nematicide application, respectively. b Rainfall (cm) is total for the respective week or day of nematicide application. c Temperatures are mean for the week or day. d Potatoes were planted during 4 weeks after application, at 25 or 26 days after application in 2020 and 2021, respectively. Potato harvest and yield collection Potatoes were grown for approximately 100 days before harvest in late April or early May (Table 2). From the central 2 rows of each plot, potatoes were harvested from 6.1 m of each row using a single-row mechanical harvester. Separately by plot, harvested potatoes were washed and graded using a mechanical size sorter. In 2021, the mechanical size sorter was supplemented by human intervention to remove culls with external quality defects. In 2020, quality culls were not removed or measured due to workforce restrictions during the COVID-19 pandemic. For each plot, potatoes were weighed for each grade category. Size classes were A3 (8.3–10.2 cm diameter), A2 (6.4–8.29 cm diameter), A1 (4.8–6.39 cm diameter), B (3.8–4.69 cm diameter), and C (1.3–3.79 cm diameter) based on local and USDA guidelines (USDA-AMS, 2011). No potatoes with diameter greater than 10.2 cm (class A4) were observed in either trial. All class A potatoes were considered marketable (USDA-AMS, 2011). In 2021, the categories of quality-related culls included tubers that were rotted, green (sun damage), had growth cracks, or were misshapen. In addition to weight by each cull category, the total weight of quality culls and the grand total cull weight (quality culls plus undersize tubers) was reported in 2021. For both years, total cull weight of tubers that were under marketable size (class B and C), as well as grand total yield is reported. For each plot, 20 tubers were quartered and inspected for internal defects in each trial, but there were negligible defects, so this data was not reported. Plant stand and vigor assessment Plant stand was estimated by counting emerged shoots from 2.4 m in each of the central 2 rows of each plot and converted to stand per 3 m row for reporting. In both 2020 and 2021, plant stand was estimated twice: (1) when shoots started to emerge and (2) when all shoots had fully emerged (Table 2). Plant vigor was rated visually 3 to 4 times during the middle of the growing season (Table 2). Vigor was assessed from central 2 rows of each plot based on a 0 to 100 scale where 0 indicates all plants dead, 25 is a very poor crop, 50 is a below-average crop, 75 is an average crop, and 100 is an exceptionally excellent crop for the given time point. Vigor assessment was based primarily on the canopy height and size. Additionally, symptoms of phytotoxicity were observed in the 2020 trial and were rated at 37 days after planting (DAP). These symptoms are described further in the results. Nematode quantification Nematode soil abundances in each plot were quantified before nematicide application (preplant), at midseason (6 weeks after planting), and near harvest each year (Table 2). Using an Oakfield tube, 12 soil cores to 25 cm depth were collected from the central 2 beds of each plot, with samples collected near plant roots at midseason and harvest. Soil was homogenized manually, and nematodes were extracted using the sucrose-centrifugation method (Jenkins, 1964). Plant-parasitic nematodes were identified morphologically and quantified by microscope. Total free-living nematode abundances were also quantified, but taxonomic identification of free-living nematodes was not completed. Statistical analysis Each variable was analyzed separately for each trial and sampling date using one-way ANOVA. Before completing ANOVA, response variables were transformed, if needed, to meet assumptions of homogeneity of variance using Levene’s test (Levene, 1960) and normality of residuals based on graphing (Cook and Weisburg, 1999). Nematicide effects were considered significant at α=0.05 in ANOVA. Nematicide treatment means were separated by Fisher’s protected LSD (α=0.05) if main effects were significant in ANOVA. Analyses were conducted in R statistical software (version 3.4.4, The R Foundation for Statistical Computing, Vienna, Austria). Results Plant stand and vigor In both 2020 and 2021, MP-A or MP+FL generally significantly reduced plant stand, particularly for the earlier assessment conducted at seedling emergence (Fig. 1). In 2020, MP-A or MP+FL significantly decreased plant stand at 30 DAP relative to all other treatments. In 2020, stand at 37 DAP—when plants had fully emerged—was significantly less for MP-A than any other treatment and less for MP+FL than FL-A. At 36 DAP in 2021, MP+FL significantly reduced stand relative to all other treatments and MP-A reduced stand relative to 1,3-D. By 44 DAP, MP+FL had less stand than 1,3-D or FL-A, but there were no other significant differences among treatments. Figure 1: Potato plant stand as affected by nematicide application in 2020 and 2021. Timing indicated in days after planting (DAP). Bars and brackets represent means (N=6) and standard errors, respectively. “UTC”, “MP-A”, “MP+FL”, “FL-A”, and “1,3-D” represent untreated control, metam potassium alone, metam potassium mixed with fluensulfone, fluensulfone alone, and 1,3-dichloropropene nematicide treatments, respectively. Within each subfigure, means with different letters are significantly different (Fisher’s protected LSD, α=0.05). Plant vigor and phytotoxicity In 2020, MP-A and MP+FL significantly reduced plant vigor compared with any other treatment at each assessment date (Figs. 2, 3). Qualitative, visual symptoms of phytotoxicity—particularly browning of shoot tips and delayed leaf emergence producing a burnt appearance—were observed for those treatments in 2020 beginning at 30 DAP (Fig. 4). When phytotoxicity was rated at 37 DAP in 2020, symptoms were only observed for MP-A and MP+FL (Fig. 2). By the next assessment date (50 DAP), no browning was observed on leaves. Figure 2: Potato plant vigor and phytotoxicity during production 2020 as affected by nematicide application. Timing indicated in days after planting (DAP). Plant vigor is a visual rating on a 0–100 scale with 100 best. Bars and brackets represent means (N=6) and standard errors, respectively. “UTC”, “MP-A”, “MP+FL”, “FL-A”, and “1,3-D” represent untreated control, metam potassium alone, metam potassium mixed with fluensulfone, fluensulfone alone, and 1,3-dichloropropene nematicide treatments, respectively. Within each subfigure, means with different letters are significantly different (Fisher’s protected LSD, α=0.05). Figure 3: Reduced stand and plant vigor in metam+fluensulfone treatment (right) compared with untreated control (left) at 36 days after planting in 2020. Figure 4: Symptoms of phytotoxicity on emerging potato shoots in treatments with metam potassium alone or in combination with fluensulfone at 36 days after planting in 2020. There is browning at tips of shoots and leaf unfolding is delayed. In 2021, for all assessment dates, plant vigor was significantly less for MP+FL than any other treatment, except that it was similar to MP-A at 59 DAP (Fig. 5). Plant vigor was also significantly less for MP-A than 1,3-D at each assessment date and less than FL-A at 36 and 59 DAP. Later in the season (50 and 59 DAP), plant vigor was significantly less for UTC than 1,3-D. Figure 5: Potato plant vigor during production in 2021 as affected by nematicide application. Timing indicated in days after planting (DAP). Plant vigor is a visual rating on a 0–100 scale with 100 best. Bars and brackets represent means (N=6) and standard errors, respectively. “UTC”, “MP-A”, “MP+FL”, “FL-A”, and “1,3-D” represent untreated control, metam potassium alone, metam potassium mixed with fluensulfone, fluensulfone alone, and 1,3-dichloropropene nematicide treatments, respectively. Within each subfigure, means with different letters are significantly different (Fisher’s protected LSD, α=0.05). Potato yield In 2020, total and marketable yields were significantly greater for 1,3-D than MP-A and UTC (Fig. 6). Total and marketable yields were also significantly greater for FL-A than MP-A in 2020. Class A1 and A2 yields were significantly affected by nematicides in 2020 and followed the same trends as marketable yield (Table 4). No other size category was significantly affected by nematicide treatments in 2020. Figure 6: Marketable and total yield as affected by nematicide application in 2020 and 2021. Marketable yield is USDA size A tubers without defects. Total yield includes marketable yield and culls due to size or quality. Bars and brackets represent means (N=6) and standard errors, respectively. “UTC”, “MP-A”, “MP+FL”, “FL-A”, and “1,3-D” represent untreated control, metam potassium alone, metam potassium mixed with fluensulfone, fluensulfone alone, and 1,3-dichloropropene nematicide treatments, respectively. Within each subfigure, means with different letters are significantly different (Fisher’s protected LSD, α=0.05). Table 4. Potato tuber yield (kg/ha) by grade and nematicide treatment.a Grade Cb B A1 A2 A3 Size cullsc Nematicide Treatment 2020 Untreated control 805 1,408 4,965 bc 13,243 bc 451 2,213 Metam potassium 637 849 3,904 c 10,613 c 303 1,486 Metam+Fluensulfone 839 1,147 6,014 abc 14,429 abc 468 1,986 Fluensulfone 773 1,832 7,954 ab 17,112 ab 547 2,606 1,3-dichlorpropene 1,039 1,969 8,948 a 18,884 a 54 3,008 Nematicide Treatment 2021 Untreated control 1,394 ab 1,914 b 6,934 7,782 c 0 3,308 b Metam potassium 1,267 b 1,940 b 7,266 11,596 a 0 3,208 b Metam+Fluensulfone 1,145 b 1,763 b 6,845 9,066 bc 0 2,908 b Fluensulfone 1,272 b 1,998 b 6,729 7,577 c 0 3,270 b 1,3-dichlorpropene 1,585 a 2,511 a 7,645 10,886 ab 0 4,095 a a Different letters in the same column indicate significantly different means based on Fisher’s protected LSD at α = 0.05. b Tuber grades of C, B, A1, A2, and A3 include harvest potatoes with diameters of 1.3–3.7, 3.8–4.7, 4.8–6.3, 6.4–8.3, and 8.4–10.2 cm respectively. Grades A1, A2, and A3 are considered marketable. c Size culls include grade C and B tubers In 2021, total yield was significantly greater for MP-A and 1,3-D than any other treatment. Marketable yield in 2021 followed a similar trend except that MP+FL had similar yield to 1,3-D. In 2021, there were no size A3 potatoes and A1 yield was not significantly affected by nematicide treatments (Table 4). Class A2 yield was significantly affected by nematicides in 2020, with yield significantly greater for MP-A than each treatment except 1,3-D and significantly greater for 1,3-D than UTC or FL-A. In 2021, the only year quality-related culls were reported, rotted and grand total cull weights were significantly affected by nematicides and generally followed the same trends as marketable yield (Table 5). In contrast, green cull weight was significantly greater for MP-A than UTC, FL, or 1,3-D and greater for MP+FL than UTC or FL-A. Growth crack, misshapen, or total quality culls were not significantly affected by nematicide treatments in 2021. Table 5. Weight (kg/ha) of tubers culled due to quality defects by category in 2021 season.a Nematicide Treatment Rotted Green Growth crack Misshapen Total quality cullsb Grand total cullsc Untreated control 469 ab 72 c 443 305 1,289 4,598 b Metam potassium 237 bc 457 a 575 291 1,561 4,768 b Metam+Fluensulfone 291 abc 299 ab 335 41 967 3,874 c Fluensulfone 145 c 65 c 941 168 1,319 4,590 b 1,3-dichlorpropene 545 a 74 bc 1,127 37 1,782 5,877 a a Different letters in the same column indicate significantly different means based on Fisher’s protected LSD at α = 0.05. Due to COVID-19 restrictions, tuber defect culls were not measured in 2020 b Total quality culls is the sum of tubers in the rotted, green, growth crack, and misshapen categories. c Grand total culls is the sum of size and quality culls. Plant-parasitic nematodes Each year, sting nematode abundances were relatively low before fumigation and did not vary significantly based on upcoming treatments (Fig. 7). Sting nematode abundances were generally greater for UTC or FL-A than treatments including MP or 1,3-D at midseason in 2020 and 2021 as well as harvest in 2021, but exact significance of treatment separation varied by season (Fig. 7). Sting nematode abundances were low and unaffected by nematicides at harvest in 2020. Figure 7: Sting nematode soil population densities at preplant (before fumigation), midseason (37 to 44 days after planting), and harvest in 2020 and 2021 as affected by nematicide treatments. Bars and brackets represent means (N=6) and standard errors, respectively. “UTC”, “MP-A”, “MP+FL”, “FL-A”, and “1,3-D” represent untreated control, metam potassium alone, metam potassium mixed with fluensulfone, fluensulfone alone, and 1,3-dichloropropene nematicide treatments, respectively. Within each subfigure, means with different letters are significantly different (Fisher’s protected LSD, α=0.05). effects for a given season (ANOVA, P < 0.05). At harvest in 2020, stunt nematode abundances were significantly greater for UTC, FL-A, or 1,3-D than MP-A or MP+FL (Fig. 8). At midseason and harvest in 2021, stunt nematode abundances were significantly greater for UTC or FL-A than any other nematicide treatment. Stunt nematode abundances were not significantly affected by nematicides at midseason in 2020 or preplant in either year. Figure 8: Stunt nematode soil population densities at preplant (before fumigation), midseason (37 to 44 days after planting), and harvest in 2020 and 2021 as affected by nematicide treatments. Bars and brackets represent means (N=6) and standard errors, respectively. “UTC”, “MP-A”, “MP+FL”, “FL-A”, and “1,3-D” represent untreated control, metam potassium alone, metam potassium mixed with fluensulfone, fluensulfone alone, and 1,3-dichloropropene nematicide treatments, respectively. Within each subfigure, means with different letters are significantly different (Fisher’s protected LSD, α=0.05). Free-living nematodes Free-living nematode soil abundances were significantly affected by nematicides only at harvest in 2020 (Fig. 9). In that season, free-living nematode abundances were significantly greater following 1,3-D than MP-A or MP+FL and greater following FL-A than MP+FL. Figure 9: Free-living nematode soil population densities at preplant (before fumigation), midseason (37 to 44 days after planting), and harvest in 2020 and 2021 as affected by nematicide treatments. Bars and brackets represent means (N=6) and standard errors, respectively. “UTC”, “MP-A”, “MP+FL”, “FL-A”, and “1,3-D” represent untreated control, metam potassium alone, metam potassium mixed with fluensulfone, fluensulfone alone, and 1,3-dichloropropene nematicide treatments, respectively. Within each subfigure, means with different letters are significantly different (Fisher’s protected LSD, α=0.05). Discussion The main objective of this study was to evaluate MP alone or in a mixture with FL for sting nematode management in potato production. Based on this study, any treatments with MP at the rate tested in this study (390 kg a.i./bedded ha) are not practical in the tested system due to risk of phytotoxicity. Metam potassium was phytotoxic to potato based on burn symptoms on shoots in 2020 as well as delayed emergence and reduced plant vigor relative to untreated in both years in this study. Phytotoxicity was generally less severe in 2021, with no qualitative symptoms (shoot browning) observed and plant stand and vigor impacts of a somewhat lesser magnitude. There was slight variation in severity of phytotoxicity between MP alone and in mixture based on some indicators, but symptoms were observed consistently enough in both treatments that it was clear that MP caused the phytotoxicity, rather than an interaction of FL and MP. Influence of MP treatments on potato yield was inconsistent, probably because MP both managed sting nematode abundances and was phytotoxic. Reduction of sting nematode abundances by both MP treatments suggests MP has potential for managing this potato pest, with no advantage of MP and FL mixture relative to MP alone, but phytotoxicity in this system must be resolved for either treatment to be practical. Fumigants—including metam-based fumigants—are known to be phytotoxic to crops if the fumigant does not dissipate sufficiently before the crop is planted (Csinos et al., 2002; Desaeger et al., 2008). However, MP use is widespread in various crops—including some use in Florida potato production—and metam has not been phytotoxic in most previous studies (Ingham et al., 2007; Desaeger et al., 2017; Watson and Desaeger, 2019). So why was there phytotoxicity in this system and could adjustments be made to avoid phytotoxicity? Many factors can contribute to fumigant dissipation time and risk of phytotoxicity, including crop type (Csinos et al., 2002), fumigation rate (Ma et al., 2001; Klose et al., 2008), plantback time after fumigation (Csinos et al., 2002), application methods (Csinos et al., 2002; Desaeger et al., 2008), soil type (Ashley et al., 1963; Triky-Dotan et al., 2007), and environmental conditions at and following fumigation (Ashley et al., 1963; Gerstl et al., 1977; Desaeger et al., 2008). Based on these factors, phytotoxicity risk will vary by cropping system and these factors also suggest options for mitigation. Regarding mitigation options, decreasing the MP application rate may decrease the risk of phytotoxicity and is the most practical option available in this cropping system. At a lower rate, there would be a lower initial fumigant concentration, so it should take less time to dissipate to a level that is not phytotoxic (Ashley et al., 1963; Klose et al., 2008). In this study, MP was applied at the maximum labelled rate (390 kg a.i./bedded ha), whereas, based on inquiries with local growers, around 170 kg a.i./bedded ha is more typical in potato production in northeast Florida (personal communication). However, to our knowledge, there are no peer-reviewed studies testing a reduced MP rate in this cropping system. Therefore, further research would be needed to determine if there is a MP rate—either alone or as a mixture with FL—in this system that is not phytotoxic but is still effective at managing sting nematodes. Other factors, among those discussed above, may have contributed to phytotoxicity in this system, but it is generally either not practical or possible to control these factors. Additionally, for parameters that can be controlled (application methods, environmental conditions at fumigation, plantback time, etc.), instructions are listed on the fumigant product label and were followed, as is legally required for applicators. Increasing time between fumigation and planting could reduce phytotoxicity risk (Csinos et al., 2002), but anything longer than the 25-day period used in this study is typically not realistic. Similarly, in this study, tilling of beds during preplant fertilizer application as well as opening furrow for planting should have helped aerate the soil to dissipate fumigant residues, a common technique to mitigate risk of phytotoxicity. Conditions were abnormally dry after fumigation, particularly in 2020, and this system is largely dependent on rainfall to water in treatments, so dry conditions may have contributed to phytotoxicity. Typically, cool wet conditions increase time for metam salts and their derivatives to dissipate and decrease crop damage risk (Ashley et al., 1963; Gerstl et al., 1977). However, in this study, phytotoxicity was more severe in 2020—when conditions were very dry—than 2021. Phytotoxicity from metam salts under abnormally dry conditions has also been reported on cantaloupe (Cucumis melo var. melo) in central Florida (J.D. Desaeger, University of Florida, personal communication) and potato in Northeast Florida, via personal communication with local growers. This suggests extremely dry conditions are also a risk factor for phytotoxicity, although further research is needed to confirm and explain the mechanisms for this. Fumigation with 1,3-D was the only consistently effective nematicide for managing sting nematode in this study considering both soil nematode abundances and potato yield. This is consistent with prior research and commercial practices as 1,3-D has typically been effective at managing sting nematode in previous studies (Weingartner and Shumaker, 1990; Crow et al., 2000a; Grabau et al., 2019) and it is a standard grower practice. Fumigation with 1,3-D was not as consistently effective at managing stunt nematodes, but this pest is not known to impact potato yield (Crow et al., 2000b). Fluensulfone was not effective at managing sting nematode abundances in this study and provided an inconsistent yield return. In a previous study in Florida potato (Grabau et al., 2019), FL—at a similar rate to this study—was as consistently effective as 1,3-D for managing sting nematode abundances and increasing yield over a 3-year study. Environmental factors—particularly rainfall—may affect efficacy of FL and other nematicides, but rainfall and temperature varied by year and did not clearly explain differences in FL efficacy between those studies. None of the nematicides tested had a consistent effect on free-living nematodes. Typically, fumigation with MP (Collins et al., 2006; Watson and Desaeger, 2019) or mixtures that include 1,3-D (Sanchez-Moreno et al., 2010; Timper et al., 2012; Watson and Desaeger, 2019) decreases free-living nematode soil abundances broadly across most trophic groups. In contrast, there is some prior information that FL has fewer non-target effects on free-living nematodes than other chemistries (Kearn et al., 2017; Waldo et al., 2019). Certain trophic groups, particularly fungivores, are often more sensitive to nematicides than other trophic groups (Fiscus and Neher, 2002; Watson and Desaeger, 2019; Grabau et al., 2020). Because only total free-living nematodes and not individual trophic group abundances were assessed in this study, it is possible that nematicides had undetected, negative impacts on individual trophic groups in this study. While infrequent, there are prior instances where fumigants had minimal impact on overall free-living nematode abundances (Desaeger et al., 2017). As a highly disturbed system with very frequent tillage and abundant inputs (fertilizers, pesticides, organic matter from crops), it is also possible that the nematode community in this system is composed predominantly of enrichment opportunists that rebound quickly after fumigation. More research would be needed to support that hypothesis. In conclusion, MP at 390 kg a.i./ha, alone or as a mixture with FL, is not acceptable in this potato production system due to phytotoxicity from MP. Further research would be needed to determine if adjusting MP application parameters—particularly reduced application rates—could mitigate phytotoxicity while retaining effective sting nematode management in this cropping system. Based on this study, fumigation with 1,3-D continues to be the most consistently effective option for managing sting nematode in potato production. On its own, FL was not consistently effective for managing sting nematode despite consistent efficacy in prior research. Nematicides did not have a negative impact on the overall free-living nematode community, but further research would be needed to assess impacts on individual groups of free-living nematodes in this system. Acknowledgements Thanks to Scott Chambers and Hastings Agricultural Education Center staff. Material or monetary support from ADAMA and Taminco Inc. This work was supported by USDA-NIFA Research Capacity Fund (Hatch Multistate) project FLA-ENY-006281. ==== Refs Literature Cited Ashley M. G. Leigh B. L. Lloyd L. S. 1963 The action of metham-sodium in soil. II.—Factors affecting the removal of methyl isothiocyanate residues Journal of the Science of Food and Agriculture 14 153 161 Collins H. P. Alva A. Boydston R. A. Cochran R. L. Hamm P. B. McGuire A. Riga E. 2006 Soil microbial, fungal, and nematode responses to soil fumigation and cover crops under potato production Biology and Fertility of Soils 42 247 257 Cook R. D. Weisburg S. 1999 Response transformations 316 333 Applied regression including computing and graphics New York City, New York Wiley–Interscience Crow W. T. Weingartner D.P. Dickson D. W. 2000a Effects of potato-cotton cropping systems and nematicides on plant-parasitic nematodes and crop yields Journal of Nematology 32 297 302 19270980 Crow W. T. Weingartner D. P. McSorley R. Dickson D. W. 2000b Damage function and economic threshold for Belonolaimus longicaudatus on potato Journal of Nematology 32 318 322 19270983 Crow W. T. Weingartner D. P. Dickson D. W. McSorley R. 2001 Effect of sorghum–sudangrass and velvetbean cover crops on plant-parasitic nematodes associated with potato production in Florida Journal of Nematology 33 285 288 19265888 Csinos A. S. Sumner D. R. Johnson W. C. Johnson A. W. McPherson R. M. Dowler C. C. 2000 Methyl bromide alternatives in tobacco, tomato and pepper transplant production Crop Protection 19 39 49 Csinos A. S. Webster T. M. Sumner D. R. Johnson A. W. Dowler C. C. Seebold K. W. 2002 Application and crop safety parameters for soil fumigants Crop Protection 21 973 982 Desaeger J. A. Csinos A. S. 2006 Root-knot nematode management in double-cropped plasticulture vegetables Journal of Nematology 38 59 67 19259431 Desaeger J. A. Seebold K. W. Csinos A. S. 2008 Effect of application timing and method on efficacy and phytotoxicity of 1, 3-D, chloropicrin and metam-sodium combinations in squash plasticulture Pest Management Science (formerly Pesticide Science) 64 230 238 Desaeger J. Dickson D. W. Locascio S. J. 2017 Methyl bromide alternatives for control of root-knot nematode (Meloidogyne spp.) in tomato production in Florida Journal of Nematology 49 140 149 28706313 Fiscus D. A. Neher D. A. 2002 Distinguishing sensitivity of free-living soil nematode genera to physical and chemical disturbances Ecological Applications 565 575 Gerstl Z. Mingelgrin U. Yaron B. 1977 Behavior of Vapam and methylisothiocyanate in soils Soil Science Society of America Journal 41 545 548 Gilreath J. P. Santos B. M. Gilreath P. R. Jones J. P. Noling J. W. 2004 Efficacy of 1,3-dichloropropene plus chloropicrin application methods in combination with pebulate and napropamide in tomato Crop Protection 23 1187 1191 Gilreath J. P. Santos B. M. Motis T. N. 2008 Performance of methyl bromide alternatives in strawberry HortTechnology 18 80 83 Grabau Z. J. Mauldin M. D. Habteweld A. Carter E. T. 2020 Nematicide efficacy at managing Meloidogyne arenaria and non-target effects on free-living nematodes in peanut production Journal of Nematology 52 2020 Grabau Z. J. Noling J. W. Navia Gine P. A. 2019 Fluensulfone and 1,3-dichloroprene for plant-parasitic nematode management in potato production Journal of Nematology 51 e2019 38 Grabau Z. J. Liu C. Sandoval-Ruiz R. 2021 Management by nematicides in tomato production Journal of Nematology 53 1 12 34179819 Holajjer P. Kamra A. Singh P. A. 2016 Influence of nematode-bacterial interactions on N and P mineralisation in soil and on decomposition of crop residues during aerobic composting Applied Ecology and Environmental Research 14 283 299 Ingham R. E. Hamm P. B. Baune M. David N. L. Wade N. M. 2007 Control of Meloidogyne chitwoodi in potato with shank-injected metam sodium and other nematicides Journal of Nematology 39 161 19259485 Jenkins W. R. 1964 A rapid centrifugal-flotation technique for separating nematodes from soil Plant Disease Reporter 48 692 Kearn J. Lilley C. Urwin P. O’Connor V. Holden-Dye L. 2017 Progressive metabolic impairment underlies the novel nematicidal action of fluensulfone on the potato cyst nematode Globodera pallida Pesticide Biochemistry and Physiology 142 83 90 29107251 Khan Z. Kim Y. H. 2005 The predatory nematode, Mononchoides fortidens (Nematoda: Diplogasterida), suppresses the root-knot nematode, Meloidogyne arenaria, in potted field soil Biological Control 35 78 82 Klose S. Ajwa H. A. Browne G. T. Subbarao K. V. Martin F. N. Fennimore S. A. Westerdahl B. B. 2008 Dose response of weed seeds, plant-parasitic nematodes, and pathogens to twelve rates of metam sodium in a California soil Plant Disease 92 1537 1546 30764440 Levene H. 1960 Robust tests for equality of variances 278 292 Olkin I. Contributions to probability and statistics Palo Alto, California Stanford University Press Ma Q. L. Gan J. Papiernik S. K. Becker J. O. Yates S. R. 2001 Degradation of soil fumigants as affected by initial concentration and temperature Journal of Environmental Quality 30 1278 1286 11476506 Navia Gine P. A. Hajihassani A. 2021 Evaluation of fluensulfone in combination with metam potassium as a broad-spectrum solution to control plant-parasitic nematodes, soil-borne diseases and weeds in fruiting vegetables Journal of Nematology 53 e2021 95 https://doi.org/10.21307%2Fjofnem-2021-095. (Abstr.) Sanchez-Moreno S. Jimenez L. Alonso-Prados J. L. Garcia-Baudin J. M. 2010 Nematodes as indicators of fumigant effects on soil food webs in strawberry crops in southern Spain Ecological Indicators 10 148 156 Santos B. M. Gilreath J. P. Motis T. N. Noling J. W. Jones J. P. Norton J. A. 2006 Comparing methyl bromide alternatives for soilborne disease, nematode and weed management in fresh market tomato Crop Protection 25 690 695 Timper P. Davis R. Jagdale G. Herbert J. 2012 Resiliency of a nematode community and suppressive service to tillage and nematicide application Applied Soil Ecology 59 48 59 Trap J. Bonkowski M. Plassard C. Villenave C. Blanchart E. 2016 Ecological importance of soil bacterivores for ecosystem functions Plant and Soil 398 1 24 Waldo B. D. Grabau Z. J. Mengistu T. M. Crow W. T. 2019 Nematicide effects on non-target nematodes in bermudagrass Journal of Nematology 51 e2019 09 Watson T. T. Desaeger J. A. 2019 Evaluation of non-fumigant chemical and biological nematicides for strawberry production in Florida Crop Protection 117 100 107 Weingartner D. P. Shumaker J. R. 1990 Control of nematodes and soil-borne diseases in Florida potatoes with aldicarb and 1,3-D Journal of Nematology 22 775 778 19287795 Triky-Dotan S. Austerweil M. Steiner B. Peretz-Alon Y. Katan J. Gamliel A. 2007 Generation and dissipation of methyl isothiocyanate in soils following metam sodium fumigation: Impact on Verticillium control and potato yield Plant Disease 91 497 503 30780692 USDA-AMS 2011 United States standards for grades of potatoes USDA Washington, DC USDA-NASS 2022 Potatoes 2021 summary (September 2022) USDA Washington DC USDA-NRCS 2005 Estimating soil moisture by feel and appearance. PA-1619 USDA Washington DC

PMC010xxxxxx/PMC10327230.txt

==== Front ArXiv ArXiv arxiv ArXiv 2331-8422 Cornell University arXiv:2306.15018v1 2306.15018 1 preprint Article Possible depth-resolved reconstruction of shear moduli in the cornea following collagen crosslinking (CXL) with optical coherence tomography and elastography Regnault Gabriel 1* Kirby Mitchell A. 1 Wang Ruikang K. 12 Shen Tueng T. 23 O’Donnell Matthew 1 Pelivanov Ivan 1 1 Department of Bioengineering, University of Washington, Seattle, USA. 2 Department of Ophthalmology, University of Washington, Seattle, USA. 3 School of Medicine, University of Washington, Seattle, USA. * gregnaul@uw.edu 26 6 2023 arXiv:2306.15018v1https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-2306.15018v1.pdf Corneal collagen crosslinking (CXL) is commonly used to prevent or treat keratoconus. Although changes in corneal stiffness induced by CXL surgery can be monitored with non-contact dynamic optical coherence elastography (OCE) by tracking mechanical wave propagation, depth dependent changes are still unclear if the cornea is not crosslinked through the whole depth. Here, phase-decorrelation measurements on optical coherence tomography (OCT) structural images are combined with acoustic micro-tapping (AμT) OCE to explore possible reconstruction of depth-dependent stiffness within crosslinked corneas in an ex vivo human cornea sample. Experimental OCT images are analyzed to define the penetration depth of CXL into the cornea. In a representative ex vivo human cornea sample, crosslinking depth varied from ~ 100 μm in the periphery to ~ 150 μm in the cornea center and exhibited a sharp in-depth transition between crosslinked and untreated areas. This information was used in an analytical two-layer guided wave propagation model to quantify the stiffness of the treated layer. We also discuss how the elastic moduli of partially CXL-treated cornea layers reflect the effective engineering stiffness of the entire cornea to properly quantify corneal deformation. NIH Grants R01-EY026532, R01-EY028753, and R01-AR077560 and the Department of Bioengineering at the University of Washington ==== Body pmc1. Introduction The cornea is the primary optical element focusing light onto the retina. It contains multiple layers, including epithelium and stroma (Figs. 1a, b). The first acts as a barrier against the external environment and the latter maintains stiffness, transparency and focusing power [1,2]. The microstructure of the stroma is composed of collagen fibrils, arranged in lamellae, lying within a protein rich, hydrated proteoglycan mesh [3,4] (Figs. 1b, c). Corneal diseases (such as keratoconus (KC)) and surgical complications from refractive surgeries (such as Laser-Assisted In Situ Keratomileusis (LASIK)) may deform the cornea (ectasia) and alter vision. The prevalence of KC in the general population is estimated to be 1.38 per 1000 [5], and nearly 1 million refractive surgeries are performed each year in the USA. Despite their overall success, however, suboptimal visual outcomes and post-refractive corneal decompensation cannot always be predicted for an individual patient. Corneal collagen crosslinking is a minimally invasive procedure that can potentially slow the progression of corneal ectasia [6–9]. Ultraviolet (UV) light modifies the microstructure of the cornea soaked in riboflavin and forms additional chemical bonds between collagen fibers in the stroma [10] (Fig. 1b). Post-treatment corneas become stiffer and more resistant to enzymatic digestion [11–13]. Although corneal topography (curvature) and thickness maps can be obtained preoperatively, and refractive corrections can be estimated, there is an unmet need to predict corneal decompensation from interventions such as LASIK and CXL therapies. Unfortunately, surgical planning cannot be customized and outcomes (e.g., postoperative corneal ectasia risks) cannot be accurately predicted without quantitatively mapping corneal elasticity. Thus, methods to quantitatively reconstruct corneal elastic moduli are needed. Ocular response analyzer (ORA - Reichert Technologies) and Dynamic Scheimpflug Analyzer (DSA - Corvis ST – Oculus Opitkgerate GmbH) are the state-of-the-art in clinical measurements of corneal mechanics. They estimate stiffness as the pressure at inward applanation divided by corneal displacement [14–16]. However, measurements induce large corneal deformations that are often clinically unacceptable, require a non-trivial IOP correction in simulations [17] and assume a simple isotropic mechanical model leading to high variability with experimental conditions. Results obtained with the Corvis ST on KC may be contradictory, and some even show no significant change in corneal stiffness pre- and post-CXL surgery [18,19]. In addition, the result is averaged over the entire cornea with no spatial resolution, and the reconstruction is questionable if corneal thickness varies. Dynamic elastography is a promising tool to probe soft tissue biomechanics. A shear wave can be launched using direct contact excitation [20–23] or radiation force-based techniques [24,25]. By tracking shear wave propagation, either using magnetic resonance imaging (MRI) [26,27] ultrasound [20–22,28] or dynamic phase-sensitive OCT [29–31], one can infer, with an appropriate mechanical description, the linear [32–35], or non-linear [36,37] stiffness moduli of the tissue. Optical coherence elastography (OCE) is particularly suited to probe corneal biomechanics non-invasively in a clinical environment [24,30,38–40], as it can be combined with non-contact excitation techniques (for example, using an air-puff or non-contact acoustic micro-tapping (AμT [24])). Because cornea is thin and bounded between air and aqueous humor, wave propagation within it is guided, leading to strong geometric dispersion [33,41]. As such, the common approach associating the Rayleigh surface wave group velocity to stiffness [24,30,38–40,42–45] is not appropriate. In addition, the stroma contains collagen lamellae running in-plane across its width. Lamellae make up approximately 90% of tissue thickness and account for most of the cornea’s mechanical structure. They are stacked vertically in approximately 200–500 separate planes [46,47], suggesting an anisotropic mechanical behavior with very different responses to in-plane versus out-of-plane loads (Fig. 1c). To account for this specific architecture we introduced a model of a nearly-incompressible transverse isotropic (NITI) medium [33], in which corneal stiffness is defined by two (in-, μ, and out-of-plane, G) shear moduli, decoupling tensile/inflation properties from shear responses. Based on this model, we developed an algorithm utilizing guided mechanical waves in a bounded NITI medium to reconstruct both moduli from AμT-OCE. The model was confirmed ex vivo in rabbit [48], porcine [49] and human [32,33] models and in vivo with rabbit models [48]. For ex vivo human corneas, we evidenced that both in- and out-of-plane post-CXL corneal shear moduli experienced an averaged two-fold increase in Young’s modulus and an almost four-fold increase for the out-of-plane shear modulus G [32], assuming the whole thickness was treated. That confirmed that CXL increases inter-corneal lamellae crosslinks but less affects corneal deformational properties, defined by the Young’s modulus which increases less and has more implications for potential refractive changes. Despite the success in quantifying CXL-induced corneal elastic properties, there is a common situation where the method [32] must be refined. For example, crosslinking is often inhomogeneous in depth due to more pronounced riboflavin penetration in the corneal anterior where the solution is applied [50], and because UVA irradiation attenuates and is less effective as it propagates through cornea [51]. Together, this generally produces a clear demarcation line between treated and untreated regions [52], suggesting a two-layer structure postoperatively. As noted above, mechanical waves generated in the cornea are typically guided. They occupy the entire thickness of the cornea and, therefore, carry depth-accumulated information. Thus, reconstructing the depth dependence of corneal moduli is difficult with wave based OCE alone without a good estimate of the CXL penetration depth into the cornea. Blackburn et al. [53] recently introduced a novel metric to track CXL penetration within the cornea using time-resolved OCT. They demonstrated that the phase decorrelation decay rate of the complex OCT signal is reduced in the CXL area, which can be used to distinguish treated from untreated areas postoperatively. In this paper, we combine the method described in [53] with AμT-OCE measurements to explore possible reconstruction of both in- and out-of-plane corneal elastic moduli over depth in a partially CXL-treated ex vivo human cornea. We developed an analytical model of guided wave propagation accounting for multiple layers, each with distinct stiffness moduli and thickness, to properly account for CXL-induced corneal layering. Baseline elastic moduli in an untreated ex vivo cornea can be determined from AμT-OCE prior to CXL, so that induced changes in the treated anterior layer can be quantified by fitting the post-CXL wave dispersion dependence in the frequency-wavenumber (f-k) domain. We also discuss how the depth-distribution of stiffness affects the effective engineering stiffness of the entire cornea and show that assessing stiffness in both layers is needed to properly predict corneal deformation and quantify surgical outcomes. 2. Method 2.1 Cornea preparation A corneal-scleral ring stored in Optisol (Chiron Ophtalmics) was obtained from CorneaGen. It came from a 26 year-old donor and was stored for less than 30 days. The corneal-scleral button was mounted on an artificial anterior chamber (Barron, CorzaMedical; see Fig. 2), connected through the inlet port to an elevated bath filled with balanced saline solution (BSS) to apply a controlled pressure mimicking intraocular pressure (IOP) on the anterior segment of the cornea. The outlet port remained closed to allow the IOP to settle at 15 mmHg within the chamber, corresponding to human physiological conditions [54]. CXL followed the Dresden protocol [6]. First, the epithelial membrane was removed. Then, the cornea was soaked in riboflavin for 30 minutes by applying a 50 μL drop of 0.1% riboflavin in 20% dextran solution every two minutes. It was then exposed to 3 mW/cm2 of 370 nm UV light for 30 minutes, while a drop was re-applied every 5 minutes. 2.2 AμT-OCE imaging system A cylindrically focused air-coupled (AμT) transducer, operating at a 1 MHz frequency, launched mechanical waves in the cornea. A spectral domain OCT system with a 46.5 kHz A-scan rate was used to track wave propagation and structural changes [24,30,32,33]. The cylindrical focus generated quasi-planar guided waves within the cornea. The OCT system operated in M-B mode, where a single push was triggered by the system while 512 consecutive A-scans were taken at a fixed location (M-scan). The M-scan sequence and push excitation were repeated for 256 locations, creating a volume with 256 x-samples, 1024 z-samples and 512 t-samples (see Fig. 3a)), with an effective imaging range of 6 mm × 1.2 mm×11 ms. The particle velocity along the probe beam direction was obtained from the optical phase difference between two consecutive A-lines at each location [55]. The spatio-temporal (x-t) surface signature of the guided wave was computed from an exponentially weighted-average of the particle velocity over the first 180 μm. As shown in Fig. 3b, the guided wave only propagated during the first 4 ms of the scans, which was used to determine the stiffness of the material by fitting the computed dispersion curve in the frequency-wavenumber domain (f-k) obtained from the 2D Fourier spectrum (Fig. 3c). This procedure is detailed in Section 2.4. On the other hand, data from the last 7 ms were used to study structural changes with phase decorrelation [53] (see Section 2.6). 2.3 Multi-layer NITI model Like most biological tissue, the cornea is nearly incompressible. In addition, its structure implies a transverse isotropy [49] and, therefore, its mechanical behavior under small deformation should be described with a NITI model [33]. In Voigt notation, Hook’s law of stress and strain for a NITI material takes the form: (1) [σxxσyyσzzτyzτxzτxy]=[λ+2μλλλλ+2μλλλλ+δGGμ][ϵxxϵyyϵzzγyzγxzγxy], where σij denotes engineering stress, ϵij denotes engineering strain, τij denotes shear stress, γij = 2 ϵij denotes shear strains, the subscripts x, y and z refer to the Cartesian axes and G, μ, λ and δ are four independent elastic constants. In previous studies [34,49], we have demonstrated that δ, which accounts for tissue tensile anisotropy, cannot be determined from the propagation of vertically polarized guided waves generated in the cornea. At the same time, the influence of δ on the in-plane Young’s modulus, ET, is minor so that it is restricted to the range of 2μ ≤ ET ≤ 3μ. Consequently, here, corneal tensile isotropy (δ = 0) is also assumed (ET = E ≅ 3μ). Because the cornea is a nearly incompressible soft tissue, its Young’s modulus does not depend on λ. Therefore, among the four elastic constants, only G and μ (respectively the out-of-plane and in-plane shear moduli) are needed to predict corneal deformation under mechanical loading. The effects of CXL on the cornea depend on depth. Several recent studies showed that postoperative CXL corneas might experience non-uniform crosslinkage with depth. The transition between crosslinked (anterior) and non-crosslinked (posterior) parts tends to be sharp rather than smooth [51,52]. This effect is also observed in our experiments (see below). Thus, a two-layer structure is considered an appropriate model to quantify postoperative corneas. CXL was also shown to change collagen fiber diameter and interfibrillar spacing [50], but nothing suggests a modification of its macroscopic anisotropic organization. Based on this observation, we developed a multi-layer model to predict wave propagation within CXL corneas (see Supplementary 1) that accounts for any arbitrary number of layers, each with a stress-strain relationship given by Eq. (1) and linked by solid-solid boundary conditions (continuity of normal components of stress and displacement across every interface). Accounting for the external boundary conditions (liquid below and air above the cornea) and the finite thickness of the medium, the dispersion relation for guided waves can be determined directly from stiffness moduli Gn and μn and the thickness ℎn of each layer. A more detailed description of the multi-layer model is in Supplementary 1. Although only 2 layers were considered here, the multi-layer model can be used to accurately capture more complicated transitions between crosslinked and non-crosslinked areas. In an untreated cornea, only the first anti-symmetric mode, referred to as A0, typically propagates in the range of frequencies that can be recorded in OCE (usually <5 kHz). Because a partial-CXL cornea consists, in our approximation, of two horizontally assembled layers, each having a vertically aligned symmetry axis, this symmetry holds for the global material. Thus, only the A0-mode is also expected in the partially crosslinked cornea. 2.4 Fitting f-k spectra pre- and post-CXL Prior to treatment, the cornea was assumed homogeneous, which in our model corresponds to a single NITI layer bounded above by air and below by water. The experimental f-k spectrum (see Fig. 3c) was obtained by computing the 2D FFT of the x-t plot. Shear moduli G and μ in pre-CXL cornea were obtained by fitting the measured f-k spectrum with the analytic dispersion relation for the A0-mode [32,33,48,49]. In the CXL-treated cornea, the thickness of both layers can be measured (see Section 3.1), and the posterior layer is assumed to still possess the original (i.e., untreated) elastic properties. Thus, the 2-layer model with known elastic moduli of the bottom (untreated) layer can be used to determine the stiffness of the top layer. To ensure reliable fitting for all cases, we computed a goodness of fit (GOF) metric Φ=Σfχfit(f)Σfχmax(f), where χfit(f) corresponds to the energy of the 2D spectrum covered by the best analytical solution (one or two layers) at a given frequency f and χmax(f) is the unconstrained maximal energy of the spectrum at frequency f. Based on recent results (see Supplemental Material in [32]), reliable fitting in human ex vivo corneas is associated with values of ϕ > 0.9. An example of a 2D-spectrum and the fitted A0-mode obtained with this procedure for the untreated case is shown in Fig. 3c. 2.5 FEM simulations We designed finite element (FEM) simulations in OnScale to determine the accuracy of our multi-layer NITI model in reconstructing stiffness along corneal depth. The geometry is shown in Fig. 4a. Corneal boundary conditions were replicated, with the material bounded above by air and below by water. The speed of sound in all layers (material and water) was fixed to avoid the reflection of compressional waves at the air-liquid and inter-layer boundaries. It also improved the absorption of compressional waves at the absorbing boundaries and, thus, avoided divergent simulations. The simulated transient excitation of broadband elastic waves closely matched that used in AμT experiments. More details about the simulations can be found in [33]. Based on phase-decorrelation measurements (see Section 2.6), we assumed that post-CXL two layers with distinct thicknesses were formed within the cornea, the top layer being stiffer than the bottom one. Stiffness values assessed from experiments were also used in the simulations. We used the top surface vertical particle velocity of the simulated wave (see Fig. 4b), and its associated f-k spectrum (see Fig. 4c), to show that the analytical solution obtained from the N-layer model (see Eq. S39 in Supplementary 1) closely matches the f-k spectrum of numerically simulated wave propagation. This confirms the accuracy of the analytical model in quantifying measurements of corneal elastic moduli and their variation with depth. 2.6 Phase Decorrelation OCT (PhD-OCT) Blackburn et al. [53] have recently introduced a metric to track CXL penetration within the cornea using time-resolved OCT. It was shown that the phase decorrelation decay rate of the complex OCT signal is reduced in CXL areas and can distinguish treated from untreated areas after the procedure. In our study, the autocorrelation function of the signal g(τ) was computed over 15 consecutive samples at 46,500 Hz for six consecutive pixels within a given A-line: (2) g(τ)=〈〈E(t)E*(t+τ)〉pixels 〈E(t)E*(t)〉pixels ×〈E(t+τ)E*(t+τ)〉pixels 〉, which is expected to follow an exponential decay [56]: (3) g(τ)=e−Γ⋅τ≈1−Γ⋅τ, where Γ is the decorrelation coefficient that is inversely proportional to the Brownian diffusion coefficient [56], meaning that the more coherent the material, the smaller the decorrelation coefficient. The procedure was performed starting at n, n + 1, n + 2, … A-lines, where n is the first time-sample used for phase-decorrelation (t(n) = 4 ms). The decorrelation coefficient Γ was then computed using the averaged g(τ) over the number of starting points by fitting with a first order polynomial (see Fig. 3e): 〈g(τ)〉 = b − Γ · τ, where 〈 〉 denotes the average over the number of starting points. In crosslinked regions of the cornea (anterior), tissue stiffens, implying that Γ should be smaller than in the untreated region (posterior). For post-processing, we rejected all fits with b < 0.95, corresponding in general to peripheral regions where the signal to noise ratio (SNR) was too low. 3. Results 3.1 Thickness of CXL layer The spatial distribution of the OCT intensity signal (Figs. 5a, b, e) and phase decorrelation images (Figs. 5c, d, e) both show clear layering in the treated cornea. The effect of CXL is not homogeneous across the cornea, with a more pronounced effect at the center (~150 μm) than at the periphery (~ 100 μm). For the present case, we estimated that about 30% of the cornea was treated effectively. The treated cornea is thinner than that prior to CXL (its thickness reduced from 575 μm to 520 μm), as generally observed in the literature [57,58]. 3.2 Stiffness of CXL-treated corneal layers AμT-OCE scans, taking approximatively 3 s to acquire and save data, were acquired prior and post-CXL. A space-time (x-t) plot of the vertically polarized particle velocity (see Figs. 3a, b) in the untreated cornea was used to compute the f-k spectrum (see Fig. 3c), which was then fitted using the procedure detailed in Section 2.4, assuming the cornea as a single homogeneous layer. The fitting routine was also detailed in our recent work [32,48]. Results for the reconstructed in- (μ) and out-of-plane (G) corneal shear moduli pre-CXL are μ = 7.6 ∓ (7,13) MPa and G = 59.5 ∓ (5,8) kPa (see Table 1). To reconstruct depth-dependent stiffness moduli after CXL, it is assumed that: i) the thickness of both the anterior and posterior layers can be measured using dynamic OCT from phase-decorrelation and/or intensity variation methods (see Fig. 5, estimated to be ℎant = 150 μm and ℎpos = 370 μm with both methods); ii) the stiffness of the posterior layer remains unchanged after CXL; iii) the effect of CXL is homogeneous in the anterior layer. Fixing known parameters (untreated cornea thicknesses and posterior stiffness moduli), stiffness moduli of the anterior cornea layer can be determined by fitting the wave dispersion curve in the f-k domain with the 2-layer model (Fig. 6b). We found an increase in both stiffness moduli Gant = 296.8 ∓ (60, 92) kPa and μant = 34.6 ∓ (20,23) MPa compared to those for the posterior region Gpos = 59.5 ∓ (5, 8) kPa and μpost = 7.6 ∓ (7, 13) MPa (see Figs. 6c, d). The goodness of fit for the 2-layer model (Φ = 0.950) remained within the range of reliable fitting. The results are summarized in Table 1. GOF variation for a given sample at a fixed IOP was previously shown to be about 1% [32]. We used this fact to build error bars for the present study, illustrated in Fig. 6d for the two-layer fitting procedure. First, we fitted the projection of the goodness of fit surface for a value of Φ = 0.99 × max (Φ) (i.e., 1% below the optimal GOF), with an ellipsoidal function that best described the shape of the iso-goodness levels. Then, we computed the uncertainty intervals as the intersection of the horizontal and vertical direction with the fitted ellipse. This produces asymmetric uncertainty intervals (particularly for μ), reflecting the asymmetric variation of Φ. 3.3 Mixing rules for the effective engineering moduli of layered materials A theory for effective moduli of multi-layer materials was developed in the early 1970’s for composite materials. It is now accepted in material science and broadly used in the development of composite structures [59–61]. The derivation of these ‘effective’ material engineering moduli is based on ‘mixing rules’ of stiffness moduli across depth using the following assumptions: i) out-of-plane stresses and in-plane strains are uniform across thickness; ii) in-plane stresses and out-of-plane strains are averaged across thickness based on layer volume fractions. Note that the solution is valid only for low-frequency material deformation, which are, fortunately, usually related to physiologically induced stresses. As such, even if these definitions do not hold for perturbations of any kind, they are appropriate for in vivo corneal response to physiological loads. Sun et al. [60] have demonstrated that an effective out-of-plane engineering modulus Geff can be computed using the inverse mixture rule for out-of-plane material constants of individual material layers: (4) Geff=(∑nhn/hGn)−1, where ℎ is a total material thickness, ℎn is a thickness of the nth layer and Gn is an out-of-plane modulus of this layer. On the other hand, an effective in-plane engineering modulus μeff can be obtained from the mixture rule: (5) μeff=∑nμn⋅hnh, where μn is an in-plane modulus of the nth layer. Based on the mixture rules described above, the effective low-frequency engineering moduli of a partially treated cornea for our case are computed to be Geff = 77.3 ∓ (6,10) kPa and μeff = 15.4 ∓ (8,11) MPa (see Table 1). 4. Discussion and conclusions In this study, we combined structural OCT with dynamic AμT-OCE to assess the penetration depth of CXL in the cornea. Analyzing the brightness of structural OCT images and the rate of image decorrelation between consecutive B-scans, we confirm a sharp transition between CXL and untreated cornea layers and measure the thicknesses of treated and untreated layers. This finding suggests a 2-layer medium model for the treated cornea that can be used to reconstruct both in- and out-of-plane elastic moduli in the treated layer. As we discussed in detail in our previous work [32,48], the A0-mode dispersion spectrum is much more sensitive to variations of out-of-plane modulus G rather than to variations of in-plane modulus μ. This results in large error bars in the reconstruction of μ (see Table 1) compared to those for G. However, here we used only a single measurement, and reconstruction accuracy can be improved by repeating AμT measurements. Note that the error bar asymmetry comes from the asymmetry of the GOF function (see Figs. 6 c, d)), which was previously described in Refs [32,48]. Using AμT-OCE, we tracked guided wave propagation in the cornea before and after CXL. Using the NITI model in each layer, we quantified depth-localized corneal stiffening with CXL. Because the A0-mode occupies the whole thickness, it carries averaged information about material stiffness (i.e., averaged over its two parts). We determined the effective moduli in the treated corneal layer by fitting the 2D-spectrum with the two-layer model using known thicknesses for each layer and known elastic moduli for the untreated part (obtained from OCE measurements pre-CXL). Effective engineering moduli of the entire cornea can be then calculated using Eqs. (4) and (5), respectively, for out-of- and in-plane moduli. As explained in Section 3.3, elastic moduli determined for the anterior (CXL-treated) corneal layer can be used to compute effective corneal engineering moduli, where μeff uses a simple mixture rule whereas Geff requires an inverse mixture rule. This implies that even if the treated layer G experience an almost six-fold increase, its effective increase for the whole tissue is only by a factor of 1.3. On the other hand, μ experiences a local five-fold increase in the anterior layer but the overall in-plane modulus, which is more directly related to deformations in response to physiological loads, increases by a factor of 2. Because the mixture rules for engineering effective moduli (Eqs. (4), (5)) assume only low-frequency perturbations, it is interesting to check if they could also describe guided wave behavior in the partially crosslinked cornea when considered as an effective homogeneous single-layer material (see Supplementary 2). We found that the effective engineering mixture rules cannot be applied to quantify ‘effective’ guided wave behavior in the layered medium and, more importantly, the guided wave behavior considered in the ‘effective’ single layer incorrectly describes effective corneal engineering moduli. Fitting the experimentally obtained f-k spectrum (Fig. 6b) with the single layer model (Supplementary Fig. S1) results in a different (incorrect) set of reconstructed engineering corneal moduli (see Table 2 below). Thus, measuring the depth of CXL penetration into cornea and implementing the multi-layer guided wave model are both required to accurately assess post-CXL corneal mechanical moduli. Although the relationship between acoustic bulk (longitudinal and shear) wave propagation speeds and mechanical moduli in multi-layer or multi-component media has been reported in several studies [62,63], these relationships have not been explored for guided waves due to their high geometric dispersion. The situation is even more complicated for anisotropic media. The lack of the complete solution to this problem does not affect the goal of this study and is definitely outside its scope. However, we would like to share an interesting observation. We empirically found that both the analytic model and ex vivo experiments in the human cornea sample suggest that a simple mixture rule (Eq. (5)) for both corneal moduli can be applied to approximately compute effective guided wave propagation in a multi-layer NITI medium, as detailed in Supplementary 2. One of the limitations of this work is that it assumes that post-CXL cornea contains two homogeneous layers, a reasonable assumption given several previous studies [50–52]. However, when a more gradual transition between CXL-corneal layers is observed, the multi-layer model introduced here can be used to compute A0-mode dispersion for more sophisticated models of CXL (e.g., accounting for a gradient in stiffness or more complex structural changes). Recent results suggest that reverberant OCE can reconstruct depth-dependent stiffness variations [23]. It would be interesting to compare it with our method in future studies. Note, however, that reverberant OCE is not currently feasible in vivo because it uses contact vibrators. This is why guided wave-based OCE is still the only method capable of in vivo non-contact measurements of corneal anisotropic elasticity, ultimately with sub-mm lateral resolution [64]. Finally, we have shown that phase-sensitive OCT combined with AμT wave excitation can assess both the structure of human cornea and the depth-dependence of moduli due to CXL. These findings are essential for building personalized models of corneal deformation following CXL and, thus, better adapt crosslinking therapy for clinical use and predict its outcomes. Further experiments on a larger group of cornea samples are required to generalize the present results. Supplementary Material 1 Acknowledgments. TODO Funding. NIH Grants R01-EY026532, R01-EY028753, and R01-AR077560 and the Department of Bioengineering at the University of Washington Data availability. Supplemental document. See Supplement 1 and 2 for supporting content and Supplement 3 for Matlab scripts of the N-layer model. Fig. 1 Illustration of the anterior section of the human eye (a), the change in corneal geometry induced by CXL (b), and three-dimensional organization of corneal lamellae (slow-axis NITI symmetry, i.e., symmetry across the lamellae) within the stroma (c). Fig. 2. Picture of the experimental set up during UV-CXL. a) Acoustic micro-tapping transducer. b) Artificial anterior chamber with c) inlet port connected to the elevated bath and outlet port closed to control IOP. Fig. 3. Diagram of spectral domain time-resolved OCT and AμT-OCE measurements in a pre-CXL cornea. a) 3D (x, z and t), wavefield after AμT excitation. The top surface wavefield of the initial time sequence (black dotted region) is used for elastic moduli reconstruction, and data at the end of the sequence are used for phase decorrelation measurements. b) x-t plot showing the top surface signature of the guided mode during the first 5 ms of the acquisition sequence. c) f-k spectrum obtained by 2D-FFT of the x-t-plot showing the dispersion dependence of the first anti-symmetric mode A0. The red curve indicates the best fit obtained with the NITI model [33]. d) Structural OCT image obtained by averaging the last 7 ms of the raw OCT signal. e) Phase decorrelation function g(τ) at the location indicated by the red square in d). Fig. 4. Finite element simulations to study the effects of a layered structure for post-CXL cornea. a) Geometry of the two-layered material used in simulations, bounded above by air and below by water. b) Top surface spatio-temporal signature (x-t plot) of the guided wave for a two-layer case with Gant = 296.8 kPa, μant = 34.6 MPa, Gpos = 59.5 kPa, μpos = 7.3 MPa, ℎant = 150 μm and ℎpos = 370 μm. c) 2D Fourier spectrum of the wave studied in b) showing the main propagating A0-mode and, in red, the analytical solution obtained from the multi-layer NITI model with identical parameters. Fig. 5. Short time decorrelation before and after CXL. Structural OCT images averaged over the last 7ms of the OCT scan for a) pre-CXL and b) post-CXL. Maps of decorrelation coefficient Γ for c) pre- and d) post-CXL. e) Profile of OCT intensity and Γ along the red dotted line shown in b) and d) for the CXL cornea. Fig. 6. Post-CXL fitting. a) Measured vertically polarized top-surface signature of the guided wave in the treated cornea. b) 2D-spectrum computed from a). c) 2D Goodness of Fit surface when the fit is performed with the 2-layer model to determine the anterior layer stiffness moduli. d) Projection of the surface plot for Φ = 0.99 × max (Φ). Error bars are computed as the intersection of the vertical and horizontal directions with the fitted ellipse. The global maximum of Φ is indicated in c) and d) by the circular marker. Table 1. Corneal thickness (h), in- (μ) and out-of-plane (G) elastic moduli and goodness of fit (Φ) pre- and post-CXL h (μm) G (kPa) μ (MPa) ϕ Pre-CXL Fitted moduli 575 59.5 ∓ (5,8) 7.6 ∓ (7,13) 0.961 Post-CXL Top layer 150 296.8 ∓ (60, 92) 34.6 ∓ (20, 23) 0.95 Bottom layer 370 59.5 ∓ (5,8) 7.6 ∓ (7,13) - Effective engineering moduli, Eqs. (4), (5) 520 77.3 ∓ (6,10) 15.4 ∓ (8,11) - Table 2. Effective moduli measured with the single or multilayer models. G (kPa) μ (MPa) 2-layer model, Eqs. (4), (5) 77.3 ∓ (6,10) 15.4 ∓ (8,11) Single layer model 127.5 ∓ (12,17) 9.3 ∓ (8,18) Disclosures. The authors declare no conflicts of interest. ==== Refs References 1. Sridhar M. , “Anatomy of cornea and ocular surface,” Indian J. Ophthalmol. 66 (2 ), 190 (2018).29380756 2. Meek K. M. and Boote C. , “The organization of collagen in the corneal stroma,” Experimental Eye Research 78 (3 ), 503–512 (2004).15106929 3. Borcherding M. S. , Blacik L. J. , Sittig R. A. , Bizzell J. W. , Breen M. , and Weinstein H. G. , “Proteoglycans and collagen fibre organization in human corneoscleral tissue,” Experimental Eye Research 21 (1 ), 59–70 (1975).124659 4. Meek K. M. and Knupp C. , “Corneal structure and transparency,” Progress in Retinal and Eye Research 49 , 1–16 (2015).26145225 5. Hashemi H. , Heydarian S. , Hooshmand E. , Saatchi M. , Yekta A. , Aghamirsalim M. , Valadkhan M. , Mortazavi M. , Hashemi A. , and Khabazkhoob M. , “The Prevalence and Risk Factors for Keratoconus: A Systematic Review and Meta-Analysis,” Cornea 39 (2 ), 263–270 (2020).31498247 6. Wollensak G. , Spoerl E. , and Seiler T. , “Riboflavin/ultraviolet-a–induced collagen crosslinking for the treatment of keratoconus,” American Journal of Ophthalmology 135 (5 ), 620–627 (2003).12719068 7. Hersh P. S. , Stulting R. D. , Muller D. , Durrie D. S. , Rajpal R. K. , Binder P. S. , Donnenfeld E. D. , Durrie D. , Hardten D. , Hersh P. , Price F. , Schanzlin D. , Stark W. , Stulting R. D. , Trattler W. , and Trokel S. , “United States Multicenter Clinical Trial of Corneal Collagen Crosslinking for Keratoconus Treatment,” Ophthalmology 124 (9 ), 1259–1270 (2017).28495149 8. Goldich Y. , Marcovich A. L. , Barkana Y. , Mandel Y. , Hirsh A. , Morad Y. , Avni I. , and Zadok D. , “Clinical and Corneal Biomechanical Changes After Collagen Cross-Linking With Riboflavin and UV Irradiation in Patients With Progressive Keratoconus: Results After 2 Years of Follow-up,” Cornea 31 (6 ), 609–614 (2012).22378112 9. Wollensak G. and Iomdina E. , “Long-term biomechanical properties of rabbit cornea after photodynamic collagen crosslinking,” Acta Ophthalmologica 87 (1 ), 48–51 (2009).18547280 10. Maier P. , Reinhard T. , and Kohlhaas M. , “Corneal Collagen Cross-Linking in the Stabilization of Keratoconus,” Dtsch Arztebl Int 116 (11 ), 184–190 (2019).31014449 11. Wollensak G. , Spoerl E. , and Seiler T. , “Stress-strain measurements of human and porcine corneas after riboflavin–ultraviolet-A-induced cross-linking,” Journal of Cataract and Refractive Surgery 29 (9 ), 1780–1785 (2003).14522301 12. Chang S.-H. , Zhou D. , Eliasy A. , Li Y.-C. , and Elsheikh A. , “Experimental evaluation of stiffening effect induced by UVA/Riboflavin corneal cross-linking using intact porcine eye globes,” PLoS ONE 15 (11 ), e0240724 (2020).33147249 13. Beshtawi I. M. , O’Donnell C. , and Radhakrishnan H. , “Biomechanical properties of corneal tissue after ultraviolet-A–riboflavin crosslinking,” Journal of Cataract and Refractive Surgery 39 (3 ), 451–462 (2013).23506922 14. Kaushik S. and Pandav S. S. , “Ocular Response Analyzer,” Journal of Current Glaucoma Practice 6 (1 ), 17–19 (2012).27990067 15. Dupps W. J. and Roberts C. J. , “Corneal biomechanics: A decade later,” Journal of Cataract and Refractive Surgery 40 (6 ), 857 (2014).24857433 16. Zhao Y. , Shen Y. , Yan Z. , Tian M. , Zhao J. , and Zhou X. , “Relationship Among Corneal Stiffness, Thickness, and Biomechanical Parameters Measured by Corvis ST, Pentacam and ORA in Keratoconus,” Front. Physiol. 10 , 740 (2019).31263429 17. Roberts C. J. , Mahmoud A. M. , Bons J. P. , Hossain A. , Elsheikh A. , Vinciguerra R. , Vinciguerra P. , and Ambrósio R. , “Introduction of Two Novel Stiffness Parameters and Interpretation of Air Puff–Induced Biomechanical Deformation Parameters With a Dynamic Scheimpflug Analyzer,” J Refract Surg 33 (4 ), 266–273 (2017).28407167 18. Bak-Nielsen S. , Bach Pedersen I. , Ivarsen A. , and Hjortdal J. , “The Rigidity of Corneas before and after Corneal Cross-linking - as measured by Corvis® ST,” Investigative Ophthalmology & Visual Science 54 (15 ), 1613 (2013).23361504 19. Sedaghat M. , Naderi M. , and Zarei-Ghanavati M. , “Biomechanical parameters of the cornea after collagen crosslinking measured by waveform analysis,” Journal of Cataract and Refractive Surgery 36 (10 ), 1728–1731 (2010).20870120 20. Urban M. W. , Pislaru C. , Nenadic I. Z. , Kinnick R. R. , and Greenleaf J. F. , “Measurement of Viscoelastic Properties of In Vivo Swine Myocardium Using Lamb Wave Dispersion Ultrasound Vibrometry (LDUV),” IEEE Trans. Med. Imaging 32 (2 ), 247–261 (2013).23060325 21. Zeng Q. , Honarvar M. , Schneider C. , Mohammad S. K. , Lobo J. , Pang E. H. T. , Lau K. T. , Hu C. , Jago J. , Erb S. R. , Rohling R. , and Salcudean S. E. , “Three-Dimensional Multi-Frequency Shear Wave Absolute Vibro-Elastography (3D S-WAVE) With a Matrix Array Transducer: Implementation and Preliminary In Vivo Study of the Liver,” IEEE Trans. Med. Imaging 40 (2 ), 648–660 (2021).33108283 22. Li R. , Qian X. , Gong C. , Zhang J. , Liu Y. , Xu B. , Humayun M. S. , and Zhou Q. , “Simultaneous Assessment of the Whole Eye Biomechanics Using Ultrasonic Elastography,” IEEE Trans. Biomed. Eng. 70 (4 ), 1310–1317 (2023).36260593 23. Zvietcovich F. , Pongchalee P. , Meemon P. , Rolland J. P. , and Parker K. J. , “Reverberant 3D optical coherence elastography maps the elasticity of individual corneal layers,” Nat Commun 10 (1 ), 4895 (2019).31653846 24. Ambroziński Ł. , Song S. , Yoon S. J. , Pelivanov I. , Li D. , Gao L. , Shen T. T. , Wang R. K. , and O’Donnell M. , “Acoustic micro-tapping for non-contact 4D imaging of tissue elasticity,” Sci. Rep. 6 (1 ), 1–11 (2016).28442746 25. Song P. , Zhao H. , Manduca A. , Urban M. W. , Greenleaf J. F. , and Chen S. , “Comb-Push Ultrasound Shear Elastography (CUSE): A Novel Method for Two-Dimensional Shear Elasticity Imaging of Soft Tissues,” IEEE Trans. Med. Imaging 31 (9 ), 1821–1832 (2012).22736690 26. Manduca A. , Oliphant T. E. , Dresner M. A. , Mahowald J. L. , Kruse S. A. , Amromin E. , Felmlee J. P. , Greenleaf J. F. , and Ehman R. L. , “Magnetic resonance elastography: Non-invasive mapping of tissue elasticity,” Medical Image Analysis 5 (4 ), 237–254 (2001).11731304 27. Kwon Oh In , Park Chunjae , Nam Hyun Soo , Woo Eung Je , Seo Jin Keun , Glaser K. J. , Manduca A. , and Ehman R. L. , “Shear Modulus Decomposition Algorithm in Magnetic Resonance Elastography,” IEEE Trans. Med. Imaging 28 (10 ), 1526–1533 (2009).19783495 28. Mellema D. C. , Song P. , Kinnick R. R. , Urban M. W. , Greenleaf J. F. , Manduca A. , and Chen S. , “Probe Oscillation Shear Elastography (PROSE): A High Frame-Rate Method for Two-Dimensional Ultrasound Shear Wave Elastography,” IEEE Trans. Med. Imaging 35 (9 ), 2098–2106 (2016).27076352 29. Liu H.-C. , Gaihre B. , Kijanka P. , Lu L. , and Urban M. W. , “Acoustic Force Elastography Microscopy,” IEEE Trans. Biomed. Eng. 70 (3 ), 841–852 (2023).36049008 30. Kirby M. A. , Pelivanov I. , Song S. , Ambrozinski Ł. , Yoon S. J. , Gao L. , Li D. , Shen T. T. , Wang R. K. , and O’Donnell M. , “Optical coherence elastography in ophthalmology,” J. Biomed. Opt. 22 (12 ), 1 (2017). 31. Liang Xing and Boppart S. A. , “Biomechanical Properties of In Vivo Human Skin From Dynamic Optical Coherence Elastography,” IEEE Trans. Biomed. Eng. 57 (4 ), 953–959 (2010).19822464 32. Kirby M. A. , Pelivanov I. , Regnault G. , Pitre J. J. , Wallace R. T. , O’Donnell M. , Wang R. K. , and Shen T. T. , “Acoustic Micro-Tapping Optical Coherence Elastography to Quantify Corneal Collagen Cross-Linking,” Ophthalmology Science 3 (2 ), 100257 (2023).36685713 33. Pitre J. J. , Kirby M. A. , Li D. S. , Shen T. T. , Wang R. K. , O’Donnell M. , and Pelivanov I. , “Nearly-incompressible transverse isotropy (NITI) of cornea elasticity: model and experiments with acoustic micro-tapping OCE,” Sci Rep 10 (1 ), 1–14 (2020).31913322 34. Kirby M. A. , Tang P. , Liou H.-C. , Kuriakose M. , Pitre J. J. , Pham T. N. , Ettinger R. E. , Wang R. K. , O’Donnell M. , and Pelivanov I. , “Probing elastic anisotropy of human skin in vivo with light using non-contact acoustic micro-tapping OCE and polarization sensitive OCT,” Sci. Rep. 12 (1 ), 1–17 (2022).34992227 35. Aleman-Castaneda L. A. , Zvietcovich F. , and Parker K. J. , “Reverberant Elastography for the Elastic Characterization of Anisotropic Tissues,” IEEE J. Select. Topics Quantum Electron. 27 (4 ), 1–12 (2021). 36. Latorre-Ossa H. , Gennisson J. , De Brosses E. , and Tanter M. , “Quantitative imaging of nonlinear shear modulus by combining static elastography and shear wave elastography,” IEEE Trans. Ultrason., Ferroelect., Freq. Contr. 59 (4 ), 833–839 (2012). 37. Goswami S. , Ahmed R. , Khan S. , Doyley M. M. , and McAleavey S. A. , “Shear Induced Non-Linear Elasticity Imaging: Elastography for Compound Deformations,” IEEE Trans. Med. Imaging 39 (11 ), 3559–3570 (2020).32746104 38. Song S. , Huang Z. , Nguyen T.-M. , Wong E. Y. , Arnal B. , O’Donnell M. , and Wang R. K. , “Shear modulus imaging by direct visualization of propagating shear waves with phase-sensitive optical coherence tomography,” J. Biomed. Opt. 18 (12 ), 121509 (2013).24213539 39. Han Z. , Aglyamov S. R. , Li J. , Singh M. , Wang S. , Vantipalli S. , Wu C. , Liu C. , Twa M. D. , and Larin K. V. , “Quantitative assessment of corneal viscoelasticity using optical coherence elastography and a modified Rayleigh–Lamb equation,” J. Biomed. Opt 20 (2 ), 2–5 (2015). 40. Larin K. V. and Sampson D. D. , “Optical coherence elastography – OCT at work in tissue biomechanics [Invited],” Biomed. Opt. Express 8 (2 ), 1172 (2017).28271011 41. Pelivanov I. , Gao L. , Pitre J. J. , Kirby M. A. , Song S. , Li D. , Shen T. T. , Wang R. K. , and O’Donnell M. , “Does group velocity always reflect elastic modulus in shear wave elastography?,” J. Biomed. Opt. 24 (07 ), 1 (2019). 42. Han Z. , Li J. , Singh M. , Aglyamov S. R. , Wu C. , Liu C. , and Larin K. V. , “Analysis of the effects of curvature and thickness on elastic wave velocity in cornea-like structures by finite element modeling and optical coherence elastography,” Appl. Phys. Lett. 106 (23 ), 233702 (2015).26130825 43. Han Z. , Li J. , Singh M. , Wu C. , Liu C. , Raghunathan R. , Aglyamov S. R. , Vantipalli S. , Twa M. D. , and Larin K. V. , “Optical coherence elastography assessment of corneal viscoelasticity with a modified Rayleigh-Lamb wave model,” Journal of the Mechanical Behavior of Biomedical Materials 66 , 87–94 (2017).27838594 44. Li C. , Guan G. , Cheng X. , Huang Z. , and Wang R. K. , “Quantitative elastography provided by surface acoustic waves measured by phase-sensitive optical coherence tomography,” Opt. Lett. 37 (4 ), 722 (2012).22344160 45. Zhu J. , Qi L. , Miao Y. , Ma T. , Dai C. , Qu Y. , He Y. , Gao Y. , Zhou Q. , and Chen Z. , “3D mapping of elastic modulus using shear wave optical micro-elastography,” Sci Rep 6 (1 ), 35499 (2016).27762276 46. Koudouna E. , Winkler M. , Mikula E. , Juhasz T. , Brown D. J. , and Jester J. V. , “Evolution of the vertebrate corneal stroma,” Progress in Retinal and Eye Research 64 , 65–76 (2018).29398348 47. Winkler M. , Shoa G. , Tran S. T. , Xie Y. , Thomasy S. , Raghunathan V. K. , Murphy C. , Brown D. J. , and Jester J. V. , “A Comparative Study of Vertebrate Corneal Structure: The Evolution of a Refractive Lens,” Invest. Ophthalmol. Vis. Sci. 56 (4 ), 2764 (2015).26066606 48. Kirby M. A. , Regnault G. , Pelivanov I. , O’Donnell M. , Wang R. K. , and Shen T. T. , “Noncontact Acoustic Micro-Tapping Optical Coherence Elastography for Quantification of Corneal Anisotropic Elasticity: In Vivo Rabbit Study,” Trans. Vis. Sci. Tech. 12 (3 ), 15 (2023). 49. Kirby M. A. , Pitre J. J. , Liou H.-C. , Li D. S. , Wang R. K. , Pelivanov I. , O’Donnell M. , and Shen T. T. , “Delineating corneal Elastic anisotropy in a porcine model using noncontact OCT elastography and ex vivo mechanical tests,” Ophthalmol. Sci. 1 (4 ), 100058 (2021).36246948 50. Akhtar S. , Smedowski A. , Masmali A. , Alkanaan A. , Khan A. A. , Almutleb E. , Mofty H. K. , Al-Debasi H. I. , Samivel R. , and Almubrad T. , “Effect of Ultraviolet-A and Riboflavin treatment on the architecture of the center and periphery of normal rat cornea: 7 days post treatment,” Experimental Eye Research 219 , 109064 (2022).35385757 51. Spadea L. , Tonti E. , and Vingolo E. , “Corneal stromal demarcation line after collagen cross-linking in corneal ectatic diseases: a review of the literature,” OPTH Volume 10 , 1803–1810 (2016). 52. Mazzotta C. , Wollensak G. , Raiskup F. , Pandolfi A. M. , and Spoerl E. , “The meaning of the demarcation line after riboflavin-UVA corneal collagen crosslinking,” Expert Review of Ophthalmology 14 (2 ), 115–131 (2019). 53. Blackburn B. J. , Gu S. , Ford M. R. , de Stefano V. , Jenkins M. W. , Dupps W. J. , and Rollins A. M. , “Noninvasive Assessment of Corneal Crosslinking With Phase-Decorrelation Optical Coherence Tomography,” Invest. Ophthalmol. Vis. Sci. 60 (1 ), 41 (2019).30601930 54. Shiose Y. , “Intraocular pressure: New perspectives,” Survey of Ophthalmology 34 (6 ), 413–435 (1990).2194306 55. Wang R. K. , Kirkpatrick S. , and Hinds M. , “Phase-sensitive optical coherence elastography for mapping tissue microstrains in real time,” Appl. Phys. Lett. 90 (16 ), 164105 (2007). 56. Chu B. , “Laser Light Scattering,” Annu Rev Phys Chem. 21 , 145–174. (1970). 57. Greenstein S. A. , Shah V. P. , Fry K. L. , and Hersh P. S. , “Corneal thickness changes after corneal collagen crosslinking for keratoconus and corneal ectasia: One-year results,” Journal of Cataract and Refractive Surgery 37 (4 ), 691–700 (2011).21420594 58. Cordeiro Barbosa M. M. V. , Barbosa J. B. , Hirai F. E. , and Hofling-Lima A. L. , “Effect of Cross-Linking on Corneal Thickness in Patients With Corneal Edema,” Cornea 29 (6 ), 613–617 (2010).20458230 59. Chou P. C. , Carleone J. , and Hsu C. M. , “Elastic Constants of Layered Media,” Journal of Composite Materials 6 (1 ), 80–93 (1972). 60. Sun C. T. and Li S. , “Three-Dimensional Effective Elastic Constants for Thick Laminates,” (n.d.). 61. Nayfeh A. H. , Wave Propagation in Layered Anisotropic Media: With Applications to Composites, North-Holland Series in Applied Mathematics and Mechanics No. v. 39 (Elsevier, 1995). 62. Benjelloun S. and Ghidaglia J.-M. , “On the sound speed in two-fluid mixtures and the implications for CFD model validation,” European Journal of Mechanics - B/Fluids 90 , 152–168 (2021). 63. Karabutov A. A. , “Laser Optoacoustic Measurement of Paper Porosity,” Acoust. Phys. 51 (5 ), 560 (2005). 64. Regnault G. , Kirby M. A. , Kuriakose M. , Shen T. , Wang R. K. , O’Donnell M. , and Pelivanov I. , “Spatial resolution in optical coherence elastography of bounded media,” Biomed. Opt. Express 13 (9 ), 4851 (2022).36187272

PMC010xxxxxx/PMC10327236.txt

==== Front ArXiv ArXiv arxiv ArXiv 2331-8422 Cornell University arXiv:2306.15113v1 2306.15113 1 preprint Article Minimum information and guidelines for reporting a Multiplexed Assay of Variant Effect Claussnitzer Melina 12† Parikh Victoria N. 3† Wagner Alex H. 456† Arbesfeld Jeremy A. 4 Bult Carol J. 7 Firth Helen V. 89 Muffley Lara A. 10 Nguyen Ba Alex N. 11 Riehle Kevin 12 Roth Frederick P. 13141516 Tabet Daniel 13141516 Bolognesi Benedetta 17* Glazer Andrew M. 18* Rubin Alan F. 1920* 1 The Novo Nordisk Foundation Center for Genomic Mechanisms of Disease, Broad Institute of MIT and Harvard, Cambridge, MA 02142, USA 2 Center for Genomic Medicine, Massachusetts General Hospital, Harvard Medical School, Cambridge, MA 02142, USA 3 Stanford Center for Inherited Cardiovascular Disease, Stanford University School of Medicine, Stanford, CA, USA 94305 4 The Steve and Cindy Rasmussen Institute for Genomic Medicine, Nationwide Children’s Hospital, Columbus, OH 43215, USA 5 Department of Pediatrics, The Ohio State University College of Medicine, Columbus, OH 43210, USA 6 Department of Biomedical Informatics, The Ohio State University, Columbus, OH 43210, USA 7 The Jackson Laboratory, Bar Harbor, ME 04609, USA 8 Wellcome Sanger Institute, Hinxton, Cambridge, UK 9 Dept of Medical Genetics, Cambridge University Hospitals NHS Trust, Cambridge UK 10 Department of Genome Sciences, University of Washington, Seattle, WA 98105, USA 11 Department of Biology, University of Toronto at Mississauga, Mississauga, Ontario, Canada 12 Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, TX 77030, USA 13 Donnelly Centre, University of Toronto, Toronto, Ontario, Canada 14 Department of Molecular Genetics, University of Toronto, Toronto, Ontario, Canada 15 Department of Computer Science, University of Toronto, Toronto, Ontario, Canada 16 Lunenfeld-Tanenbaum Research Institute, Sinai Health, Toronto, Ontario, Canada 17 Institute for Bioengineering of Catalunya (IBEC), The Barcelona Institute of Science and Technology, Barcelona, Spain 18 Vanderbilt University Medical Center, Nashville, TN 37232, USA 19 Bioinformatics Division, WEHI, Parkville, Victoria, Australia 20 Department of Medical Biology, University of Melbourne, Parkville, Victoria, Australia † These authors contributed equally to this work. Authors’ contributions MC, VNP, CJB, ANNB, FPR, DT, AMG, and AFR conceptualized the study. AHW, FPR, and AFR developed the methodology. AHW and KR developed software. AHW performed validation. MC, AHW, BB, AMG, and AFR performed formal analysis. AHW conducted investigations. VNP and DT provided resources. VNP and ANNB curated data. MC, AHW, JAA, BB, AMG, and AFR wrote the original draft.MC, VNP, AHW, CJB, HVF, LAM, ANNB, DT, BB, AMG, and AFR reviewed and edited the manuscript. LAM, BB, AMG, and AFR supervised the team. VNP and LAM performed project administration. * To whom correspondence should be addressed bbolognesi@ibecbarcelona.eu, andrew.m.glazer@vumc.org, alan.rubin@wehi.edu.au. 26 6 2023 arXiv:2306.15113v1https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License, which allows reusers to distribute, remix, adapt, and build upon the material in any medium or format, so long as attribution is given to the creator. The license allows for commercial use. nihpp-2306.15113v1.pdf Multiplexed Assays of Variant Effect (MAVEs) have emerged as a powerful approach for interrogating thousands of genetic variants in a single experiment. The flexibility and widespread adoption of these techniques across diverse disciplines has led to a heterogeneous mix of data formats and descriptions, which complicates the downstream use of the resulting datasets. To address these issues and promote reproducibility and reuse of MAVE data, we define a set of minimum information standards for MAVE data and metadata and outline a controlled vocabulary aligned with established biomedical ontologies for describing these experimental designs. genomics standards genetic variants multiplexed assays of variant effect MAVE deep mutational scanning DMS MC received funding from the Novo Nordisk Foundation (NNF21SA0072102) and NIH/NIDDK grant UM1DK126185. VNP received funding from NIH/NHLBI grants K08HL143185 and R01HL164675. AHW received funding from NIH/NHGRI grant R35HG011949. LAM, FPR, and AFR received funding from NIH/NHGRI grants UM1HG011969 and RM1HG010461. ANNB acknowledges funding from CIHR, NSERC, and the University of Toronto. KR receives funding from NIH/NHGRI grant U24HG009649. FPR received funding from a CIHR Foundation Grant. BB received funding from La Caixa Research Foundation grant LCF/PR/HR21/52410004 and the Spanish Ministry of Science, Innovation and Universities grants PID2021-127761OB-I00 and RYC2020-028861-I. AMG received funding from NIH/NHGRI grant R00HG010904 and NIH/NHLBI grant R01HL164675. This work was supported by the Australian government. ==== Body pmcBackground The emergence of high-throughput genomic technologies has revolutionized our ability to study the impact of genetic variants at a grand scale. A prominent example of these innovative methods is Multiplexed Assays of Variant Effect (MAVEs). MAVEs are a family of experimental methods combining saturation mutagenesis with a multiplexed assay to interrogate the effects of thousands of genetic variants in a given functional element in parallel [1,2]. The output of a MAVE is a variant effect map quantifying the consequences of all single nucleotide (or single amino acid) variants in a target functional element, even variants not yet observed in the population. MAVEs have been applied to coding sequences as well as noncoding elements like splice sites and regulatory regions across various organisms. Variant effect maps have broad applications including clinical variant interpretation [2,3], understanding sequence/structure/function relationships [4,5], and investigating molecular mechanisms of evolution [6,7]. The MAVE field is growing rapidly, leading to the formation of organizations such as the Atlas of Variant Effects (AVE). AVE consists of over 400 researchers from over 30 countries who perform, interpret, and apply MAVE experiments. The rapid growth and adoption of MAVE technologies across many fields has led to an excess of overlapping definitions, complicating discovery and interpretation. Minimum information standards in other research areas have increased the reporting, archiving, and reuse of biological data [8–11]. To promote reuse and FAIR data sharing [12], minimum information standards and a controlled vocabulary for describing MAVE experiments and variant effect maps are needed. Here, we—members of the AVE Experimental Technology and Standards and Data Coordination and Dissemination workstreams—provide a comprehensive structured vocabulary and recommendations for data release for MAVE datasets. Uptake of these recommendations by the MAVE community will greatly improve the usability and longevity of MAVE datasets, enabling novel insights and applications. Results and Discussion All MAVEs share a core pipeline: generation of a variant library, delivery of the library into a model system, separation of variants based on function, quantification of variant frequency by high-throughput DNA sequencing, and carrying out data analysis and score calculation [1,2,13]. Accurate and consistent metadata describing each of these steps is the basis for the interpretability of MAVE functional scores and is a requirement for any advanced quantitative analysis, such as comparing and combining scores. To systematize these metadata, we have defined and implemented a computable controlled vocabulary that covers the majority of current and emerging MAVE techniques (Figure 1) [14]. This vocabulary captures the major steps of the MAVE experimental process including project scope, library generation, library integration/expression, assay type, and sequencing method. The vocabulary also contains terms to describe the biological and disease relevance of the assay. In addition to releasing scores and other datasets in published papers, we recommend sharing MAVE datasets through MaveDB, an open-source platform to distribute and interpret MAVE data [15,16]. Researchers should communicate the target sequence, the method used to generate library diversity, and the method of variant delivery into the assay system using terms from the controlled vocabulary. Metadata about the variant generation method should include terms for either editing at the endogenous locus or in vitro variant library generation. It should also specify the model system as defined by NCBI Taxonomy ID [17] and Cell Line Ontology (CLO) [18] terms where available. It is essential for the target sequence to be linked to a reference genome database or similar by including a versioned stable identifier from a widely-used resource such as RefSeq [19], Ensembl [20], or UniProt [21]. We also recommend that researchers designing a study choose a reference-identical allele when it does not otherwise affect the study design, particularly for clinically-relevant targets. The entire target sequence used in the assay must be provided to allow MaveDB and other systems to generate globally unique identifiers (sha512t24u computed identifiers [22]) as used by the Global Alliance for Genomics and Health (GA4GH) [23] refget [24] and Variation Representation Specification (VRS) [25] standards. We recommend that variant libraries are exchanged using VRS and stored using a VRS-compatible information model, including the aforementioned computed identifiers, inter-residue sequence location data, and VOCA-normalized allele representation [25,26]. This allows variants to be defined in terms of both the variant on the target sequence and the homologous variant on the linked reference sequences with an appropriate variant mapping relation, such as the homologous_to relation from the sequence ontology [27]. Descriptions of variants on target sequences should follow the MAVE-HGVS nomenclature conventions [16]. Homologous variants on linked reference sequences should describe variants following conventions typical for the target organism, e.g. using the Human Genome Variation Society (HGVS) variant nomenclature [28] for variants on human reference sequences. An example of these sequence variant recommendations in practice are described by Arbesfeld et al. [29], where they enable interoperability with downstream resources including the Ensembl Variant Effect Predictor (VEP) [30], UCSC Genome Browser [31], the Genomics to Proteins resource [32], the ClinGen Allele Registry [33] and ClinGen Linked Data Hub. The phenotypic assay is the most unique aspect of a MAVE compared to other data types for which minimum information standards have been established. There is a tremendous diversity in functional assays in terms of both the assay readout and the biology the assay was designed to interrogate. For assay readout, we have identified a subset of phenotypic readouts in the Ontology for Biomedical Investigations (OBI) [34] that are commonly used in variant effect maps. Because OBI has over 2,500 terms, we hope that this “short list” will help researchers identify the most relevant terms to describe their experiments. Neverthelesswe also welcome the use of other OBI terms if necessary to describe new assays. Researchers should also detail any environmental variables (such as the addition of small molecules) and use the appropriate controlled vocabulary term for the high-throughput sequencing method used for variant quantification. We strongly recommend that raw sequence reads be deposited in a suitable repository, such as the Sequence Read Archive (SRA) [35] or Gene Expression Omnibus (GEO) [36], along with a description of each file (e.g. time point and sample information). We recommend that researchers investigating clinical phenotypes use terms from the Mondo Disease Ontology (Mondo) [37] or Online Mendelian Inheritance in Man (OMIM) [38] to help clinicians and other stakeholders retrieve relevant functional data. Particular care is needed for genes encoding proteins with multiple functional domains and where loss of function and gain of function variants are associated with different diseases. Ideally, each MAVE should be associated with a particular gene-disease entity that describes the mechanism of disease such as those defined by G2P [39] and how the MAVE assay recapitulates or is relevant to the mechanism of disease. Some genes or functional domains may require multiple MAVE assays, each probing a different function or attribute of the gene product, to accurately model different disease entities. Although it is not within the scope of this controlled vocabulary, it is still crucial to detail the data analysis performed to produce a variant effect map. This includes steps to generate variant counts, including sequence read processing, quality filtering, alignment, and variant identification, as well as further statistical and bioinformatic processing to calculate scores and associated error estimates. Researchers should describe the analysis pipeline used for these calculations, including software versions. Several well-documented tools are available for this purpose and the field continues to advance rapidly [40–42]. Custom code should be shared using GitHub or a similar platform and archived using Zenodo or a similar archival service that mints a DOI. In addition to processed variant scores, we urge researchers to share raw counts for each dataset, as these have tremendous utility for downstream users who want to reanalyze datasets or develop new statistical methods. Similarly, we recommend that researchers also report scores prior to normalization or imputation, and MaveDB supports the deposition of counts, scores, normalized/imputed scores, and sequence metadata for the same dataset (Table 1). Conclusions Minimum data standards are important to guide researchers who want their datasets to be used and cited broadly. We anticipate that this document will enhance the readability and discoverability of current and future datasets by defining a vocabulary that can be adopted across the many fields where MAVEs are being performed and where the resulting datasets are being used. Ensuring a minimum set of available metadata that uses a shared set of terms enables new types of analysis, such as machine learning methods to combine large numbers of disparate, high-dimensional datasets like MAVEs. Large-scale meta-analyses of multiple MAVE datasets have already been implemented in several contexts, including computational prediction of variant effects [43,44] and clinical variant reclassification [45]. In the near term, the controlled vocabulary will be implemented as part of MaveDB records, creating a large set of rich metadata annotations that can be searched and mined. We believe that the MAVE community should share datasets and resources responsibly, and that accessibility is real only when it ensures usability and reproducibility. Methods The initial draft of the controlled vocabulary was developed collaboratively using Google Docs. The controlled vocabulary schema is defined using JSON Schema Draft 2020–12. Acknowledgements The authors would like to thank Michael Boettcher and Melissa S. Cline for thoughtful discussion of this work and comments on the manuscript. We would also like to thank Alex Hopkins for administrative support. The images in Figure 1 were created in Biorender. Funding MC received funding from the Novo Nordisk Foundation (NNF21SA0072102) and NIH/NIDDK grant UM1DK126185. VNP received funding from NIH/NHLBI grants K08HL143185 and R01HL164675. AHW received funding from NIH/NHGRI grant R35HG011949. LAM, FPR, and AFR received funding from NIH/NHGRI grants UM1HG011969 and RM1HG010461. ANNB acknowledges funding from CIHR, NSERC, and the University of Toronto. KR receives funding from NIH/NHGRI grant U24HG009649. FPR received funding from a CIHR Foundation Grant. BB received funding from La Caixa Research Foundation grant LCF/PR/HR21/52410004 and the Spanish Ministry of Science, Innovation and Universities grants PID2021-127761OB-I00 and RYC2020-028861-I. AMG received funding from NIH/NHGRI grant R00HG010904 and NIH/NHLBI grant R01HL164675. This work was supported by the Australian government. Availability of data and materials The controlled vocabulary implementation is available on GitHub from the AVE Data Coordination and Dissemination workstream repository located at https://github.com/ave-dcd/mave_vocabulary and Zenodo at https://doi.org/10.5281/zenodo.8049231 [14]. Figure 1: A structured vocabulary of terms relevant to the technical development, execution and recording of multiplexed assays of variant effects (MAVEs). Examples of each level of controlled vocabulary term are depicted using a published MAVE dataset [46] shown in blue with alternative options shown in gray. Table 1: Recommendation locations for MAVE data deposition Type of data Deposition location Processed scores, unprocessed scores, raw counts MaveDB[15,16] Raw sequence reads Sequence Read Archive[35]/Gene Expression Omnibus[36] Target sequence MaveDB[15,16] Linked sequence references MaveDB[15,16] Sequence metadata / digests MaveDB[15,16]/SeqRepo[22] Variant library MaveDB[15,16] Analysis code GitHub/Zenodo Structured vocabulary description This work/MaveDB[15,16] Competing interests The authors declare that they have no competing interests ==== Refs References 1. Gasperini M , Starita L , Shendure J . The power of multiplexed functional analysis of genetic variants. Nat Protoc. 2016;11 :1782–7.27583640 2. Tabet D , Parikh V , Mali P , Roth FP , Claussnitzer M . Scalable Functional Assays for the Interpretation of Human Genetic Variation. Annu Rev Genet. 2022;56 :441–65.36055970 3. Starita LM , Ahituv N , Dunham MJ , Kitzman JO , Roth FP , Seelig G , Variant Interpretation: Functional Assays to the Rescue. Am J Hum Genet. 2017;101 :315–25.28886340 4. Stein A , Fowler DM , Hartmann-Petersen R , Lindorff-Larsen K . Biophysical and Mechanistic Models for Disease-Causing Protein Variants. Trends Biochem Sci. 2019;44 :575–88.30712981 5. Kinney JB , McCandlish DM . Massively Parallel Assays and Quantitative Sequence–Function Relationships. Annu Rev Genomics Hum Genet. 2019;20 :null. 6. Starr TN , Picton LK , Thornton JW . Alternative evolutionary histories in the sequence space of an ancient protein. Nature. 2017;549 :409–13.28902834 7. Gallego Romero I , Lea AJ . Leveraging massively parallel reporter assays for evolutionary questions. Genome Biol. 2023;24 :26.36788564 8. Brazma A , Hingamp P , Quackenbush J , Sherlock G , Spellman P , Stoeckert C , Minimum information about a microarray experiment (MIAME)-toward standards for microarray data. Nat Genet. 2001;29 :365–71.11726920 9. Taylor CF , Paton NW , Lilley KS , Binz P-A , Julian RK , Jones AR , The minimum information about a proteomics experiment (MIAPE). Nat Biotechnol. 2007;25 :887–93.17687369 10. Brazma A , Ball C , Bumgarner R , Furlanello C , Miller M , Quackenbush J , MINSEQE: Minimum Information about a high-throughput Nucleotide SeQuencing Experiment - a proposal for standards in functional genomic data reporting. Zenodo, 2012. 10.5281/zenodo.5706412 11. Füllgrabe A , George N , Green M , Nejad P , Aronow B , Fexova SK , Guidelines for reporting single-cell RNA-seq experiments. Nat Biotechnol. 2020;38 :1384–6.33188371 12. Wilkinson MD , Dumontier M , Aalbersberg IJJ , Appleton G , Axton M , Baak A , The FAIR Guiding Principles for scientific data management and stewardship. Sci Data. 2016;3 :160018.26978244 13. Fowler DM , Fields S . Deep mutational scanning: a new style of protein science. Nat Methods. 2014;11 :801–7.25075907 14. Wagner AH , Rubin AF . ave-dcd/mave_vocabulary: 0.1.0 (v0.1.0). Zenodo; 2023. 10.5281/zenodo.8049231 15. Esposito D , Weile J , Shendure J , Starita LM , Papenfuss AT , Roth FP , MaveDB: an open-source platform to distribute and interpret data from multiplexed assays of variant effect. Genome Biol. 2019;20 :223.31679514 16. Rubin AF , Min JK , Rollins NJ , Da EY , Esposito D , Harrington M , MaveDB v2: a curated community database with over three million variant effects from multiplexed functional assays. 2021;2021.11.29.470445. 17. Schoch CL , Ciufo S , Domrachev M , Hotton CL , Kannan S , Khovanskaya R , NCBI Taxonomy: a comprehensive update on curation, resources and tools. Database J Biol Databases Curation. 2020;2020:baaa062. 18. Sarntivijai S , Lin Y , Xiang Z , Meehan TF , Diehl AD , Vempati UD , CLO: The cell line ontology. J Biomed Semant. 2014;5 :37. 19. O’Leary NA , Wright MW , Brister JR , Ciufo S , Haddad D , McVeigh R , Reference sequence (RefSeq) database at NCBI: current status, taxonomic expansion, and functional annotation. Nucleic Acids Res. 2016;44 :D733–745.26553804 20. Cunningham F , Allen JE , Allen J , Alvarez-Jarreta J , Amode MR , Armean IM , Ensembl 2022. Nucleic Acids Res. 2022;50 :D988–95.34791404 21. UniProt Consortium. UniProt: the Universal Protein Knowledgebase in 2023. Nucleic Acids Res. 2023;51 :D523–31.36408920 22. Hart RK , Prlić A . SeqRepo: A system for managing local collections of biological sequences. PloS One. 2020;15 :e0239883.33270643 23. Rehm HL , Page AJH , Smith L , Adams JB , Alterovitz G , Babb LJ , GA4GH: International policies and standards for data sharing across genomic research and healthcare. Cell Genomics. 2021;1 :100029.35072136 24. Yates AD , Adams J , Chaturvedi S , Davies RM , Laird M , Leinonen R , Refget: standardized access to reference sequences. Bioinforma Oxf Engl. 2021;38 :299–300. 25. Wagner AH , Babb L , Alterovitz G , Baudis M , Brush M , Cameron DL , The GA4GH Variation Representation Specification: A computational framework for variation representation and federated identification. Cell Genomics. 2021;1 :100027.35311178 26. Holmes JB , Moyer E , Phan L , Maglott D , Kattman B . SPDI: data model for variants and applications at NCBI. Bioinforma Oxf Engl. 2020;36 :1902–7. 27. Eilbeck K , Lewis SE , Mungall CJ , Yandell M , Stein L , Durbin R , The Sequence Ontology: a tool for the unification of genome annotations. Genome Biol. 2005;6 :R44.15892872 28. den Dunnen JT , Dalgleish R , Maglott DR , Hart RK , Greenblatt MS , McGowan-Jordan J , HGVS Recommendations for the Description of Sequence Variants: 2016 Update. Hum Mutat. 2016;37 :564–9.26931183 29. dcd-mapping [Internet]. ave-dcd; 2023 [cited 2023 May 30]. Available from: https://github.com/ave-dcd/dcd_mapping 30. McLaren W , Gil L , Hunt SE , Riat HS , Ritchie GRS , Thormann A , The Ensembl Variant Effect Predictor. Genome Biol. 2016;17 :122.27268795 31. Nassar LR , Barber GP , Benet-Pagès A , Casper J , Clawson H , Diekhans M , The UCSC Genome Browser database: 2023 update. Nucleic Acids Res. 2023;51 :D1188–95.36420891 32. Iqbal S , Pérez-Palma E , Jespersen JB , May P , Hoksza D , Heyne HO , Comprehensive characterization of amino acid positions in protein structures reveals molecular effect of missense variants. Proc Natl Acad Sci U S A. 2020;117 :28201–11.33106425 33. Pawliczek P , Patel RY , Ashmore LR , Jackson AR , Bizon C , Nelson T , ClinGen Allele Registry links information about genetic variants. Hum Mutat. 2018;39 :1690–701.30311374 34. Brinkman RR , Courtot M , Derom D , Fostel JM , He Y , Lord P , Modeling biomedical experimental processes with OBI. J Biomed Semant. 2010;1 Suppl 1 :S7. 35. Leinonen R , Sugawara H , Shumway M . The Sequence Read Archive. Nucleic Acids Res. 2011;39 :D19–21.21062823 36. Barrett T , Wilhite SE , Ledoux P , Evangelista C , Kim IF , Tomashevsky M , NCBI GEO: archive for functional genomics data sets—update. Nucleic Acids Res. 2013;41 :D991–5.23193258 37. Mungall CJ , McMurry JA , Köhler S , Balhoff JP , Borromeo C , Brush M , The Monarch Initiative: an integrative data and analytic platform connecting phenotypes to genotypes across species. Nucleic Acids Res. 2017;45 :D712–22.27899636 38. Hamosh A , Amberger JS , Bocchini C , Scott AF , Rasmussen SA . Online Mendelian Inheritance in Man (OMIM®): Victor McKusick’s magnum opus. Am J Med Genet A. 2021;185 :3259–65.34169650 39. Thormann A , Halachev M , McLaren W , Moore DJ , Svinti V , Campbell A , Flexible and scalable diagnostic filtering of genomic variants using G2P with Ensembl VEP. Nat Commun. 2019;10 :2373.31147538 40. Bloom JD . Software for the analysis and visualization of deep mutational scanning data. BMC Bioinformatics. 2015;16 :168.25990960 41. Rubin AF , Gelman H , Lucas N , Bajjalieh SM , Papenfuss AT , Speed TP , A statistical framework for analyzing deep mutational scanning data. Genome Biol. 2017;18 :150.28784151 42. Faure AJ , Schmiedel JM , Baeza-Centurion P , Lehner B . DiMSum: an error model and pipeline for analyzing deep mutational scanning data and diagnosing common experimental pathologies. Genome Biol. 2020;21 :207.32799905 43. Wu Y , Li R , Sun S , Weile J , Roth FP . Improved pathogenicity prediction for rare human missense variants. Am J Hum Genet. 2021;108 :1891–906.34551312 44. Frazer J , Notin P , Dias M , Gomez A , Min JK , Brock K , Disease variant prediction with deep generative models of evolutionary data. Nature. 2021;599 :91–5.34707284 45. Fayer S , Horton C , Dines JN , Rubin AF , Richardson ME , McGoldrick K , Closing the gap: Systematic integration of multiplexed functional data resolves variants of uncertain significance in BRCA1, TP53, and PTEN. Am J Hum Genet. 2021;108 :2248–58.34793697 46. Seuma M , Lehner B , Bolognesi B . An atlas of amyloid aggregation: the impact of substitutions, insertions, deletions and truncations on amyloid beta fibril nucleation. Nat Commun. 2022;13 :7084.36400770

PMC010xxxxxx/PMC10336092.txt

==== Front Nat Commun Nat Commun Nature Communications 2041-1723 Nature Publishing Group UK London 37433782 39221 10.1038/s41467-023-39221-x Article Global forest fragmentation change from 2000 to 2020 http://orcid.org/0000-0003-3412-7766 Ma Jun ma_jun@fudan.edu.cn http://orcid.org/0009-0005-2176-3861 Li Jiawei http://orcid.org/0000-0002-0581-9774 Wu Wanben http://orcid.org/0000-0002-1923-5964 Liu Jiajia liujiajia@fudan.edu.cn grid.8547.e 0000 0001 0125 2443 Ministry of Education Key Laboratory for Biodiversity Science and Ecological Engineering, Coastal Ecosystems Research Station of the Yangtze River Estuary, Institute of Biodiversty Science, School of Life Sciences, Fudan University, #2005 Songhu Road, Shanghai, 200438 China 11 7 2023 11 7 2023 2023 14 375218 11 2022 2 6 2023 © The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. A comprehensive quantification of global forest fragmentation is urgently required to guide forest protection, restoration and reforestation policies. Previous efforts focused on the static distribution patterns of forest remnants, potentially neglecting dynamic changes in forest landscapes. Here, we map global distribution of forest fragments and their temporal changes between 2000 and 2020. We find that forest landscapes in the tropics were relatively intact, yet these areas experienced the most severe fragmentation over the past two decades. In contrast, 75.1% of the world’s forests experienced a decrease in fragmentation, and forest fragmentation in most fragmented temperate and subtropical regions, mainly in northern Eurasia and South China, declined between 2000 and 2020. We also identify eight modes of fragmentation that indicate different recovery or degradation states. Our findings underscore the need to curb deforestation and increase connectivity among forest fragments, especially in tropical areas. Forest losses and gains are highly dynamic processes. Here, the authors present a forest fragmentation index to map distribution and temporal changes of forest fragments globally, revealing major trends and patterns during the first two decades of the 21st century. Subject terms Forest ecology Conservation biology Sustainability https://doi.org/10.13039/501100001809 National Natural Science Foundation of China (National Science Foundation of China) 32271659 U2106209 Ma Jun National Key Research and Development Program of China (2022YFF0802400)issue-copyright-statement© Springer Nature Limited 2023 ==== Body pmcIntroduction Forest fragmentation is a major driver of global biodiversity loss and ecosystem degradation1–4. Identifying which areas have the most severe forest fragmentation is a fundamental task in ecology, but the findings from investigations into the patterns of global forest fragmentation have been inconsistent5–7. Forests in tropical regions are regarded as more contiguous than those in other regions5,8 and Morreale et al.9 concluded that temperate forests are 1.5 times more fragmented than tropical forests. In contrast, other studies implied that tropical forests are undergoing the most severe forest fragmentation because of an acceleration in deforestation in these regions10–12. For example, the fraction of forest edge area to total forest area in the tropics increased from 27% in 2000 to 31% in 201012, and there was a net loss of forest cover in the tropics in the same period, while many countries in temperate regions achieved net forest gains13,14. These contrasting findings may have been due to a variation in how forest fragmentation was defined on different temporal scales. To gain a better understanding of the consequences of forest fragmentation, a comprehensive quantification of global forest fragmentation is urgently needed. Forest fragmentation is a landscape-level process of forest change, and mainly occurs over time as decreased patch size, increased patch number of patches, and more forest edges15. Accurate assessments and maps of forest fragmentation are crucial for exploring its effects on biodiversity and ecosystem functions. However, previous studies on global forest fragmentation mapping have relied on evaluating static landscape patterns5,6,8 and have ignored the fact that forest fragmentation is a dynamic process over time. Moreover, although the effects of forest fragmentation can persist for up to a century16, the effects are mostly immediate and obvious for only the first several decades after formation17,18. Therefore, accurate quantification of the current patterns and dynamics of forest fragmentation is key to preventing future biodiversity loss and ecosystem degradation. However, the lack of quantification of global forest fragmentation over recent decades from a dynamic perspective, represents an important knowledge gap. Over the past several decades, the assessment of forest fragmentation has become increasingly complex because deforestation and afforestation is highly dynamic. Forest fragmentation is generally thought to be associated with forest loss5,15, and it is thus generally assessed in terms of deforested area. As such, most previous studies on forest fragmentation dynamics focused on tropical forests with high deforestation rates6,11,12. However, forest fragmentation and forest cover changes are relatively independent and do not always proceed in the same direction7. Many temperate and subtropical countries have gained forest cover19, yet forest cover gains do not reverse the trend of forest fragmentation7. Consequently, the credibility of large-scale estimates of forest landscape pattern dynamics will be reduced when considering only either fragmentation or coverage. Thus, there is an urgent need to effectively integrate and analyze changes in forest cover and fragmentation patterns to inform forest management decisions. Here, we constructed a synthetic forest fragmentation index (FFI) to represent the main characteristics of forest fragmentation, including edge, isolation, and patch size effects. We used the FFIs in 2000 and 2020 and their differences (ΔFFI) to determine the static and dynamic patterns of global forest fragmentation. These two indexes were compared among various regions globally to explore which index more effectively and accurately reflects the fragmentation status. We also analyzed key processes and potential causes for some global hotspots of forest fragmentation, and combined changes in forest coverage (ΔFC) and fragmentation to derive a two-dimensional framework for the assessment of forest landscape dynamic patterns. Our study reveals that 75.1% the world’s forest landscapes experienced decreased fragmentation but those in tropical regions experienced increased fragmentation during the first two decades of the 21st century. Results Differences in static and dynamic forest fragmentation index values We calculated the static FFI using the average weighted values of normalized edge density (ED), patch density (PD), and mean patch area (MPA, using 1-normalized MPA) for 2000 and 2020 (Fig. 1), while the values for the individual normalized ED, PD, and MPA layers are shown in Supplementary Fig. 1. Forest landscapes with low static fragmentation (FFI < 0.2) were mainly in the tropics, western Canada, western Siberia, and Far East Russia, while forests with high static fragmentation (FFI > 0.8) were mainly in eastern North America, southern Europe, central and South China, and along the edges of tropical forests. The area proportion of forest landscapes with different fragmentation levels remained stable during 2000–2020. There was a slight increase in the percentage of forest landscapes with low static fragmentation from 2000 to 2020 (17% to 19%), and the percentage of forest landscapes with high static fragmentation also decreased slightly from 17% in 2000 to 13% in 2020.Fig. 1 Global distributions of the static forest fragmentation index (FFI) and the dynamic FFI (ΔFFI) for global forest landscapes. a Static FFI in 2000, (b) static FFI in 2020, (c) ΔFFI from 2000 to 2020; and comparisons of (d) FFI values for 2000 and 2020, and (e) ΔFFI values across climatic zones. d, e The bar heights showed mean value, and error bars showed one standard deviation. The significance of the differences in static FFIs and ΔFFI among climatic zones was tested using the two-side Tukey-HSD test, which adjusted for multiple comparisons, and letters in each bar showed post-hoc differences in mean static FFIs and ΔFFI with P < 0.001. d The numbers of forest pixels (n) from tropical to boreal zones are 581,649, 381,768, 1,031,907, and 1,384,718 in 2000, and 569,260, 378,518, 1,055,401, and 1,385,159 in 2020, respectively. e The numbers of forest pixels (n) from tropical to boreal zones are 553,655, 363,858, 1,021,172, and 1,320,564, respectively. However, the dynamic FFI (ΔFFI) for the period from 2000 to 2020 exhibited a dramatically different pattern from the static FFIs. Approximately 75.1% of global forest landscapes showed a decline in fragmentation (ΔFFI < 0) especially in the western Canada, western and the Far East Russia, and central and South China (Fig. 1c). Forest landscapes with increased fragmentation trends (ΔFFI > 0) were mainly in tropical areas, especially the southeastern Amazon, the Congo Basin, Indochina Peninsula, and some regions in western North America and central Siberia. Meanwhile, fragmentation was relatively stable (ΔFFI ~ 0) in the central Amazon, central and eastern Europe, and the southeastern US. We then compared the static FFI and ΔFFI across climate zones (Fig. 1d) and found that the world’s most fragmented forests were distributed in the subtropics, while the most intact forests were in the tropics and boreal regions. Forests in the subtropical zone had significantly higher static FFI values than those in the other zones in both years (0.64 ± 0.34 in 2000 and 0.62 ± 0.31 in 2020, P < 0.001), while forests in the tropical regions (0.43 ± 0.38, P < 0.001) and boreal regions (0.41 ± 0.20, P < 0.001) had the lowest static FFI values in 2000 and 2020, respectively. However, the mean ΔFFI value in the tropics (0.01 ± 0.104, P < 0.001) was significantly higher than in the other zones (−0.06 to −0.02). In addition, the ΔFFI generally had a significant negative relationship with the FFI2000 across all climatic zones (using a generalized additive model, P < 0.001 for all four zones; Supplementary Fig. 2). Furthermore, the static FFIs were significantly and positively correlated with altitude, while there was a significantly negative correlation between ΔFFI and altitude (Supplementary Fig. 3), indicating lowland forests were relatively intact but experienced more severe fragmentation during 2000–2020. Modes and potential causes of forest fragmentation processes We identified eight modes of forest fragmentation based on the possible combinations of change (increase or decrease) in the three individual components of FFI (ED, PD, and MPA) and detected the area composition proportions of the eight modes for areas with decreased (ΔFFI < 0) and increased (ΔFFI > 0) forest fragmentation, respectively. Among the areas with decreased FFI, the EDdownPDdownMPAup, with an area proportion of 69.8%, was the most common mode and widely distributed all over the world (Fig. 2). Other modes, such as EDdownPDdownMPAdown and EDupPDdownMPAup had area proportions of 15.4% and 8.6% and were mainly dominant in the central Amazon and eastern Europe, respectively. In terms of the areas with increased FFI, the EDupPDupMPAdown was the most common mode and accounted for 53.3% of the total, which occurred mainly in the tropics, western North America, northern Europe, and central Siberia. The EDupPDdownMPAdown and EDdownPDdownMPAdown modes had area proportions of 23.6% and 8.7%, respectively, and were predominant in the tropics, Russia, and western Africa.Fig. 2 Spatial distributions and composition proportions of eight forest fragmentation process modes for forest fragmentation decreased areas (the upper row) and forest fragmentation increased areas (the below row). ED, PD, and MPA mean the individual components of the synthetic forest fragmentation index (FFI). The marks of “up” and “down” after each FFI component represent an increase and decrease trend during 2000–2020, respectively. The pie charts on the right represent the area proportions of eight forest fragmentation process modes in FFI decreased (the number of forest pixels, n = 2,470,511) and increased areas (n = 820,937). Three hotspots in the most obvious FFI decreased and increased areas were selected respectively to evaluate the changes in each component of FFI. Hotspots D1-D3 were in western Canada (n = 7007), southern Europe (n = 3610), and central China (n = 5180), and hotspots I1-I3 were in the southeastern Amazon (n = 2710), the Congo Basin (n = 5102), and central Siberia (n = 6640), respectively. The boxes of hotspots display the median value, lower 25% and upper 75% quartiles; the dots represent the mean value; and the whiskers are extended to the limit of the 1.5-fold interquartile ranges (IQRs). We also detected the changes in individual fragmentation-related metrics for hotspots in areas with decreased and increased FFI, respectively. Overall, hotspots where FFI decreased had declines in ED and PD, and increases in MPA. Specifically, MPA increased significantly by 73% in western Canada, by 38% in southern Europe, and by 50% in central China from 2000 to 2020 (Fig. 2). Conversely, increased ED and PD played a more important role in hotspots where FFI increased. ED increased dramatically by 41% in the southeastern Amazon, by 81% in the Congo Basin, and by 90% in central Siberia, while the increments of PD in these hotspots were 32%, 186%, and 78%, respectively. However, MPA decreased slightly (−8% to −31%) in these hotspots where forest fragmentation increased. Using generalized linear models, we further explored the relationships between ΔFFI and explanatory factors (see Methods) for the globe and the six hotspots. Although ΔFFI was not significantly correlated with any explanatory variables at the global scale (Supplementary Fig. 4), we found that anthropogenic activity factors (nighttime light, nighttime light change, cropland coverage, and cropland change) dominated the changes in FFI during 2000–2020 in the most developed areas, such as the eastern US, Europe, and South China (Supplementary Fig. 5). Moreover, wildfire mainly controlled the ΔFFI of some areas in Canada, Far East Russia, the southeastern Amazon, tropical Africa, and Australia. In addition, for hotspots with decreased FFI (Fig. 3a–c), ΔFFI was most strongly related to wildfire frequency (P < 0.001, standardized coefficient = 0.061) in western Canada, while the most important driving factors of ΔFFI in southern Europe and central China were mean cropland coverage (P < 0.001, standardized coefficient = 0.244) and cropland coverage change (P < 0.001, standardized coefficient = −0.132), respectively. For hotspots with increased FFI (Fig. 3d–f), wildfire frequency was the strongest driving factor of ΔFFI in the southeastern Amazon (P < 0.001, standardized coefficient = 0.299) and central Siberia (P < 0.001, standardized coefficient = 0.466), while ΔFFI in the Congo Basin was significantly affected by all factors except nighttime light.Fig. 3 Standardized correlation coefficients of the dynamic forest fragmentation index (ΔFFI) for six hotspots. Relative effects of anthropogenic activity (the mean and difference values of cropland coverage and nighttime light, yellow color), demographic pressure (the mean and difference values of population density, blue color), and natural disturbance (fire frequency, red color) on dynamics of ΔFFI in (a–c) forest fragmentation decreased hotspots (D1–D3, n = 7661, 3663, and 5271 forest pixels) and in (d–f) forest fragmentation increased hotspots (I1-I3, n = 4044, 5104, and 7689 forest pixels). The dots represent standardized coefficient estimates with 95% (thin segments, ±1.960 standard errors) and 90% (thick segments, ±1.645 standard errors) confidence intervals in generalized linear models. Locations of the six hotspots can be checked in Fig. 2. A new framework for the assessment of global forest landscape dynamics As changes in areas and patterns are two key indicators when assessing forest landscape dynamics, we developed a two-dimensional framework based on the changes in FC and FFI between 2000 and 2020 to obtain a comprehensive understanding of forest landscape dynamics. We found that the forest landscapes with a pattern of FCupFFIdown were distributed worldwide but were concentrated in the western Canada, the northeastern US, northern Eurasia and central China (Fig. 4a). FCupFFIdown generally accounted for the highest forest area percentage in temperate (50.0%) and boreal (59.2%) regions (Fig. 3b). In contrast, the percentage of forest landscape area that exhibited the FCdownFFIup pattern in tropical regions (39.8%) was much higher than that in subtropical (27.9%), temperate (14.3%) or boreal (10.6%) regions. Forest landscapes with a pattern of FCdownFFIup were distributed mostly in the tropics, northern Europe, and central Siberia. Moreover, forest landscapes with patterns of FCupFFIup and FCdownFFIdown, which accounted for 5.7–7.8% and 24.5–34.2% of the total area, were mainly distributed in northern Europe, the central Amazon, and south tropical Africa respectively.Fig. 4 Spatial distributions of different forest landscape dynamic patterns and their area percentages among climatic zones. Landscape dynamic pattern is defined by the changes in forest fragmentation index (FFI) and forest coverage (FC), and the marks of “up” and “down” after FFI or FC represent an increase and decrease trend during 2000–2020, respectively. a Global spatial distribution of four forest landscape dynamic patterns, (b) relative area percentages of the four-forest landscape dynamic patterns among climatic zones, and (c) the relationship between ΔFFI and ΔFC for forest landscapes at national scale using the Pearson’s linear correlation (n = 131 countries). The statistical significance in (c) was obtained with a two-side Student’s T-test. We also used this framework to assess forest landscapes dynamic patterns for forested countries worldwide. Forest landscapes in most countries generally exhibited the FCupFFIdown (n = 40), FCdownFFIup (n = 32) or FCdownFFIdown (n = 54), while those in only five countries predominantly displayed the FCupFFIup pattern. At the national scale, the ΔFFI of a country was significantly and negatively correlated with the ΔFC (R2 = 0.35, P < 0.001) (Fig. 3c). In addition, among the 10 countries with the largest forest area, the overall forest landscape dynamics pattern was exhibited as either FCupFFIdown (Russia, China, and India), FCdownFFIup (Brazil, Australia, the Democratic Republic of the Congo, and Peru), or FCdownFFIdown (Canada, the US, and Indonesia) (Table 1). The FCupFFIdown pattern was found for more than half of the forest landscapes in Russia (59%), Canada (56%), the US (53%) and China (58%), while a considerable proportion of total forest landscapes in Brazil (42%), the Democratic Republic of Congo (54%) and Peru (43%) had a FCdownFFIup pattern. In particular, China (ΔFC = 1.21%, ΔFFI = −0.07) had relatively high ΔFC and low ΔFFI values, while Brazil (ΔFC = −3.22%, ΔFFI = 0.014) had relatively low ΔFC and high ΔFFI values. Moreover, for forest landscapes in the world’s major forested countries, the FCdownFFIdown pattern occupied high proportions (23% − 35%) of the forested areas, while FCupFFIup pattern only accounted for very low proportions (4–15%).Table 1 Areas, area percentages, and mean values of the changes in forest coverage and fragmentation of four forest landscape dynamic patterns in the ten countries with the largest forest area globally Country Area and area percentages △FC △FFI FCupFFIdown FCupFFIup FCdownFFIdown FCdownFFIup 105 km2 % 105 km2 % 105 km2 % 105 km2 % Mean Mean Russian federation 173.6 58.8 18.2 6.2 70.6 23.9 32.6 11.1 0.57 −0.05 Brazil 7.7 16.5 3.1 6.7 16.1 34.5 19.8 42.3 −3.22 0.01 Canada 52.2 56.1 4.8 5.2 26.0 28.0 10.0 10.7 −0.20 −0.07 United States of America 38.3 53.0 4.1 5.6 20.1 27.8 9.9 13.6 −0.43 −0.04 China 24.7 58.5 1.6 3.7 14.1 33.4 1.9 4.4 1.21 −0.07 Australia 4.3 33.0 1.9 14.6 3.3 25.1 3.5 27.3 −0.48 0.03 Democratic Republic of the Congo 2.9 15.8 1.4 7.8 4.1 22.6 9.8 53.8 −2.35 0.01 Indonesia 4.3 25.6 1.5 8.8 5.8 34.7 5.2 30.9 −1.54 −0.01 Peru 1.2 14.7 0.7 7.9 2.8 34.1 3.6 43.3 −0.97 0.01 India 3.0 37.0 1.1 13.2 2.4 29.8 1.6 20.0 0.12 −0.02 FCupFFIdown: increased coverage and decreased fragmentation, FCupFFIup: increased coverage and increased fragmentation, FCdownFFIdown: decreased coverage and decreased fragmentation, and FCdownFFIup decreased coverage and increased fragmentation. Discussion We developed an integrated FFI and evaluated the static and dynamic patterns of forest fragmentation between 2000 and 2020, and the results identified the world’s most fragmented forests and those that experienced the most severe fragmentation. Consistent with previous studies5,20,21, we found that forest landscapes were relatively intact in the Amazon, the Democratic Republic of the Congo, Borneo, and New Guinea, which host some of the highest biodiversity in the world22. However, these areas have also experienced the most severe forest fragmentation over the last two decades according to their higher positive ΔFFI values. For example, tropical regions suffered from more intensive conditions that drove the conversion of intact forests into fragmented forests7,10–12. In contrast, although forest landscapes in Europe and South China are highly fragmented, they are recovering as a result of afforestation and effective protection efforts, which has significantly improved forest landscapes. These findings demonstrated that the dynamic FFI, which reflects the nature of fragmentation (changes in forest distribution pattern), may be more appropriate in evaluating forest fragmentation than the static FFI. Our approach separates forest fragmentation from the distribution of forest fragments, which are often conflated in forest fragmentation assessments. The static FFI mainly reveals forest distribution patterns that are the long-term consequence of climate, topography, and historical land cover changes since the onset of the Anthropocene23–26, and the dynamic FFI represents the processes of forest fragmentation more accurately. In addition, the two-dimensional assessment framework we developed based on changes in forest fragmentation and forest coverage further reduced the uncertainties in the evaluation of forest landscape dynamic patterns due to the inconsistencies between these two factors. For example, we found that the FFI decreased in some areas in central Canada and the central Amazon between 2000 and 2020, but forest losses were still found in those areas. These findings further emphasize the value of our proposed two-dimensional assessment framework in the evaluation of forest landscape pattern dynamics, and it also provides reasonable approaches for evaluating forest fragmentation and its dynamics at regional-national-global scales. Thus, our approach can be used to link these changes in the spatiotemporal pattern of forest fragmentation to forest management policies27,28. We found that 75.1% of global forests experienced a decline in fragmentation during the first 20 years of the 21st century, which suggested that most global forest landscapes were generally improving. However, forest fragmentation exhibited divergent patterns in different regions of the world. On the one hand, we found an extensive decline in forest fragmentation in the world’s most densely populated and economically developed regions (the eastern US, Europe, and South China). On the other hand, the remarkable increase in FFI and decrease in FC in tropical areas presumably reflected the fact that forests in these regions are under tremendous pressure from human beings11,29,30. If these trends continue, forest fragmentation in the tropics will be further exacerbated, and the ecological functions and values of these forest landscapes will further decline and seriously undermine the role of these forests in international climate agreements and biodiversity conservation. In our study, various forest landscape dynamic patterns, generated by the two-dimensional assessment framework, reflected different recovery or degradation states of forests worldwide. In the early stage of forest degradation, small or irregular forest patches are cleared first, resulting in an FCdownFFIdown pattern. As forest loss spreads to large intact patches, more forest edges are created, which causes the forest to exhibit the FCdownFFIup pattern and enter a deep degradation stage. In contrast, in the forest recovery (forest protection and afforestation) scenario, the introduction of additional small patches until they are connected into large intact forest patches causes forests to exhibit FCupFFIup (more forest patches) and FCupFFIdown (fewer forest patches) patterns, which represent the initial and deep stages of the forest recovery, respectively. Our findings indicated that there was a large contrast in the relationship between forests and human beings globally. The widespread deep recovery of forests in some subtropical regions (particularly in South China) and forests with deep degradation in some tropical regions (particularly in the Brazilian Amazon) were both mainly attributed to government forest policies (afforestation and deforestation)31–35. For example, the substantial decrease in FFI and increase in FC in central and South China since 2000 have been mainly attributed to the implementation of various ecological protection projects36–38. In contrast, the land policy promulgated by the Brazilian government and its associated wildfire disturbances led to sharp forest losses39 and significantly exacerbated forest fragmentation in the southeastern Amazon. Similarly, large forest losses in southeast Asia and tropical Africa were driven by increasing human pressures according to satellite-based evidence14,40–42. However, considering that the early recovery and early degradation stages accounted for considerable percentages of all global forest landscapes, there is a large opportunity for society to improve the forest landscape dynamic patterns by adjusting policies. Particularly in the tropics, timely conservation and restoration efforts will prevent further damage and maintain the important functions of these forests in global biodiversity conservation and climate change mitigation. Identifying the distribution and composition of forest fragmentation modes enhances our insights for understanding the forest fragmentation processes and drivers. We found that the most typical FFI decrease mode (EDdownPDdownMPAup) and FFI increase mode (EDupPDupMPAdown) accounted for more than half of the world’s relevant forest landscapes, which indicated that edge, isolation, and patch size effects changed synergistically with forest cover change for most global forest landscapes. However, we also found some atypical FFI change modes. For example, in the central Amazon, MPA decreased in some areas where FFI decreased, while PD decreased in some areas where FFI decreased, which indicated that the processes of forest fragmentation were extremely complex. Therefore, efforts in detecting the underlying mechanism of forest fragmentation change should be site-specific, and focus on the relationship between explanatory factors and forest landscape patterns. By coupling the changes in individual FFI components, we investigated ΔFFI and its associations with anthropogenic and natural factors and identified revealed the possible causes for forest fragmentation dynamics for some hotspots. For hotspots with increased FFI in the southeastern Amazon, large intact forest patches have been converted into multiple small patches under the mixed pressures from commercial harvest, cropland expansion and fire disturbances39, causing serious forest losses and an increase in fragmentation (Supplementary Fig. 6). In central Siberia, however, forest losses due to fire disturbances, especially in forest edges, directly increased forest fragmentation during 2000–2020. For hotspots with decreased FFI in subtropical regions, especially in central China, the decrease in fragmentation was highly related to the implementation of ecological restoration projects under rapid economic development. For example, afforestation efforts under the “grain to green” project increased forest area and connected discrete forest patches43. Also, for hotspots in western Canada and eastern Europe, the decline of FFI was mainly attributed to fire disturbances and changes in cropland area, respectively. These factors increase MPA, reduce ED and PD, and ultimately reduce forest fragmentation by smoothing forest edges and reducing small forest patches. Although climate change is unlikely to cause changes in forest distribution on a 20-year timescale in most regions of the world, it is possible that climate may still have certain impacts on forest fragmentation dynamics in some regions. For example, the decrease in FFI in northern Eurasia between 2000 and 2020 may also be attributed to forest expansion caused by climate warming at high latitudes44, which caused the conversion of small patches into large patches and reduced forest fragmentation. In contrast, more frequent fires, caused by climate change in Canada, Far East Russia, the Brazilian Amazon, tropical Africa and coastal Australia (Supplementary Fig. 5) have resulted in remarkable forest losses and intensified forest fragmentation45–48. However, the complexities of forest fragmentation processes and the causes also remind us that forest fragmentation studies should be targeted and localized. The availability of precise forest distribution data and a thorough understanding of forest landscape dynamic drivers, including land policy, climate change, and international trade, are essential conditions for research on the patterns, causes, ecological consequences, and coping strategies of forest fragmentation. Moreover, it should be noted that the use of bi-temporal forest cover data also makes it impossible to fully assess the continuous dynamics, especially in some areas of sub-tropical forestry or humid tropical shifting cultivation. Therefore, more comprehensive analyses should be conducted that consider specific species characteristics, vegetation types, and multi-temporal forest cover data in the detection of the causes of forest fragmentation dynamics. These analyses also form the basis for applying the assessment of patterns and causes of forest fragmentation to biodiversity conservation and carbon cycle feedback mechanisms. By coupling landscape changes and multiple features of fragmentation, our approach overcomes the problem of considering only the static state of fragmentation instead of dynamic change, and thus it improves our understanding of the patterns of global forest fragmentation. It also more effectively reflects the reality of forest landscape changes and is valuable for the timely formulation and adjustment of relevant policies. In addition, negative states of forests exist for most of the 10 countries with the largest forest areas, which demonstrates remarkable agricultural land expansion, timber harvest and forest fire disturbance in recent years, as well as an overall increase in other disturbances. These changes result in the further loss of forest area, the intensification of fragmentation, and the degradation of ecosystem functions28,49,50. However, the significant negative relationship between ΔFFI and ΔFC at the national scale suggested that efforts that aimed to increase forest area were still effective in mitigating fragmentation. Targeted afforestation and protection measures are important approaches to prevent further deterioration of fragmentation globally. Our findings highlight that an understanding of deforestation and fragmentation dynamics needs to be incorporated into the policy-making process in these countries to minimize irreversible damage to vital forest ecosystems and to redirect the course of development toward sustainability. Methods Datasets Global forest cover maps High-resolution (30 m) global forest cover data for 2000 and 2020 were obtained from the global land cover and land use (GLCLU) change dataset51. The GLULC forest cover data defines a pixel with tree height ≥5 m at the Landsat pixel scale as a forest pixel, which agrees with the definition of forest by the Food and Agriculture Organization of the United Nations (FAO). The 30 m forest cover data were processed into binary forest maps for 2000 and 2020 and were used to calculate three fragmentation-related landscape metrics based on the global 5000 m grid (see calculation details below). The 5000 m grid was also used to calculate the forest coverage (FC, the percentage of forest area to total area in a particular space) in each grid cell and to generate 5000 m resolution FC maps for 2000 and 2020. Climatic zones The climate zones data were from the world climate regions (WCR) map produced in 2020 (Supplementary Fig. 7a)52, with a spatial resolution of 250 m. We considered global forests to occur in one of four zones: tropical, subtropical, temperate or boreal, which represent the major climatic zones of relevance to forest distribution and spatial patterns of forest fragmentation. Digital elevation model (DEM) Raster-based global altitude data were obtained from the Global Land One-kilometer Base Elevation (GLOBE) dataset (Supplementary Fig. 7b)53. GLOBE is a global digital elevation model (DEM) with a latitude-longitude grid spacing of 30 arc-seconds. The GLOBE DEM dataset was aggregated to a spatial resolution of 5000 m to match the FFI and FC data, and the continuous DEM data were divided into 12 altitudinal categories (<0, 0–100 m, 100–200 m, 200–300 m, 300–400 m, 400–500 m, 500–600 m, 600–700 m, 700–800 m, 800–900 m, 900–1000 m, and >1000 m) to analyze the relationship between forest fragmentation and altitude. Estimates of global forest fragmentation Global grid extent We developed a series of grids, 5000 m × 5000 m in size, for 2000 and 2020 to cover the global forest area. For each grid cell, the fragmentation-related landscape pattern metrics were calculated based on the forest/non-forest binary maps. In total 3,413,077 and 3,422,375 grid cells were considered for 2000 and 2020, respectively, which mainly covered all forest landscapes worldwide. The total area of forest landscapes was larger than the actual global forest area, because each forest landscape contains a certain proportion of non-forest areas. The overlaid parts of the grids were ultimately used in this study to analyze the changes in forest fragmentation. Forest landscape metrics related to fragmentation Edge effect, isolation effect, and patch size effect were the most important features of forest fragmentation5, and they can be quantified by three landscape pattern metrics, including edge density (ED), patch density (PD) and mean patch area (MPA), respectively. The three landscape pattern metrics were used to assemble a synthetic forest fragmentation index in our study. These metrics were calculated as follows:1 ED=∑k=1meikA*10,000 2 PD=niA*10000*100 3 MPA=meanAREApatchij where eik is the total edge length in meters, ni is the number of patches, A is the total landscape (a grid cell was regarded as a landscape) area in square meters, and AREA [patchij] is the area of each patch in hectares. These three landscape pattern metrics were calculated at the class level (class 0: non-forest, class 1: forest) using the “landscapemetrics” package54 in R software based on the two binary maps described above, and the values were converted into raster layers for subsequent analysis. Static and dynamic forest fragmentation indexes To obtain a comprehensive understanding of the multiple dimensions of fragmentation, we constructed a synthesized forest fragmentation index (FFI) using the normalized single fragmentation metrics in ArcGIS software. During the normalization of ED, PD, and MPA, both the directions of the three metrics in reflecting forest fragmentation and the comparability of the FFI in different years were considered (Supplementary Note 1 in Supplementary Information). We used the difference in the FFI between 2020 and 2000 to construct the dynamic forest fragmentation index (ΔFFI) using the follow equation:4 △FFI=FFI2020−FFI2000 where the range of ∆FFI is −1-1. The negative and positive values of ∆FFI indicated decreased and increased fragmentation, respectively. Spatial patterns of forest fragmentation To evaluate the spatial patterns of the static FFI and ΔFFI values, we compared the FFI2000, FFI2020, and ΔFFI among the different climatic zones and altitudinal categories defined previously. One-way ANOVA was used to test for significant differences in FFI2000, FFI2020, and ΔFFI values, and the LSD test was used for pairwise comparisons (Fig. 1d). Moreover, a generalized additive model (GAM) was used to detect the relationships between ΔFFI and FFI2000 for the different climatic zones (Supplementary Fig. 2). Linear correlations were used to explore the relationships between altitude and FFI2000, FFI2020, and ΔFFI, and the correlation coefficient and the slope of the regression line were both used to assess the strength of these relationships. Spatial patterns of the modes of fragmentation processes Since the ΔFFI was calculated by combining ED, PD and MPA, the directions of change in these metrics can reflect the different processes forest fragmentation can proceed. Therefore, we identified eight modes of forest fragmentation processes (EDupPDupMPAdown, EDdownPDupMPAdown, EDupPDdownMPAdown, EDdownPDdownMPAdown, EDupPDupMPAup, EDdownPDupMPAup, EDupPDdownMPAup, and EDdownPDdownMPAup) by considering all possible combinations of an increase or decrease in the ED, PD and MPA between 2000 and 2020 (Fig. 2). Moreover, to characterize the forest fragmentation processes that occurred in areas where fragmentation decreased or increased, we evaluated the global spatial distributions and composition proportions of the different fragmentation process modes where the ΔFFI values were negative or positive, respectively. To better analyze the processes driving changes in forest fragmentation, we selected three hotspots (western Canada, southern Europe, and central China) where fragmentation was remarkably decreased and three hotspots (the southeastern Amazon, the Congo Basin, and central Siberia) where fragmentation was obviously increased. For each hotspot, we analyzed the composition proportions of the eight modes of fragmentation processes (Fig. 2). We then used a T-test to compare the values of EDnor, PDnor and MPAnor between 2000 and 2020, and the percentages of increase or decrease for each of these values were used to reflect their contributions to the changes in the FFI. Drivers of forest fragmentation processes for the globe and hotspots Explanatory variables Anthropogenic activities, demographic pressure, and natural disturbances are considered as the main drivers of global forest loss39. Considering the relationship between human and forest cover, anthropogenic activities were further divided into agricultural activity and socio-economic intensity. Therefore, factors regarding agricultural activity (mean cropland coverage and cropland coverage change), socio-economic intensity (mean nighttime light and nighttime light change), demographic pressure (mean population density and population density change), and natural disturbance (fire frequency) were adopted in our study to explain the dynamics of FFI for six hotspots and the globe. Agricultural activity and natural disturbance are direct influencing factors for the changes in forest distribution and forest fragmentation, while socio-economic intensity and demographic pressure affect forest cover and forest fragmentation by indirect pathway. For variables of agricultural activity, the mean cropland coverage and cropland coverage change were regarded as two important indicators that represent the magnitude and variation of cropland area during 2000–2020. The 30 m resolution cropland extent maps from the Global Land Cover and Land Use Change dataset55 for 2003 and 2019 were adopted in our study and were processed into cropland coverage data by calculating the ratio of cropland pixel to total pixel in a 5000 m size grid. The mean and the difference values of the cropland coverage between 2003 and 2019, respectively, were directly used as agricultural activity factors. In addition, the 500 m resolution nighttime light data, derived from the global NPP-VIIRS-like nighttime light dataset56, for 2000 and 2020 were directly used to represent the socio-economic intensity factors. Similarly, the mean nighttime light and nighttime light change were two important indicators that represent the magnitude and variation of socio-economic intensity during 2000–2020, respectively, and were calculated by the mean and difference values of the NPP-VIIRS-like nighttime light between 2000 and 2020. Furthermore, the mean population density and the population density change during 2000–2020 reflect the magnitude and variation of demographic pressure, respectively, and were used as demographic pressure factors in this study. The WorldPop global gridded population count datasets57 for 2000 and 2020 with a spatial resolution of 1000 m were firstly aggregated into 5000 m resolution population density data by calculating the mean value for each 5000 m size grid of the two periods. The two demographic pressure variables were obtained by the mean and the difference values of the 5000 m resolution gridded population density layers for 2000 and 2020. Finally, for the natural disturbance variable, fire frequency during 2000–2020 was selected as an important indicator that represented the total fire disturbance. The MODIS monthly global burned area dataset (MCD64A1 version 6) with a spatial resolution of 500 m were used in this study, and the fire frequency was obtained by counting the ratio of burned pixel during 2000–2020 for 5000 m size grids during 2000–2020. The resolution of raster layers of all independent variables was aggregated into 5000 m to match the ΔFFI. Correlation analysis The general linear models were used to detect the relationship between ΔFFI and the seven explanatory factors, and all dependent and independent variables were standardized into the range of 0-1 during statistical analysis. The impact of each explanatory factor on the ΔFFI was quantified by the standardized coefficient estimates and P values from the standardized multiple linear models, and the corresponding confidence intervals were also incorporated to help analyze the drivers of ΔFFI. We performed the general linear models for each hotspot (Fig. 3) and the globe (Supplementary Fig. 4) based on dependent and independent variables in their respective scopes. In addition, we identified the major driver of ΔFFI, represented by the factor with the highest absolute value of coefficient estimates, for a series of 50 km × 50 km grids and finally generated the ΔFFI major driver map (Supplementary Fig. 5) with a spatial resolution of 50 km at the global scale. Two-dimensional framework for the assessment of forest landscape dynamics We constructed a two-dimensional assessment framework of global forest landscape dynamics in which all forest landscapes were categorized into four types (FCupFFIdown, FCupFFIup, FCdownFFIdown, and FCdownFFIup) based on increases (positive values) or decreases (negative values) in FC and FFI from 2000 to 2020. These four types of forest landscape dynamic patterns represent forest landscapes in the deep recovery stage, early recovery stage, early degradation stage and deep degradation stage, respectively. We mapped the spatial distributions of these different forest landscape dynamic patterns (Fig. 3a) by overlaying the ΔFFI layer and the ΔFC layer at 5000 m resolution. We also evaluated the variation in the relative percentages of the four patterns across climatic zones (Fig. 3b) and the altitudinal gradient (Supplementary Fig. 8). Furthermore, we compared the mean values of ΔFFI and ΔFC among all countries worldwide (Supplementary Data 1) and used Pearson’s linear correlations to fit the relationship (Fig. 3c). Specifically, to evaluate how changes in forest cover and fragmentation between 2000 and 2020 varied among the ten countries with the largest forest area worldwide, we calculated the area percentages representing each pattern and the mean values of ΔFC and ΔFFI (Table 1). The administrative boundaries of each country were determined by the FAO (https://data.apps.fao.org/map/catalog/static/search?format=shapefile). Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Peer review file Description of additional supplementary files Supplementary Data 1 Reporting Summary Supplementary information The online version contains supplementary material available at 10.1038/s41467-023-39221-x. Acknowledgements This research was supported by the National Key Research and Development Program of China (2022YFF0802400, J.M.) and the Natural Science Foundation of China (32271659, J.M. and U2106209, J.M.). Author contributions J.M. and J.J.L. conceived the idea and designed the methodology; J.M., J.W.L., and W.B.W. conducted the data analysis; and J.M. wrote the manuscript with contributions from J.W.L., and J.J.L. All authors contributed critically to the interpretation of the results and gave final approval for publication. Peer review Peer review information Nature Communications thanks Victor Danneyrolles, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Data availability All data used in the analysis are publicly accessible. The global land cover and land use change dataset is available at https://glad.umd.edu/dataset/GLCLUC2020 (including forest cover data and cropland coverage data); The global climate zones dataset is available at https://storymaps.arcgis.com/stories/61a5d4e9494f46c2b520a984b2398f3b; The global altitude dataset is available at https://ngdc.noaa.gov/mgg/topo/gltiles.html; The global NPP-VIIRS-like nighttime light dataset is available at https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/YGIVCD; The worldPop global gridded population count dataset is available at https://hub.worldpop.org/project/categories?id=3; The MODIS monthly global burned area dataset is available at https://lpdaac.usgs.gov/products/mcd64a1v006/. The global administrative boundary dataset is available at https://data.apps.fao.org/map/catalog/static/search?format=shapefile. The Forest Fragmentation Index (FFI) data generated in this study have been deposited in the Figshare repository at https://figshare.com/s/21dbf1f50250aeb7f5a0. Code availability The code used to calculate the landscape pattern index in this study can be found in the Figshare repository at https://figshare.com/s/21dbf1f50250aeb7f5a0. Competing interests The authors declare no competing interests. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Jun Ma, Jiawei Li. ==== Refs References 1. Krauss J Habitat fragmentation causes immediate and time-delayed biodiversity loss at different trophic levels Ecol. Lett. 2010 13 597 605 10.1111/j.1461-0248.2010.01457.x 20337698 2. Pfeifer M Creation of forest edges has a global impact on forest vertebrates Nature 2017 551 187 191 10.1038/nature24457 29088701 3. Brinck K High resolution analysis of tropical forest fragmentation and its impact on the global carbon cycle Nat. Commun. 2017 8 14855 10.1038/ncomms14855 28303883 4. Pimm SL The forest fragment classic Nature 1998 393 23 24 10.1038/29892 5. Haddad NM Habitat fragmentation and its lasting impact on Earth’s ecosystems Sci. Adv. 2015 1 e1500052 10.1126/sciadv.1500052 26601154 6. Taubert F Global patterns of tropical forest fragmentation Nature 2018 554 519 522 10.1038/nature25508 29443966 7. Liu J Forest fragmentation in China and its effect on biodiversity Biol. Rev. 2019 94 1636 1657 10.1111/brv.12519 31058438 8. Wade TG Riitters KH Wickham JD Jones KB Distribution and causes of global forest fragmentation Conserv. Ecol. 2003 7 7 9. Morreale LL Thompson JR Tang X Reinmann AB Hutyra LR Elevated growth and biomass along temperate forest edges Nat. Commun. 2021 12 7181 10.1038/s41467-021-27373-7 34893596 10. Gibbs HK Tropical forests were the primary sources of new agricultural land in the 1980s and 1990s Proc. Natl. Acad. Sci. USA 2010 107 16732 16737 10.1073/pnas.0910275107 20807750 11. Hansen MC The fate of tropical forest fragments Sci. Adv. 2020 6 eaax8574 10.1126/sciadv.aax8574 32195340 12. Fischer R Accelerated forest fragmentation leads to critical increase in tropical forest edge area Sci. Adv. 2021 7 eabg7012 10.1126/sciadv.abg7012 34516875 13. Song X-P Global land change from 1982 to 2016 Nature 2018 560 639 643 10.1038/s41586-018-0411-9 30089903 14. Hoang NT Kanemoto K Mapping the deforestation footprint of nations reveals growing threat to tropical forests Nat. Ecol. Evol. 2021 5 845 853 10.1038/s41559-021-01417-z 33782576 15. Fahrig L Effects of habitat fragmentation on biodiversity Annu. Rev. Ecol. Evol. Syst. 2003 34 487 515 10.1146/annurev.ecolsys.34.011802.132419 16. Vellend M Extinction debt of forest plants persists for more than a century following habitat fragmentation Ecology 2006 87 542 548 10.1890/05-1182 16602283 17. Laurance WF Ecosystem decay of Amazonian forest fragments: A 22-year investigation Conserv. Biol. 2002 16 605 618 10.1046/j.1523-1739.2002.01025.x 18. Gibson L Near-complete extinction of native small mammal Fauna 25 years after forest fragmentation Science 2013 341 1508 1510 10.1126/science.1240495 24072921 19. Chen C China and India lead in greening of the world through land-use management Nat. Sustain. 2019 2 122 129 10.1038/s41893-019-0220-7 30778399 20. Chaplin-Kramer R Degradation in carbon stocks near tropical forest edges Nat. Commun. 2015 6 10158 10.1038/ncomms10158 26679749 21. Grantham HS Anthropogenic modification of forests means only 40% of remaining forests have high ecosystem integrity Nat. Commun. 2020 11 5978 10.1038/s41467-020-19493-3 33293507 22. Raven PH The distribution of biodiversity richness in the tropics Sci. Adv. 2020 6 eabc6228 10.1126/sciadv.abc6228 32917691 23. Laurance WF An Amazonian rainforest and its fragments as a laboratory of global change Biol. Rev. 2018 93 223 247 10.1111/brv.12343 28560765 24. Bastin J-F The global tree restoration potential Science 2019 365 76 79 10.1126/science.aax0848 31273120 25. Alkama R Cescatti A Biophysical climate impacts of recent changes in global forest cover Science 2016 351 600 604 10.1126/science.aac8083 26912702 26. Geldmann J Manica A Burgess ND Coad L Balmford A A global-level assessment of the effectiveness of protected areas at resisting anthropogenic pressures Proc. Natl Acad. Sci. USA 2019 116 23209 23215 10.1073/pnas.1908221116 31659036 27. Agnoletti M Rural landscape, nature conservation and culture: Some notes on research trends and management approaches from a (southern) European perspective Landsc. Urban Plan. 2014 126 66 73 10.1016/j.landurbplan.2014.02.012 28. Lawler JJ Projected land-use change impacts on ecosystem services in the United States Proc. Natl Acad. Sci. USA 2014 111 7492 7497 10.1073/pnas.1405557111 24799685 29. Qin Y Carbon loss from forest degradation exceeds that from deforestation in the Brazilian Amazon Nat. Clim. Chang. 2021 11 442 448 10.1038/s41558-021-01026-5 30. Trondoli Matricardi EA Long-term forest degradation surpasses deforestation in the Brazilian Amazon Science 2020 369 1378 1382 10.1126/science.abb3021 32913104 31. Tong X Forest management in southern China generates short term extensive carbon sequestration Nat. Commun. 2020 11 1 10 10.1038/s41467-019-13798-8 31911652 32. Brandt M Satellite-observed major greening and biomass increase in South China Karst during recent decade Earth’s Future 2018 6 1017 1028 10.1029/2018EF000890 33. Ellis EC People have shaped most of terrestrial nature for at least 12,000 years Proc. Natl Acad. Sci. USA 2021 118 e2023483118 10.1073/pnas.2023483118 33875599 34. Achard F Determination of tropical deforestation rates and related carbon losses from 1990 to 2010 Glob. Change Biol. 2014 20 2540 2554 10.1111/gcb.12605 35. Sloan S Sayer JA Forest Resources Assessment of 2015 shows positive global trends but forest loss and degradation persist in poor tropical countries For. Ecol. Manage. 2015 352 134 145 10.1016/j.foreco.2015.06.013 36. Piao S Characteristics, drivers and feedbacks of global greening Nat. Rev. Earth Environ. 2020 1 14 27 10.1038/s43017-019-0001-x 37. Jiang Z Lian Y Qin X Rocky desertification in Southwest China: Impacts, causes, and restoration Earth-Sci. Rev. 2014 132 1 12 10.1016/j.earscirev.2014.01.005 38. Deng L Past and future carbon sequestration benefits of China’s grain for green program Glob. Environ. Change-Human Policy Dimens. 2017 47 13 20 10.1016/j.gloenvcha.2017.09.006 39. Curtis PG Slay CM Harris NL Tyukavina A Hansen MC Classifying drivers of global forest loss Science 2018 361 1108 1111 10.1126/science.aau3445 30213911 40. Feng Y Upward expansion and acceleration of forest clearance in the mountains of Southeast Asia Nat. Sustain. 2021 4 892 899 10.1038/s41893-021-00738-y 41. Pendrill F Persson UM Godar J Kastner T Deforestation displaced: trade in forest-risk commodities and the prospects for a global forest transition Environ. Res. Lett. 2019 14 055003 10.1088/1748-9326/ab0d41 42. Grogan K Pflugmacher D Hostert P Mertz O Fensholt R Unravelling the link between global rubber price and tropical deforestation in Cambodia Nat. Plants 2019 5 47 53 10.1038/s41477-018-0325-4 30598534 43. Lu F Effects of national ecological restoration projects on carbon sequestration in China from 2001 to 2010 Proc. Natl Acad. Sci. USA 2018 115 4039 4044 10.1073/pnas.1700294115 29666317 44. Tchebakova NM Parfenova EI Korets MA Conard SG Potential change in forest types and stand heights in central Siberia in a warming climate Environ. Res. Lett. 2016 11 035016 10.1088/1748-9326/11/3/035016 45. Barlow J Anthropogenic disturbance in tropical forests can double biodiversity loss from deforestation Nature 2016 535 144 147 10.1038/nature18326 27362236 46. Tyukavina A Types and rates of forest disturbance in Brazilian Legal Amazon, 2000-2013 Sci. Adv. 2017 3 e1601047 10.1126/sciadv.1601047 28439536 47. van Wees D The role of fire in global forest loss dynamics Glob. Change Biol. 2021 27 2377 2391 10.1111/gcb.15591 48. Tyukavina A Global trends of forest loss due to fire from 2001 to 2019 Front. Remote Sens. 2022 3 825190 10.3389/frsen.2022.825190 49. Lewis SL Edwards DP Galbraith D Increasing human dominance of tropical forests Science 2015 349 827 832 10.1126/science.aaa9932 26293955 50. Potapov P The last frontiers of wilderness: tracking loss of intact forest landscapes from 2000 to 2013 Sci. Adv. 2017 3 e1600821 10.1126/sciadv.1600821 28097216 51. Potapov P The global 2000-2020 land cover and land use change dataset derived from the landsat archive: first results Front. Remote Sens. 2022 3 856903 10.3389/frsen.2022.856903 52. Sayre R An assessment of the representation of ecosystems in global protected areas using new maps of World Climate Regions and World Ecosystems Glob. Ecol. Conserv. 2020 21 e00860 10.1016/j.gecco.2019.e00860 53. NOAA National Geophysical Data Center. Global land one-kilometer base elevation (GLOBE) v.1. 10.7289/V52R3PMS (1999). 54. Hesselbarth MHK Sciaini M With KA Wiegand K Nowosad J landscapemetrics: an open-source R tool to calculate landscape metrics Ecography 2019 42 1648 1657 10.1111/ecog.04617 55. Potapov P Global maps of cropland extent and change show accelerated cropland expansion in the twenty-first century Nat. Food 2022 3 19 28 10.1038/s43016-021-00429-z 37118483 56. Chen, Z. et al. An extended time series (2000-2018) of global NPP-VIIRS-like nighttime light data from a cross-sensor calibration. Earth Syst. Sci. Data 13, 889–906 (2021). 57. Lloyd CT Global spatio-temporally harmonised datasets for producing high-resolution gridded population distribution datasets Big Earth Data 2019 3 108 139 10.1080/20964471.2019.1625151 31565697

PMC010xxxxxx/PMC10338005.txt

==== Front 100971395 22591 J Fluid Mech J Fluid Mech Journal of fluid mechanics 0022-1120 37441053 10.1017/jfm.2022.799 nihpa1901731 Article Buoyancy-modulated Lagrangian drift in wavy-walled vertical channels as a model problem to understand drug dispersion in the spinal canal http://orcid.org/0000-0001-5346-9273 Alaminos-Quesada J. 1 http://orcid.org/0000-0002-2693-6574 Coenen W. 2 http://orcid.org/0000-0003-1123-2002 Gutiérrez-Montes C. 3 http://orcid.org/0000-0003-1349-9843 Sánchez A.L. 1† 1 Department of Mechanical and Aerospace Engineering, University of California San Diego, La Jolla, CA 92093-0411, USA 2 Grupo de Mecánica de Fluidos, Universidad Carlos III de Madrid, Leganés 28911, Spain 3 Andalusian Institute for Earth System Research, University of Jaén, Jaén 23071, Spain † correspondence: als@ucsd.edu 18 5 2023 25 10 2022 06 10 2022 12 7 2023 949 A48https://creativecommons.org/licenses/by/4.0/ This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited. This paper investigates flow and transport in a slender wavy-walled vertical channel subject to a prescribed oscillatory pressure difference between its ends. When the ratio of the stroke length of the pulsatile flow to the channel wavelength is small, the resulting flow velocity is known to include a slow steady-streaming component resulting from the effect of the convective acceleration. Our study considers the additional effect of gravitational forces in configurations with a non-uniform density distribution. Specific attention is given to the slowly evolving buoyancy-modulated flow emerging after the deposition of a finite amount of solute whose density is different from that of the fluid contained in the channel, a relevant problem in connection with drug dispersion in intrathecal drug delivery (ITDD) processes, involving the injection of the drug into the cerebrospinal fluid that fills the spinal canal. It is shown that when the Richardson number is of order unity, the relevant limit in ITDD applications, the resulting buoyancy-induced velocities are comparable to those of steady streaming. As a consequence, the slow time-averaged Lagrangian motion of the fluid, involving the sum of the Stokes drift and the time-averaged Eulerian velocity, is intimately coupled with the transport of the solute, resulting in a slowly evolving problem that can be treated with two-time-scale methods. The asymptotic development leads to a time-averaged, nonlinear integro-differential transport equation that describes the slow dispersion of the solute, thereby circumventing the need to describe the small concentration fluctuations associated with the fast oscillatory motion. The ideas presented here can find application in developing reduced models for future quantitative analyses of drug dispersion in the spinal canal. biomedical flows ==== Body pmc1. Introduction The steady Lagrangian drift generated in oscillatory viscous flows in pipes and channels is known to play an important role in different heat and mass transport processes, including those occurring in extracorporeal membrane oxygenators (Bellhouse et al. 1973), pulmonary high-frequency ventilation devices (Grotberg 1994), compact heat exchangers (Mackley & Stonestreet 1995) and drug dispersion in the spinal canal (Lawrence et al. 2019). For configurations with slowly varying cross-section, the lubrication approximation can be used to derive insightful analytical results, with seminal analyses including those of Hall (1974), who considered flow in a pipe subject to a harmonic pressure difference, and Grotberg (1984), who investigated flow in a tapered channel subject to a prescribed oscillating stroke volume. More recent analytical studies pertaining to channels include those of Lo Jacono, Plouraboué & Bergeon (2005) and Guibert, Plouraboué & Bergeon (2010), involving three-dimensional wavy-walled configurations, and that of Larrieu, Hinch & Charru (2009), who considered an oscillating Couette flow over a wavy bottom. All of these analytical investigations of oscillating slender flows addressed configurations with weak inertia, corresponding to small values of the ratio ε of the stroke length to the characteristic longitudinal length, with ε−1≫1 representing the relevant Strouhal number. The asymptotic analysis for ε≪1 reveals that the velocity, resulting from a balance between the local acceleration and the pressure and viscous forces, is harmonic at leading order, with the small corrections arising from the convective acceleration providing a small steady-streaming component of order ε (Riley 2001). This steady streaming determines, together with the Stokes drift associated with the leading-order harmonic flow, the Lagrangian drift experienced by the fluid particles, with both contributions having in general comparable orders of magnitude (Larrieu et al. 2009). As discussed by Guibert et al. (2010), the fundamental pulsatile-flow investigations mentioned previously are relevant in connection with the motion of cerebrospinal fluid (CSF) along the spinal subarachnoid space, a slender annular canal surrounding the spinal cord, depicted in figure 1. The flow features an oscillatory velocity driven by the pressure pulsations induced by the cardiac and respiratory cycles (Linninger et al. 2016). The dynamics of this flow and its associated Lagrangian transport are fundamental in understanding the role of CSF as a vehicle for metabolic-waste clearance (Klarica, Radoš & Orešković 2019) and also to quantify drug dispersion in intrathecal drug delivery (ITDD) (Onofrio 1981; Lynch 2014; Fowler et al. 2020), a medical procedure used for treatment of some cancers (Lee et al. 2017), infections (Remeš et al. 2013) and pain (Bottros & Christo 2014). The standard ITDD protocol involves the placement of a small catheter along the lumbar section of the spinal canal to continuously pump the drug or to release a finite dose at selected times. The transport of the drug depends fundamentally on its physical properties, including molecular diffusivities κ that are much smaller than the kinematic viscosity v of CSF. The flow in the spinal canal has been investigated computationally in numerous studies addressing different aspects of the problem, as summarised in a recent paper by Khani et al. (2018), although corresponding theoretical analyses are more scarce. Exact solutions for pulsatile viscous flow in a straight elliptic annulus have been proposed as a representation for the oscillatory flow in the spinal canal (Gupta, Poulikakos & Kurtcuoglu 2008). More recent studies, modelling the canal as a linearly elastic annular pipe of slowly varying section, have employed the lubrication approximation in the asymptotic limit ε≪1 to quantify steady streaming (Sánchez et al. 2018) and to investigate the dispersion of a solute (Lawrence et al. 2019). For the large values of the Schmidt number v/κ~1000 corresponding to the molecular diffusivities of all ITDD drugs, the analysis of Lawrence et al. (2019) showed that the Lagrangian mean flow is the key mechanism responsible for the dispersion of the solute, whereas shear-enhanced Taylor dispersion has a negligibly small effect. An important outcome of the asymptotic analysis is a time-averaged transport equation that has been recently validated by means of comparisons with results of direct numerical simulations (DNS) spanning hundreds of oscillation cycles (Gutiérrez-Montes et al. 2021), as needed to generate significant dispersion of the solute. The comparisons demonstrate the accuracy of the reduced description, which is seen to provide excellent fidelity at a fraction of the computational cost involved in the DNS. Our previous analysis of solute dispersion (Lawrence et al. 2019) assumed the density of the solute ρs and the density of the carrier fluid ρ to be identical, thereby neglecting the small density differences ρ−ρS found in ITDD applications, for which the values of ρ−ρs typically range from positive values of order ρ−ρs~ρ/1000 for drugs diluted with water to negative values of order ρ−ρs/~−ρ/100 for drugs diluted with dextrose (Nicol & Holdcroft 1992; Lui, Polis & Cicutti 1998). These relative density differences ρ−ρs/ρ≪1 between the drug and the CSF, although extremely small, are known by clinicians to play an important role in the dispersion rate of ITDD drugs for patients in a sitting position, when buoyancy forces are nearly aligned with the canal. It has been seen that for a hyperbaric (dense) solution, injection while the patient is seated for some time before moving to a supine position leads to an initial restriction in the transport of the anaesthesia (Mitchell et al. 1988; Povey, Jacobsen & Westergaard-Nielsen 1989; Veering et al. 2001). (In spinal anaesthesia, baricity is the term used to refer to the density of the anaesthetic relative to that of the CSF. Thus, an anaesthetic is said to be hyperbaric/hypobaric when its density is higher/lower than that of the CSF, whereas the term isobaric describes anaesthetics whose density matches exactly that of the CSF.) Conversely, for a hypobaric (light) solution, the sitting injection position leads to more rapid cephalad spread of the anaesthesia as compared to a lateral injection position (Richardson et al. 1996). As could be expected, the density of the drug is inconsequential when injection occurs in the lateral position (Hallworth et al. 2005) and, similarly, positioning has no effect on the spread rate when the solution density matches that of CSF (Wildsmith et al. 1981). Given the abundance of clinical evidence on the importance of buoyancy forces on the drug dispersion rate, there is interest in developing a quantitative description; the present paper, focused on a simplified geometry, is a necessary first step in that direction. In looking for a simplified geometrical model, we follow Guibert et al. (2010) in noting that the width ho~1−2mm of the annular spinal canal is smaller than the spinal-cord diameter ~1cm, with the consequence that a two-dimensional channel can be used to describe many aspects of the flow. The channel is placed in a vertical position, as is appropriate in describing buoyancy effects for a patient in a sitting position. As indicated in figure 1, the quasi-periodic variation of the canal section, associated with the presence of the vertebrae, will be modelled by including a wavy boundary whose wavelength λ mimics the inter-vertebral distance. The channel will be assumed to be slender in that λ≫ho, a good approximation in the spinal canal, where λ~2−4cm and ho/λ≃0.05. For simplicity, the total channel length is taken to be an integer multiple of the wavelength, so that the channel contains a finite number of identical cells. As in the seminal analysis of Hall (1974), an oscillating pressure difference with angular frequency ω will be imposed between the channel ends, resulting in a pulsating flow. We investigate the buoyancy-modulated dispersion of a bolus of solute released inside the channel when the buoyancy-induced acceleration is comparable to the convective acceleration of the pressure-driven flow, those being the conditions of interest in ITDD applications, as explained later below (2.6). The rest of the paper is organised as follows. The problem is formulated in dimensionless form in §2, which includes the identification of the relevant non-dimensional parameters and a discussion of the essential features of the subsequent asymptotic analysis, including the existence of a long time scale ε−2ω−1 for solute dispersion, additional to the much smaller oscillation time ω−1. The asymptotic description of the velocity field is presented next in §3, with the time-averaged Eulerian velocity including the familiar steady-streaming contribution stemming from the convective acceleration along with an additional buoyancy-induced component that depends on the distribution of solute. This velocity field is used in §4 to analyse solute dispersion with use of a two-time scale asymptotic analysis, resulting in a time-averaged transport equation that describes the evolution of the flow in the long-time scale ε−2ω−1. The reduced description stemming from the asymptotic analysis is validated in §5 through comparisons with DNS. In addition, the model is used to quantify effects of buoyancy-induced motion on the solute dispersion for different values of the controlling parameters. Finally, concluding remarks are given in §6. 2. Problem formulation 2.1. Governing equations Consider a vertical wavy channel of average gap size ho filled with a Newtonian fluid of density ρ and kinematic viscosity v (for CSF, ρ≃103kgm−3 and v≃0.7×10−6m2s−1). The channel, open at both ends, is bounded by a flat surface and a wavy wall of wave length λ≫ho, so that the resulting channel width is h=ho1+βcos 2πx*/λ, where x* is the longitudinal distance measured from the upper end and β<1 is the relative amplitude of the wall undulation. The total channel length is nλ, with n representing a general integer number, so that the channel comprises n identical cells. The flow is described using cartesian coordinates x*,y*, with y* measured from the flat surface, and corresponding velocity components v*=u*,v*. The Navier–Stokes equations describing the planar unsteady flow are written in the Boussinesq approximation (2.1) ∇*⋅v*=0, (2.2) ∂v*∂t*+v*⋅∇*v*=−1ρ∇*p*+ν∇*2v*−ρ−ρsρgcex, (2.3) ∂c∂t*+v*⋅∇*c=κ∇*2c, where p* is the sum of the pressure difference from the upper end and the constant-density hydrostatic component −ρgx*,c is the solute volume concentration, κ is the solute diffusivity, ∇*=∂/∂x*,∂/∂y* and ex is the unit vector aligned with the gravitational acceleration. A pressure difference nΔpcos ωt* oscillating harmonically in time is prescribed between the upper and lower ends of the canal, driving a periodic fluid motion with angular frequency ω. The resulting slender flow is characterised by longitudinal velocities of order uc=Δp/(ρωλ), as follows from a balance between the local acceleration ∂u*/∂t*~ucω and the pressure gradient ρ−1∂p*/∂x*~Δp/(λρ), and much smaller transverse velocities of order vc=ho/λuc≪uc, as follows from the continuity balance ∂u*/∂x*~∂v*/∂y*. 2.2. Controlling parameters The analysis assumes that the viscous time across the channel ho2/v is comparable to the characteristic oscillation time ω−1, resulting in Womersley numbers (2.4) α=ωho2v1/2, of order unity. The limit α~1 is instrumental in analysing cardiac-driven CSF flow ω=2πs−1 in the spinal canal, for which typical values of α are in the range 3≲α≲6, as can be seen by evaluating the above expression with ho≃1−2mm and v=0.7×10−6m2s−1. In the lumbar region, the typical drug-delivery site in ITDD procedures, the average CSF speeds are of order uc~1cms−1, so that the associated stroke lengths uc/ω are much smaller than the characteristic longitudinal distance ≃2−4cm. Their ratio (2.5) ε=uc/ωλ, of order ε≃0.05 for spinal CSF flow, defines the small parameter employed in the following asymptotic description. As shown earlier (Hall 1974; Grotberg 1984), the solution at leading order is determined by a balance between the pressure gradient, the local acceleration and the viscous forces, with the convective acceleration introducing small corrections of relative magnitude ε. Although the leading-order motion is harmonic, the velocity corrections include a non-zero steady-streaming component. The familiar periodic channel flow described previously is altered by gravitational forces when a solute of density ρs≠ρ is introduced in the channel. The extent of the resulting buoyancy-induced motion can be measured by the associated Richardson number (2.6) Ri=ρ−ρsρgλuc2, which compares the order of magnitude of the effective gravitational acceleration g(ρ−ρs/ρ with that of the convective acceleration v*⋅∇*v*~uc2/λ. Our analysis addresses the limit Ri~1, which is relevant for drug dispersion in ITDD procedures, as can be seen by evaluating (2.6) with λ≃2cm and uc~1cms−1 for density differences in the range 10−3≲ρ−ρs/ρ≲10−2. Also motivated by ITDD applications, we consider solutes with diffusivities κ much smaller than the kinematic viscosity, that always being the case of diffusion in liquid phase. As v/κ~1000 and ε≃0.05 in ITDD applications, the following analysis of solute dispersion will specifically address the distinguished limit κ/ν~ε2, with solute diffusion correspondingly characterised by the reduced Schmidt number (2.7) σ=ε2νκ, assumed to be of order unity. 2.3. Non-dimensional formulation We address the motion that follows from the deposition of the solute inside an intermediate cell along the channel. The problem is non-dimensionalised using the scales identified previously to give the dimensionless variables (2.8a–f) t=ωt*,x=x*λ,y=y*ho,u=u*uc,v=v*vc,p=p*Δp and associated conservation equations (2.9) ∂u∂x+∂v∂y=0, (2.10) ∂u∂t+εv⋅∇u=−∂p∂x+1α2∂2u∂y2−εRic, (2.11) ∂p∂y=0, (2.12) ∂c∂t+εv⋅∇c=ε2α2σ∂2c∂y2, where v=(u,v) and ∇=(∂/∂x,∂/∂y). In writing (2.9)–(2.12) from (2.1)–(2.3) we have used the slender-flow approximation resulting from the limit ho≪λ. Thus, the terms representing longitudinal diffusion of momentum and mass have been neglected in (2.10) and (2.12), because they are a factor ho/λ2 smaller than those associated with transverse diffusion. At the same level of approximation, the transverse component of the momentum equation takes the reduced form (2.11). The velocity and concentration must satisfy the boundary conditions (2.13) u=v=∂c∂y=0 at y=0y=H=1+βcos (2πx), corresponding to non-permeable no-slip surfaces, whereas the reduced pressure p(x,t), independent of y, is identically zero at x=0 and takes the value p=ncos t at the lower end x=n. In the absence of buoyancy (i.e. for Ri=0), the solution for the velocity is periodic in time, including a steady component of order ε, and also periodic in space, so that the velocity distribution found in each cell is identical. On the other hand, for Ri≠0 the motion is coupled to the solute transport, albeit weakly, with the result that the velocity necessarily evolves in time following the dispersion of the solute, which is driven partly by the steady streaming motion, with characteristic velocities εuc. It can be anticipated that the characteristic time for the slow evolution is that associated with the dispersion of the solute inside the deposition cell λ/εuc=ε−2ω−1, much larger than the characteristic oscillation time ω−1. These considerations suggest the introduction of a second time variable τ=ε2t for describing the slow evolution, additional to the fast time-scale variable t describing the oscillatory motion. In the two-time-scale formalism, the time derivatives in (2.10) and (2.12) are replaced with ∂/∂t+ε2∂/∂τ and the different variables are expressed in terms of the power expansions (2.14) u=u0(x,y,t,τ)+εu1(x,y,t,τ)+⋯, (2.15) v=v0(x,y,t,τ)+εv1(x,y,t,τ)+⋯, (2.16) p=p0(x,t,τ)+εp1(x,t,τ)+⋯, (2.17) c=c0(x,y,t,τ)+εc1(x,y,t,τ)+⋯, with all functions assumed to be 2π periodic in the fast time scale t. The asymptotic procedure leads to a hierarchy of problems that can be solved sequentially, as shown in the following. 3. Velocity description We begin by describing the velocity field in the asymptotic limit ε≪1, following the procedure used in previous steady-streaming investigations of slender flows (Larrieu et al. 2009; Guibert et al. 2010; Sánchez et al. 2018). The solution at leading order and also the first-order corrections associated with convective acceleration are similar to those found earlier in three-dimensional wavy-walled channels (Guibert et al. 2010) and annular canals (Sánchez et al. 2018). These previous analyses did not address, however, effects of buoyancy forces, which are investigated here for order-unity values of the Richardson number Ri, leading to a velocity correction that will be seen to be expressible in terms of integrals of the solute concentration. 3.1. Leading-order oscillatory flow Convective acceleration and buoyancy are negligible at leading order, so that the velocity v0=u0,v0 satisfies a linear problem that can be solved in terms of the reduced variables (3.1 a–c) u0=Re ieitU,v0=Re ieitV,p0=Re eitP, where the complex functions U(x,y),V(x,y), and P(x) satisfy (3.2a,b) ∂U∂x+∂V∂y=0 and −U=−dPdx+iα2∂2U∂y2. The second equation above can be integrated with boundary conditions U=0 at y=0,H to give (3.3) U=dPdxG(x,y), where (3.4a,b) G=1−cosh [Λ(2y/H−1)]cosh Λ and Λ=α21+i2H(x). The result can be used to integrate the first equation in (3.2a,b) subject to V=0 at y=0, yielding (3.5) V=−∂∂x∫0yUdyˆ=−∂∂xdPdx∫0yGdyˆ, where (3.6) ∫0yGdyˆ=y−H2Λsinh [Λ(2y/H−1)]+sinh Λcosh Λ, with yˆ representing a dummy integration variable. Note that both velocity components U and V are spatially periodic in x through the function H=1+βcos (2πx). The determination of the longitudinal pressure gradient dP/dx that completes the solution begins by using the condition V=0 at y=H in the first equation of (3.5) to give (3.7) ddx∫0HUdy=0, indicating that the reduced flow rate (3.8) Q=∫0HUdy=dPdx∫0HGdy is constant. Further progress requires use of the conditions P(0)=P(n)−n=0, consistent with the boundary values p(0,t)=0 and p(n,t)=ncos t stated below (2.13). Using (3.6) to evaluate the integral ∫0HGdy leads to the equation (3.9) Q=dPdxH1−Λ−1tanh Λ, which can be integrated subject to P(0)=0 to yield the pressure distribution (3.10) P=Q∫0xdxˆH1−Λ−1tanh Λ. Using now the condition P(n)=n provides (3.11) Q=n∫0ndxH1−Λ−1tanh Λ−1. Owing to the spatial periodicity of H and Λ, it follows that (3.12) ∫0ndxH1−Λ−1tanh Λ=n∫01dxH1−Λ−1tanh Λ, thereby finally yielding (3.13) Q=∫01dxH1−Λ−1tanh Λ−1 and, from (3.9), (3.14) dPdx=H1−Λ−1tanh Λ∫01dxH1−Λ−1tanh Λ−1, independent of n. It is worth pointing out that, because at this order the velocity is harmonic, the associated time-averaged values u0 and v0 with ⟨⋅⟩=(1/2π)∫tt+2πdt are identically zero, so that the steady bulk motion of the fluid occurs through the velocity corrections at the following order. 3.2. First-order corrections Collecting terms of order ε in (2.9) and (2.10) yields (3.15) ∂u1∂x+∂v1∂y=0, (3.16) ∂u1∂t+∂∂xu02+∂∂yu0v0=−∂p1∂x+1α2∂2u1∂y2−Ric0, to be integrated with boundary conditions u1=v1=0 at y=0,H and p1=0 at x=0,n. There is interest in computing the corresponding time-averaged velocity correction v1=u1,v1. Taking the time average of (3.15) and (3.16) provides (3.17 a,b) ∂u1∂x+∂v1∂y=0 and F(x,y)=−∂p1∂x+1α2∂2u1∂y2−Ric0, where the known function F=∂u02/∂x+∂u0v0/∂y can be expressed in terms of the complex velocities U and V defined above in the form (3.18) F=12Re ∂∂x(UU‾)+∂∂y(VU‾), a result following from the identity Re ieitf1Re ieitf2=Re f1f‾2/2, which applies to any generic time-independent complex functions f1 and f2, with the bar denoting complex conjugates. In writing (3.17a,b), we have anticipated that, at leading order, the solute concentration is independent of the fast time scale t, as follows from (2.12) when ε≪1, so that its time-averaged value c0 reduces simply to c0=c0. Also of interest is that, because of the symmetry of H(x), the periodic function F defined in (3.18) is antisymmetric with respect to x=1/2, so that F(x,y)=−F(1−x,y). As can be concluded from (3.16), the velocity corrections arise partly owing to the convective acceleration and partly owing to the buoyancy force. In computing the corresponding time average, it is convenient to compute both contributions separately by introducing v1=vSS+vB and p1=pSS+pB, where vSS(x,y)=uSS,vSS and pSS(x) describe the familiar steady-streaming associated with the nonlinear convective terms, which is independent of time and periodic in x, and vB(x,y,τ)=uB,vB and pB(x,τ) describe the buoyancy-induced corrections, which evolve in the long time scale τ as the solute spreads in the channel. 3.3. Steady streaming The solution procedure needed to compute the velocity corrections parallels that followed earlier at leading order. Thus, the longitudinal component of the steady-streaming velocity (3.19) uSSα2=−dppSSdx12(H−y)y+y∫0yFdyˆ−∫0yFyˆdyˆ−y∫0HF1−yHdy is determined by integrating the momentum equation (3.16) written for Ri=0 subject to the boundary conditions uSS=0 at y=0,H. The result can be substituted into (3.15) to give (3.20) vSSα2=∂∂x[dpSSdxy22(H2−y3)+y22∫HyF(1−yˆH)dyˆ+y(1−y2H)∫y0Fyˆdyˆ−12∫y0Fyˆ2dyˆ] upon integrating with vSS=0 at y=0. To determine the unknown pressure gradient dpSS/dx we begin by using vSS=0 at y=H in the above equation to give (3.21) ddx∫0HuSSα2dy=ddxdpSSdxH312+12∫0HFy(H−y)dy=0, which indicates that the flow rate (3.22) QSS=∫0HuSSdy=α2dpSSdxH312+12∫0HFy(H−y)dy is constant. Its value can be determined by integrating a second time (3.22) subject to pSS(n)=pSS(0)=0 to yield (3.23) QSS=α2∫01H−3∫0HFy(H−y)dydx2∫01H−3dx, when account is taken of the spatial periodicity of H and F. Since H(x) is symmetric about x=1/2 while F(x,y) is antisymmetric, the double integral in the numerator of the above equation is identically zero, so that (3.24) QSS=∫0HuSSdy=0, in agreement with previous findings regarding steady streaming in tubes (Hall 1974) and channels (Lo Jacono et al. 2005; Guibert et al. 2010). Using the condition QSS=0 in (3.22) finally yields (3.25) dpSSdx=−6H−3∫0HFy(H−y)dy, for the pressure gradient, thereby completing the solution. 3.4. Buoyancy-induced velocity The corresponding solution for the buoyancy-induced velocity can be obtained by simply replacing F(x,y) with Ri c0(x,y,τ) in (3.19) and (3.20), yielding (3.26) uBα2Ri=−1Ri∂pB∂x12(H−y)y+y∫0yc0dyˆ−∫0yc0yˆdyˆ−y∫0Hc01−yHdy and (3.27) vBα2Ri=∂∂x1Ri∂pB∂xy22H2−y3+y22∫yHc01−yˆHdyˆ+y1−y2H∫0yc0yˆdyˆ−12∫0yc0yˆ2dyˆ. Using the condition vB=0 at y=H in the above equation and integrating once gives (3.28) QBα2Ri=1Ri∂pB∂xH312+12∫0Hc0y(H−y)dy for the buoyancy-induced flow rate QB=∫0HuB dy. Integrating a second time with pB=0 at x=0,n to give (3.29) QB(τ)=α2Ri∫0nH−3∫0Hc0y(H−y)dydx2n∫01H−3dx, finally determines the pressure gradient (3.30) 1Ri∂pB∂x=6H3∫0nH−3∫0Hc0y(H−y)dydxn∫01H−3dx−∫0Hc0y(H−y)dy, which can be used in (3.26) and (3.27) to complete the determination of the buoyancy-induced velocity. Note that, because c0 is not spatially periodic, the solution carries a dependence on the channel length n through the pressure gradient ∂pB/∂x. 4. Solute dispersion The flow velocity is coupled with the solute concentration c through the dependence on c0 present in vB=uB,vB. The computation of c0 involves substitution of the expansion c=c0+εc1+ε2c2+⋯ into (2.12) with the time derivative replaced by the two-time-scale expression ∂c/∂t+ε2∂c/∂τ. At leading order we find the result ∂c0/∂t=0, anticipated earlier when writing (3.17a,b). Collecting terms of order ε yields (4.1) ∂c1∂t+v0⋅∇c0=0, which can be integrated to provide the concentration correction (4.2) c1=c1(x,y,τ)−∫v0dt⋅∇c0, where ∫v0dt=Re eit(U,V), as follows from (3.1a-c). It is worth noting that, because the solute diffusivity takes small values of order κ/ν~ε2, effects of diffusion are absent at the first two orders in the asymptotic analysis. These effects are present in the equation that arises at the following order, (4.3) ∂c2∂t+∂c0∂τ+v0⋅∇c1+v1⋅∇c0=1α2σ∂2c0∂y2, which can be time-averaged to give (4.4) ∂c0∂τ+∫v0dt⋅∇v0+v1⋅∇c0=1α2σ∂2c0∂y2. In deriving the second term in (4.4) from the third term in (4.3) use has been made of (4.2). Since c1 is independent of t and v0=0, the contribution of the former to the resulting time average v0⋅∇c1=v0⋅∇c1 is identically zero. The leading-order solute concentration c0 is also independent of t, so that the contribution arising from the second term in (4.2) can be written in the form (4.5) −v0⋅∇∫v0dt⋅∇c0=−v0⋅∫∇u0dt∂c∂x−v0⋅∫∇v0dt∂c∂y−u0∫u0dt∂2c∂x2−v0∫u0dt+u0∫v0dt∂2c∂x∂y−v0∫v0dt∂2c∂y2. With the time averages of any two harmonic functions A and B satisfying A∫Bdt=−∫AdtB and A∫Adt=B∫Bdt=0, it follows that the terms in the second line of the above equation are identically zero, whereas the remaining two terms on the right-hand side can be cast in the compact form shown in (4.4). As seen in (4.4), convective transport in the long time scale relies on the time-averaged Lagrangian velocity, given by the sum of the time-averaged Eulerian velocity v1=vSS+vB and the Stokes drift vSD=uSD,vSD=∫v0dt⋅∇v0 (see, e.g., Larrieu et al. 2009 for a discussion on Lagrangian transport in a similar wall-bounded flow). The latter contribution can be evaluated in terms of the complex functions U and V with use of the expressions (4.6a,b) uSD=12Im ∂∂x(UU‾)+∂∂y(VU‾) and vSD=12Im ∂∂x(UV‾)+∂∂y(VV‾), which follow from the identity Re eitf1Re iitf2=Im f1f‾2/2. The function uSD, which is related to the function F defined earlier in (3.18), is identically zero at x=0,1/2,1,3/2,…, so that the associated constant volumetric flow rate is simply (4.7) QSD=∫0HuSDdy=0. As our asymptotic description is limited to the leading-order term in the asymptotic expansion (2.17) for the solute concentration, to summarise the results of the asymptotic analysis one may replace c0 with c when rewriting the final transport (4.4) in the form (4.8) ∂c∂τ+uSD+uSS+uB∂c∂x+vSD+vSS+vB∂c∂y=1α2σ∂2c∂y2. The description of the solute dispersion following its deposition in the channel reduces to the integration of the above equation with initial condition c=ci(x,y) at τ=0 and boundary conditions ∂c/∂y=0 at y=0,H. In the integration, the time-independent Stokes-drift and steady-streaming velocities are computed with use of (4.6a,b) and of (3.19), (3.20) and (3.25), respectively, whereas the time-varying buoyancy-induced velocity is evaluated in terms of the solute concentration through the integral expressions (3.26), (3.27) and (3.30) with c0=c. Observation of (4.8) reveals that gravitational forces modify the character of the solution in a non-trivial way. Owing to the dependence of uB and vB on the concentration distribution c, the time-averaged transport equation that governs the dispersion of the solute, which for Ri=0 reduces to a linear partial-differential equation with time-independent coefficients, turns into a complicated nonlinear integro-differential equation in the presence of buoyancy. It is worth noting that, whereas the volumetric flow rates QSS=∫0HuSSdy and QSD=∫0HuSDdy associated with steady streaming and Stokes drift are identically zero, as discussed previously, that induced by buoyancy is in general non-zero, its value QB=∫0HuBdy evolving in time according to (3.29). Note that writing (4.8) in conservative form and integrating across the channel with use of ∂c/∂y=0 and vSD=vSS=vB=0 at y=0,H yields the relation (4.9) ∂C∂τ+∂ϕ∂x=0 between the amount of solute per unit channel length C(x,τ)=∫0Hcdy and the solute flux ϕ(x,τ)=∫0HuSD+uSS+uBcdy, whereas a second integral between x=0 and x=1 gives (4.10) ∂∂τ∫01Cdx+ϕ(1,τ)−ϕ(0,τ)=0, which naturally reduces to the expected conservation law (4.11) ∫01∫0Hcdydx=∫01∫0Hcidydx, when the solute flux at the two ends is zero. An important aspect of the reduced description developed above is that the nonlinear integro-differential equation (4.8) targets directly the evolution of the flow in the long time scale τ~1, relevant for solute dispersion over distances of order λ (i.e. dimensionless distances x of order unity), thereby circumventing the need to describe the small concentration fluctuations occurring in the short time scale t=ε−2τ. As a consequence, the model predictions involve computational times that can be expected to be a factor ε2 smaller than those required in DNS, because to describe solute dispersion the DNS must track the flow over a large number of cycles ~ε−2, larger for smaller ε. 5. Selected numerical results The reduced flow description is to be utilised to investigate the influence of buoyancy on solute dispersion. To facilitate the computation, the conservation (4.8) was written in terms of the normalised coordinate η=y/H(x), so that the integration domain becomes 0<x<n and 0<η<1. The numerical scheme utilises a fourth-order compact centred finite-difference approximation for the spatial discretisations of the viscous terms and a second-order upwind scheme for the nonlinear terms. A third-order TVD Runge–Kutta scheme is used for the time marching, whereas the integral expressions (3.26), (3.27) and (3.30) are evaluated with a simple trapezoidal rule. The accuracy of the model predictions, derived in the asymptotic limit ε≪1 for a slender channel with ho/λ≪1, is tested through comparisons with two-dimensional, unsteady simulations of the fluid motion and solute dispersion for small but finite values of ho/λ and ε. The DNS, involving the complete (2.1)–(2.3) written in dimensionless form, span thousands of oscillation cycles, as needed to generate significant dispersion of the solute. The numerical integration was performed with the finite-volume solver Ansys Fluent (release 20.2), assuring second-order accuracy in time and in space. Computations employing upwind and central-differencing schemes for the convective terms were found to yield virtually indistinguishable results, with the former discretisation used in generating the figures shown in the following. A coupled algorithm was used for the pressure–velocity coupling. In addition to the boundary conditions used in integrating the slender flow (2.9)–(2.12), additional conditions of developed flow ∂u/∂x=∂c/∂x=0 at the upper and lower open boundaries are incorporated in integrating (2.1)–(2.3). The computations presented in the following correspond to a canal of total length n=3 and aspect ratio ho/λ=1/20 with the imposed pressure difference yielding a dimensionless stroke length ε=0.02. The time-periodic DNS results were averaged in time to determine the mean Eulerian velocity ⟨v⟩=(1/2π)∫tt+2πvdt, of order ε, to be compared with the steady-streaming velocity vSS, as explained later. In addition, tracer particles are used to compute the Lagrangian velocity vL by following their displacement over a cycle, i.e. if the particle located at (x,y) at time t moves to occupy the new location (x+δx,y+δy) at time t+2π, then the Lagrangian velocity at (x,y) and time t is defined as vL=(δx,δy)/(2π). 5.1. Buoyancy-free flow As mentioned previously, in the absence of buoyancy, the flow induced by the imposed pressure gradient is periodic in time and space. The steady Lagrangian motion for ε≪1 is given in this case by the sum of the steady-streaming velocity vSS and the Stokes-drift velocity vSD. These two contributions as well as their sum are shown in figure 2 for β=0.4 and two different values of α. As the flow in each cell is identical, it suffices to show the solution for 0≤x≤1, symmetric with respect to the centre line x=0.5. For each value of α, streamlines are plotted using a fixed increment δψ of the streamfunction ψ, defined in the usual way (e.g. ∂ψ/∂y=uSS and ∂ψ/∂x=−vSS for steady streaming) with ψ=0 along the wall, so that the interline spacing provides a measure of the local velocity. To further quantify the motion, colour contours are used to represent the associated vorticity Ω=ho/λ2∂v/∂x−∂u/∂y, which reduces to Ω=−∂u/∂y in the slender flow approximation. The spatially periodic, time-independent, steady-streaming velocity computed with (3.19) and (3.20) supplemented with (3.25) is shown in the second column of figure 2. The results are qualitatively similar to those presented in Guibert et al. (2010). For α=4, the flow structure of each half cell exhibits two counter-rotating vortices, whereas for α=16 the flow develops an additional, much weaker vortex, located near the section with largest width. As expected, the vorticity, having peak values of order unity for α=4, increases with increasing flow frequency as a result of augmented wall production to reach peak values exceeding Ω=40 for α=16. The steady-streaming results are compared with time-averaged velocity fields obtained in DNS with ε=0.02. In the comparison, the time-averaged DNS velocity is expressed in the rescaled form ⟨v⟩/ε~1, consistent with the scaling employed in defining vSS. The two functions vSS and ⟨v⟩/ε are seen to be almost identical, thereby giving additional confidence in the mathematical development. For instance, the peak values of the stream function and vorticity corresponding to the time-averaged DNS velocity ⟨v⟩/ε are ψPEAK=±(0.0115,0.1680) and ΩPEAK=±(1.4465,40.786) for α=(4,16), whereas the corresponding values for the steady-streaming motion are ψPEAK=±(0.0115,0.1699) and ΩPEAK=±(1.4474,40.787). The small relative differences remain below about 1%, as is consistent with the order of the asymptotic development. The third column in figure 2 displays the Stokes-drift velocity field evaluated with (4.6a,b). As it is clear from a quantitative comparison with the corresponding steady-streaming results, both bulk-flow velocities have comparable magnitude for α=4, whereas for α=16 the Stokes drift provides a much smaller relative contribution to the Lagrangian drift. The dominant role of steady streaming in flows at high Womersley numbers is found also away from the wall in oscillating flow over circular cylinders (Holtsmark et al. 1954; Raney, Corelli & Westervelt 1954), for example. The mean Lagrangian velocity vSS+vSD corresponding to the asymptotic limit ε≪1 compares favourably with the velocity vL/ε obtained numerically by following tracer particles in the DNS computation for ε=0.02, shown in the last column of figure 2, although the relative errors are somewhat larger than those of the Eulerian velocity. For instance, the peak values of the stream function and vorticity corresponding to vL/ε are ψPEAK=±(0.0235,0.1674) and ΩPEAK=±(2.0454,45.9497) for α=(4,16), while the corresponding values for vSS+vSD are ψPEAK=±(0.0235,0.1491) and ΩPEAK=±(1.9906,40.787). 5.2. Buoyancy-free solute dispersion The reduced transport (4.8) resulting from the two-time-scale asymptotic analysis indicates that the solute relies on the Lagrangian drift for longitudinal dispersion. As a consequence, the existence of the closed recirculating vortices displayed in figure 2 implies that in oscillatory buoyancy-free channel flow a solute released in a given cell would be unable to reach their neighbouring cells, thereby precluding its progression along the canal. To illustrate this important feature of the flow, we show in figure 3 the temporal evolution of a bolus of solute with reduced Schmidt number σ=1 released at the initial instant of time in the central cell of a three-cell canal. The initial concentration is given by the truncated Gaussian distribution (5.1) ci=min1,32exp −16x−x0δ2, which represents a band of solute with characteristic width δ centred at xo having a saturated core flanked by thin layers across which the concentration decays to zero. Results obtained by integration of (4.8) for x0=1.75 and δ=0.2 are compared in figure 3 at different instants of time in the interval 0≤τ≤8 with DNS computations. Note that, with τ=ε2t, for the value ε=0.02 used in the DNS, this interval of time corresponds to 0≤t≤20000 (i.e. about 20000/2π≃3200 oscillatory cycles). As seen in figure 3 the model accurately reproduces the DNS results. To facilitate the quantitative comparisons, in addition to colour contours showing the solute concentration, the figure includes side plots for the amount of solute per unit channel length C=∫0Hc dy at different instants of time, with the initial distribution Ci=H(x)ci(x) included for reference as a dotted curve. The model predictions lie very close to the DNS results, in that the normalised value of the integrated departure ∫0nCDN−CMODELdx/∫0nCidx, which provides a metric for the accuracy of the model, remains below 0.003 over the entire range of times considered in the figure. For the buoyancy-free conditions considered in figure 3, the steady Lagrangian motion is seen to stir the solute about the deposition location, uniformising its concentration within the recirculating cell. The effect of longitudinal diffusion, present in the DNS results, is found to be rather limited, in that, even at the latest instant of time considered, the presence of the solute in the adjacent cells is negligibly small. This tendency of the solute to remain trapped inside Lagrangian vortices has potential implications concerning the drug-dispersion rate in ITDD procedures. Although the Lagrangian flow in the spinal canal does not exhibit the spatial periodicity of the canonical configuration investigated here, closed recirculating vortices, associated with the changes in the eccentricity of the spinal cord along the canal, have been found to characterise the CSF bulk motion (Coenen et al. 2019). Typically, there are three main vortices, extending along the cervical, thoracic and lumbar regions. As ITDD injection occurs in the lumbar region, the buoyancy-free results in figure 3 seem to indicate that, when the density of the drug matches exactly the CSF density, the drug is bound to linger in the lumbar vortex near the injection site without reaching the thoracic region. This could be advantageous in applications involving pain medication, which is meant to be delivered to the spinal cord, but not in applications involving anticancer drugs targeting brain tumors, for example. As shown in the following, buoyancy-induced motion has the potential to drastically change the associated transport rate, in accordance with clinical observations (Mitchell et al. 1988; Povey et al. 1989; Richardson et al. 1996; Veering et al. 2001). 5.3. Slowly varying buoyancy-induced motion As reasoned previously, buoyancy forces, acting on solutes with density ρs≠ρ, alter the steady Lagrangian drift by adding an additional component that varies slowly in the long time scale τ following the solute dispersion, so that the flow and the solute transport are intimately coupled, as described by (4.8) supplemented with (3.26), (3.27) and (3.30). The corresponding behaviour is characterised in figure 4 for a light solute with Ri=1 spreading upwards. Note that, because of the problem symmetry, results corresponding to a heavy solute with Ri=−1 can be generated by simply reversing the direction of the gravity vector, i.e. by rotating the figure 180°. As in the buoyancy-free flow depicted in figure 3, the solution in figure 4 includes Lagrangian streamlines, colour contours of solute concentration, and streamwise distributions of integrated solute concentration C=∫0Hc dy along the canal. Buoyancy has a dramatic effect on the dispersion of the solute, as is apparent by comparing the results in both figures. Gravitational forces acting on the light solute induce a longitudinal pressure gradient that modifies drastically the resulting Lagrangian drift, as can be seen by comparing the streamlines in figure 3 with those in figure 4. The pattern of symmetric recirculating cells with unconnected streamlines existing for Ri=0 is replaced for Ri=1 by a more complicated streamline pattern featuring a net upward flow rate QB(τ) (see the solid curves in figure 5, to be discussed later). As can be seen, although the flow rate and the associated streamline pattern vary slowly in time, the observed changes are not very pronounced. Also of interest is that the quantitative agreement between the streamlines predicted by the model and the DNS results is again remarkable, thereby giving additional confidence in our development. The changes in the Lagrangian motion have a dramatic reflection in the solute dispersion. As shown in figure 4, the solute is transported upwards following the Lagrangian streamlines connecting the cells, enabling its upward progression. The bolus of solute is distorted by the recirculating flow as it travels upwards, driven by the buoyancy-induced draft. The variation with time of the concentration distribution predicted with the model is in excellent agreement with the DNS results. The model is shown to predict not only the mean location of the bolus but also its shape and elongation. The relative departures, measured by ∫0nCDNS−CMODELdx/∫0nCidx, remain below 0.009 over the entire range of times shown in the figure. In assessing the potential benefits of the time-averaged formulation, it is important to emphasise that, although the integration of the integro-differential equation (4.8) over times τ~1 can be completed in a few hours using a laptop computer, generating the DNS results shown in figure 4, spanning over 3000 cardiac cycles, required 10 days in a computational cluster using a total of 72 cores. 5.4. Parametric dependence of the results The case shown in figure 4 , corresponding to β=0.2,α=4,σ=Ri=1 is used in figures 5 and 6 as a basis to investigate the influence of the different parameters on the dispersion of a buoyant solute. To that end, results are generated with use of (4.8) by modifying one of the four controlling parameters at a time, while keeping the other three fixed at the values selected earlier. Figure 5 shows the variation with time of the buoyancy-induced flow rate QB, whereas figure 6 shows instantaneous solute-concentration distribution and associated Lagrangian streamlines at a fixed time, namely, τ=(6,6,6,2) for figures 6(a)–6(d), with corresponding results for the base case at these times shown in two of the subpanels of figure 4. We begin by discussing the effect of the channel geometry. As shown in figure 6(a), increasing the undulation of the channel from β=0.1 to β=0.4 tends to increase the magnitude of the buoyancy-induced longitudinal velocity in the region of minimum cross-sectional area, where streamlines are closely spaced together for larger β, but these changes result in only moderately small variations of the flow rate QB, as shown in figure 5(a). As a result, the bolus of solute becomes more elongated as β increases, but advances upward at approximately the same rate, so that the maximum concentration occupies approximately the same location at τ=6, as shown in figure 6(a). The effect of the Schmidt number σ, entering in the formulation only through the factor affecting the transverse diffusion rate on the right-hand side of (4.8), is investigated in figures 5(b) and 6(b). The changes observed in streamline pattern and flow rate when changing the Schmidt number from σ=0.25 to σ=8 are not very significant, so that the differences in solute evolution in figure 6(b) are attributable to the direct effect of transverse diffusion (or lack thereof). The snapshot corresponding to σ=8 displays the behaviour expected at large Schmidt numbers, for which fluid particles maintain a nearly constant concentration in their slow Lagrangian evolution, as described by the limiting form of (4.8) for σ≫1. In the opposite limit σ≪1, solute diffusion leads to a rapid uniformisation of the concentration, as can be seen by integrating ∂2c/∂y2=0 (i.e. the reduced form of (4.8) when σ≪1) subject to the non-permeability condition ∂c/∂y=0 at y=0,H to give c=c(x,τ). As a result, the bolus remains relatively compact as it moves along the channel with a velocity determined by the flow rate. The reduced transport equation governing the transport of solutes with σ≪1 can be derived from (4.9) by noting that in this limit the integrated solute concentration C=∫0Hcdy becomes C=H(x)c(x,τ) while the solute flux ϕ(x,τ)=∫0HuSD+uSS+uBc dy reduces to ϕ=QBc. Substituting these simplified expressions into (4.9) and using (3.29) to evaluate QB finally yields (5.2) ∂c∂τ+α2Ri∫0ncdx12nH∫01H−3dx∂c∂x=0 as the limiting form of (4.8) for σ≪1 As is to be expected from the velocity expressions derived in §3.4, the Richardson number and the Womersley number have a pronounced effect on the mean Lagrangian motion. As shown in figures 5(c) and 5(d), the flow rate exhibits dependences on Ri and α that are approximately linear and approximately quadratic, respectively, consistent with (3.29). These dependences have a reflection on the evolution of the solute bolus shown in figures 6(c) and 6(d). With limited updraft for Ri=0.25, the bolus is seen to spread about the injection location, without significant upward progression at the instant of time τ=6 considered in the figure. An increase in Ri promotes the displacement of the bolus, but its longitudinal extent remains approximately equal in all three cases. By way of contrast, an increase in α increases the upward displacement and also enhances bolus distortion. The reason for the latter is that larger values of α hinder transverse diffusion, as can be inferred from (4.8), with the result that fluid particles travel following the Lagrangian recirculating paths with a nearly constant concentration, rapidly deforming the compact concentration distribution of the initial bolus. 6. Conclusions Solute dispersion in a wavy-walled vertical channel subject to an oscillating pressure gradient has been used as a canonical model to investigate the effect of buoyancy on the transport of ITDD drugs, characterised by large values of the Schmidt number and order-unity values of the Richardson number. The mean Lagrangian velocity determined in the asymptotic limit of small stroke lengths, responsible for the convective transport of the solute, displays a buoyancy component whose local value depends on the solute concentration through integral expressions, resulting in a nonlinear integro-differential transport equation. The predictive capabilities of the reduced description are tested through comparisons with DNS computations involving thousands of oscillating cycles. The validation exercise reveals that the model provides accurate predictions of solute dispersion at a fraction of the computational cost involved in the DNS. In contrast to the motion observed in the buoyancy-free case investigated in figures 2 and 3, characterised by the existence of a series of closed Lagrangian vortices distributed periodically along the channel, the buoyancy-modulated mean Lagrangian flow shown in figure 4 includes a streamwise draft connecting neighbouring cells that promotes the longitudinal dispersion of the solute. The buoyancy-enhanced transport rate revealed in our channel computations is consistent with previous clinical observations pertaining to dispersion of light drugs for patients in a sitting injection position (Richardson et al. 1996). The simple canonical flow considered here has served to unveil some of the key aspects of the solute-dispersion problem, including the enhanced transport associated with the buoyancy-modulated mean Lagrangian velocity. Future work should consider application of the two-time-scale asymptotic analysis delineated previously to the description of ITDD processes, with account taken of the three-dimensional morphology of the spinal canal, possibly including the effect of microanatomical features such as arachnoid trabeculae, which are thin strands of connective tissue that form a web-like structure stretching across the spinal canal. The presence of these fine anatomical structures, which has been shown to have an important effect on pressure loss (Tangen et al. 2015), can be accounted for by treating the spinal subarachnoid space as a Brinkman porous medium, as done in previous investigations (Gupta et al. 2009; Kurtcuoglu, Jain & Martin 2019; Salerno, Cardillo & Camporeale 2020; Sincomb et al. 2022). The future developments envisioned here can potentially provide a reduced transport equation, possibly similar to (4.8), to be used in combination with magnetic resonance imaging characterisations of the canal anatomy (Coenen et al. 2019) to describe the transport of the drug in the relevant dispersion time scale. The ultimate goal of such efforts is the development of computationally effective subject-specific predictive tools for drug delivery to a target site from injection by a lumbar puncture with account taken of the specific anatomy and physiological conditions of the individual patient as well as for the molecular characteristics and injection rate of the drug, as needed in guiding clinical treatments. Acknowledgement. We thank Dr Jenna Lawrence for insightful discussions regarding clinical observations reported in the literature. Funding. This work was supported by the National Institute of Neurological Disorders and Stroke through contract number 1R01NS120343-01 and by the National Science Foundation through grant number 1853954. The work of WC was partially supported by the Comunidad de Madrid through contract CSFFLOW-CM-UC3M and by the Spanish MICINN through the coordinated project PID2020-115961RB, whereas that of CGM was partially supported by Junta de Andalucía and European Funds through grant P18-FR-4619 and by the Spanish MICINN through the coordinated project PID2020-115961RB. Figure 1. A schematic view of the spinal canal, showing the location of (a) ITDD and (b, c) of the two-dimensional channel flow investigated here. Figure 2. Streamlines and colour contours of vorticity corresponding to the steady-streaming velocity vSS, Stokes-drift velocity vSD and steady mean Lagrangian velocity vSS+vSD in a canal with β=0.4 for (a) α=4 and (b) α=16. Results of DNS of a non-buoyant flow (i.e. Ri=0) with ho/λ=1/20 and ε=0.02 are also shown, including the rescaled time-averaged Eulerian velocity field ⟨v⟩/ε (first column) and the rescaled Lagrangian velocity vL/ε (fifth column), the latter determined by following tracer particles, as explained in the text. To facilitate the comparisons, fixed constant streamline spacings δψ=0.003 and δψ=0.03 are used for the upper and lower plots, respectively. Figure 3. Snapshots of solute concentration for Ri=0,β=0.2,α=4 and σ=1 as obtained at five different instants of time from the reduced model and from DNS for ho/λ=1/20 and ε=0.02. In addition to colour contours of local concentration, the figure shows distributions of solute per unit channel length ∫0Hcdy for the model (solid curve) and for the DNS (dot-dashed curves), with the initial distribution ∫0Hcidy shown as a dotted curve. For reference, the figure shows streamlines with constant spacing δψ=0.003 for the Lagrangian mean drift, which is characterised using the asymptotic prediction vSS+vSD for the model results and the value of vL/ε determined numerically for the DNS results. Figure 4. Snapshots of solute concentration for Ri=1,β=0.2,α=4 and σ=1 as obtained at five different instants of time from the reduced model and from DNS for ho/λ=1/20 and ε=0.02. In addition to colour contours of local concentration, the figure shows distributions of solute per unit channel length ∫0Hcdy for the model (solid curve) and for the DNS (dot-dashed curves), with the initial distribution ∫0Hcidy shown as a dotted curve. The plots include streamlines with constant spacing δψ=0.003 for the varying Lagrangian mean drift, which is evaluated with use of vSS+vSD+vB (model results) and from the displacement of the tracer particles (DNS results). Figure 5. Influence of (a) the contraction ratio β, (b) reduced Schmidt number σ, (c) Richardson number Ri and (d) Womersley number α on the temporal evolution of the buoyancy-induced flow rate QB. The values of the parameters in each case are: (a) α=4,σ=Ri=1;(b)α=4,Ri=1 and β=0.2;(c)α=4,σ=1 and β=0.2;(d)σ=Ri=1 and β=0.2. Figure 6. Influence of (a) the contraction ratio β, (b) reduced Schmidt number σ, (c) Richardson number Ri and (d) Womersley number α on the solute-concentration distribution and associated Lagrangian streamlines. The snapshots are taken at τ=6 for (a−c) and at τ=2 for (d). The values of the parameters in each case are: (a) α=4,σ=Ri=1; (b) α=4,Ri=1 and β=0.2; (c) α=4,σ=1 and β=0.2; (d) σ=Ri=1 and β=0.2. The streamline spacing for the mean Lagrangian velocity is δψ=0.003 for (a-c) and δψ=0.005 for (d). Declaration of interests. The authors report no conflict of interest. ==== Refs References Bellhouse BJ , Bellhouse FH , Curl CM , MacMillan TI , Gunning AJ , Spratt EH , MacMurray SB & Nelems JM 1973 A high efficiency membrane oxygenator and pulsatile pumping system, and its application to animal trials. ASAIO J. 19 (1 ), 72–79. Bottros MM & Christo PJ 2014 Current perspectives on intrathecal drug delivery. J. Pain Res 7 , 615–626.25395870 Coenen W , Gutiérrez-Montes C , Sincomb S , Criado-Hidalgo E , Wei K , King K , Haughton V , Martínez-Bazán C , Sánchez AL & Lasheras JC 2019 Subject-specific studies of CSF bulk flow patterns in the spinal canal: implications for the dispersion of solute particles in intrathecal drug delivery. Am. J. Neuroradiol 40 (7 ), 1242–1249.31196863 Fowler MJ , Cotter JD , Knight BE , Sevick-Muraca EM , Sandberg DI & Sirianni RW 2020 Intrathecal drug delivery in the era of nanomedicine. Adv. Drug Deliv. Rev 165 , 77–95.32142739 Grotberg JB 1984 Volume-cycled oscillatory flow in a tapered channel. J. Fluid Mech 141 , 249–264. Grotberg JB 1994 Pulmonary flow and transport phenomena. Annu. Rev. Fluid Mech 26 (1 ), 529–571. Guibert R , Plouraboué F & Bergeon A 2010 Steady streaming confined between three-dimensional wavy surfaces. J. Fluid Mech 657 , 430–455. Gupta S , Poulikakos D & Kurtcuoglu V 2008 Analytical solution for pulsatile viscous flow in a straight elliptic annulus and application to the motion of the cerebrospinal fluid. Phys. Fluids 20 (9 ), 093607. Gupta S , Soellinger M , Boesiger P , Poulikakos D & Kurtcuoglu V 2009 Three-dimensional computational modeling of subject-specific cerebrospinal fluid flow in the subarachnoid space. Trans. ASME J. Biomech. Engng 131 (2 ), 021010. Gutiérrez-Montes C , Coenen W , Lawrence JJ , Martínez-Bazán C , Sánchez AL & Lasheras JC 2021 Modelling and direct numerical simulation of flow and solute dispersion in the spinal subarachnoid space. Appl. Math. Model 94 , 516–533. Hall P 1974 Unsteady viscous flow in a pipe of slowly varying cross-section. J. Fluid Mech 64 (2 ), 209–226. Hallworth SP , Fernando R , Columb MO & Stocks GM 2005 The effect of posture and baricity on the spread of intrathecal bupivacaine for elective cesarean delivery. Anesth. Analg 100 (4 ), 1159–1165.15781538 Holtsmark J , Johnsen I , Sikkeland T & S Kavlem S 1954 Boundary layer flow near a cylindrical obstacle in an oscillating, incompressible fluid. J. Acoust. Soc. Am 26 (1 ), 26–39. Khani M , Sass LR , Xing T , Sharp MK , Balédent O & Martin BA 2018 Anthropomorphic model of intrathecal cerebrospinal fluid dynamics within the spinal subarachnoid space: spinal cord nerve roots increase steady-streaming. Trans. ASME J. Biomech. Engng 140 (8 ), 081012. Klarica M , Radoš M & Orešković D 2019 The movement of cerebrospinal fluid and its relationship with substances behavior in cerebrospinal and interstitial fluid. Neuroscience 414 , 28–48.31279048 Kurtcuoglu V , Jain K & Martin BA 2019 Modelling of cerebrospinal fluid flow by computational fluid dynamics. In Biomechanics of the Brain, pp. 215–241. Springer. Larrieu E , Hinch EJ & Charru F 2009 Lagrangian drift near a wavy boundary in a viscous oscillating flow. J. Fluid Mech 630 , 391–411. Lawrence JJ , Coenen W , Sánchez AL , Pawlak G , Martínez-Bazán C , Haughton V & Lasheras JC 2019 On the dispersion of a drug delivered intrathecally in the spinal canal. J. Fluid Mech 861 , 679–720. Lee YC , Hsieh CC , Chuang JP & Li CY 2017 The necessity of intrathecal chemotherapy for the treatment of breast cancer patients with leptomeningeal metastasis: a systematic review and pooled analysis. Curr. Probl. Cancer 41 , 355–370.28779849 Linninger AA , Tangen K , Hsu CY & Frim D 2016 Cerebrospinal fluid mechanics and its coupling to cerebrovascular dynamics. Аnnu. Rev. Fluid Mech 48 , 219–257. Lo Jacono D , Plouraboué F & Bergeon A 2005 Weak-inertial flow between two rough surfaces. Phys. Fluids 17 (6 ), 063602. Lui ACP , Polis TZ & Cicutti NJ 1998 Densities of cerebrospinal fluid and spinal anaesthetic solutions in surgical patients at body temperature. Can. J. Anaesth 45 (4 ), 297–303.9597201 Lynch L 2014 Intrathecal drug delivery systems. Contin. Edu. Anaesth. Crit. Care Pain 14 , 27–31. Mackley MR & Stonestreet P 1995 Heat transfer and associated energy dissipation for oscillatory flow in baffled tubes. Chem. Engng Sci 50 (14 ), 2211–2224. Mitchell RWD , Bowler GMR , Scott DB & Edström HH 1988 Effects of posture and baricity on spinal anaesthesia with 0.5% bupivacaine 5ml. Br. J. Anaesth 61 (2 ), 139–143.3415884 Nicol ME & HoldCroft A 1992 Density of intrathecal agents. Br. J. Anaesth 68 (1 ), 60–63.1346744 Onofrio BM , Yaksh TL & Arnold PG 1981 Continuous low-dose intrathecal morphine administration in the treatment of chronic pain of malignant origin. Mayo Clin. Proc 56 , 516–520.6894954 Povey HMR , Jacobsen J & Westergatrd-Nielsen J 1989 Subarachnoid analgesia with hyperbaric 0.5 % bupivacaine: effect of a 60-min period of sitting. Acta Anaesth. Scand 33 (4 ), 295–297.2655366 Raney WP , Corelli JC & Westervelt PJ 1954 Acoustical streaming in the vicinity of a cylinder J. Acoust. Soc. Am 26 (6 ), 1006–1014. Remeš F , Tomáš R , Jindrák V , Vaniš V & Setlík M 2013 Intraventricular and lumbar intrathecal administration of antibiotics in postneurosurgical patients with meningitis and/or ventriculitis in a serious clinical state. J. Neurosurg 119 , 1596–1602.23952688 Richardson MG , Thakur R , Abramowicz JS & Wissler RN 1996 Maternal posture influences the extent of sensory block produced by intrathecal dextrose-free bupivacaine with fentanyl for labor analgesia. Anesth. Analg 83 (6 ), 1229–1233.8942591 Riley N 2001 Steady streaming. Annu. Rev. Fluid Mech 33 (1 ), 43–65. Salerno L , Cardillo G & Camporeale C 2020 Aris-Taylor dispersion in the subarachnoid space. Phys. Rev. Fluids 5 (4 ), 043102. Sánchez AL , Martínez-Bazán C , Gutiérrez-Montes C , Criado-Hidalgo E , Pawlak G , Bradley W , Haughton V & Lasheras JC 2018 On the bulk motion of the cerebrospinal fluid in the spinal canal. J. Fluid Mech 841 , 203–227. Sincomb S , Comnen W , Gutiérrez-Montes C , Martínez-Bazán C , Haughton V & Sánchez AL 2022 A one-dimensional model for the pulsating flow of cerebrospinal fluid in the spinal canal. J. Fluid Mech 939 , A26.36337071 Tangen KM , Hsu Y , Zhu DC & Linninger AA 2015 Cns wide simulation of flow resistance and drug transport due to spinal microanatomy. J. Biomech 48 (10 ), 2144–2154.25888012 Veering BT , Immink-Speet TTM , Burm AGL , Stienstra R & van Kleef JW 2001 Spinal anaesthesia with 0.5% hyperbaric bupivacaine in elderly patients: effects of duration spent in the sitting position. Br. J. Anaesth 87 (5 ), 738–742.11878525 Wildsmith JAW , Mcclure JH , Brown DT & Scott DB 1981 Effects of posture on the spread of isobaric and hyperbaric amethocaine. Br. J. Anaesth 53 (3 ), 273–278.6894092

PMC010xxxxxx/PMC10338034.txt

==== Front 9918351182406676 51422 Aggregate (Hoboken) Aggregate (Hoboken) Aggregate (Hoboken, N.J.) 2766-8541 2692-4560 37441005 10.1002/agt2.189 nihpa1862534 Article Bubble-pen lithography: Fundamentals and applications Nanoscience: Special Issue Dedicated to Professor Paul S. Weiss Kollipara Pavana Siddhartha 1 Mahendra Ritvik 2 http://orcid.org/0000-0003-0827-9758 Li Jingang 3 http://orcid.org/0000-0002-9168-9477 Zheng Yuebing 123 1 Walker Department of Mechanical Engineering, The University of Texas at Austin, Austin, Texas, USA 2 Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, Texas, USA 3 Material Science and Engineering Program, Texas Materials Institute, The University of Texas at Austin, Austin, Texas, USA AUTHOR CONTRIBUTIONS The authors thank Anand Swain for proofreading the manuscript. Pavana Siddhartha Kollipara and Ritvik Mahendra contributed equally to this work. Correspondence: Yuebing Zheng, Walker Department of Mechanical Engineering, The University of Texas at Austin, Austin, TX 78712, USA., zheng@austin.utexas.edu 3 1 2023 8 2022 08 3 2022 12 7 2023 3 4 e189https://creativecommons.org/licenses/by/4.0/ This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Developing on-chip functional devices requires reliable fabrication methods with high resolution for miniaturization, desired components for enhanced performance, and high throughput for fast prototyping and mass production. Recently, laser-based bubble-pen lithography (BPL) has been developed to enable sub-micron linewidths, in situ synthesis of custom materials, and on-demand patterning for various functional components and devices. BPL exploits Marangoni convection induced by a laser-controlled microbubble to attract, accumulate, and immobilize particles, ions, and molecules onto different substrates. Recent years have witnessed tremendous progress in theory, engineering, and application of BPL, which motivated us to write this review. First, an overview of experimental demonstrations and theoretical understandings of BPL is presented. Next, we discuss the advantages of BPL and its diverse applications in quantum dot displays, biological and chemical sensing, clinical diagnosis, nanoalloy synthesis, and microrobotics. We conclude this review with our perspective on the challenges and future directions of BPL. additive manufacturing capillary flow lithography marangoni convection microbubbles sensing ==== Body pmc1 | INTRODUCTION Recent years have witnessed an exponential growth in the development and implementation of functional miniaturized devices, such as point-of-care sensors,[1–4] high-density transistor chips,[5] planar batteries,[6] augmented reality glasses,[7,8] and micro-electro-mechanical systems.[9] Conventional top-down lithographical techniques, such as photolithography,[10–12] nanoimprinting lithography,[13–15] electron beam lithography, and focused-ion beam lithography,[16,17] have been developed to fabricate these devices with high fidelity. However, they generally suffer from expensive instruments, complex operations, and high materials wastage. In addition, the high-throughput fabrication of multi-component and three-dimensional (3D) structures is challenging. To overcome these limitations, researchers proposed bottom-up methods as promising alternatives for the fabrication of functional devices.[18] In comparison to conventional top-down techniques, bottom-up approaches provide additional flexibility in controlling the composition and configuration of functional structures.[19] A typical bottom-up approach like printing involves two steps: (1) synthesis of functional materials as inks,[20–22] and (2) printing of the inks into designed configurations on substrates for functional devices.[23–26] Many techniques, such as hydrothermal synthesis,[27] sol-gel synthesis,[28,29] physical or chemical vapor deposition,[30] and pyrolysis,[31,32] have been developed to produce inks that are solutions with colloidal particles of precisely controllable sizes, shapes, compositions, and properties. These colloidal particles can serve as the building blocks for the construction of diverse functional structures where the properties can be tuned by both the discrete particles and the inter-particle interactions. Bottom-up fabrication starting from these building blocks is promising for producing multi-component and multi-functional structures with single-particle resolution and low materials wastage. Several bottom-up technologies have been developed, such as self-assembly[33–35] and inkjet printing.[36,37] However, self-assembly lacks precise control over the patterning geometries and is limited to thermodynamically stable configurations. Inkjet printing suffers from poor resolution, strict requirements on ink parameters, and the coffee ring effect.[38] Laser-based fabrication techniques such as laser-induced hydrothermal synthesis[39–42] and laser sintering[43–45] offer unique advantages of high diffraction-limited resolutions and on-demand two-dimensional (2D) and 3D structure generation. However, laser-induced hydrothermal synthesis is limited by long laser exposure times and poor control of the 3D structure morphology. Laser-sintering provides a good prototyping alternative for electrical and pharmaceutical applications.[46,47] Typically, laser sintering uses a femtosecond laser to fuse the particles to achieve the required 2D/3D patterns on diverse substrates. Although scalability is achieved through larger beam size and fast laser-scanning speeds, challenges such as the resolution limit due to particle size, high surface roughness and high porosity of the finished products, thermal shrinking and warping of the printed structures, and thermal requirements on materials limit the wide-applicability of laser sintering technology.[48] An emerging technique, bubble-pen lithography (BPL), has been developed to overcome some of the limitations in the current bottom-up fabrication methods.[49] In brief, BPL exploits laser heating of the substrate or solution to generate light-controlled microbubbles for the concentration and immobilization of ink materials (e.g., colloids, polymers, biomolecules, and ions) on the substrate.[50–52] In addition, the concentrated ink materials at the bubble interfacial regions that act as local nanoreactors facilitate the in situ synthesis and structuring of unconventional functional materials.[53] The bubble size and printing resolution in BPL can be controlled by the laser power and exposure time, which is desired for the patterning of multiscale structures. In this review, we highlight the current progress in fundamentals, technologies, and applications of BPL. Specifically, we discuss the fundamentals of microbubble formation from both experimental and theoretical aspects and illustrate the laser-bubble-patterning parametric relationship in detail. In addition, we present the implementation of BPL in diverse applications and provide a broad perspective on the challenges and prospects of BPL. 2 | WORKING PRINCIPLE OF BPL BPL is achieved through the control of optically generated microbubbles at the substrate-liquid interface (Figure 1A). Localized optical heating of the substrate/particle increases the temperature beyond the boiling point of the solvent, which leads to vaporization of the solvent and the formation of a bubble at the laser spot.[54] The temperature difference across the microbubble corresponds to a steep temperature gradient of ~50–70 K/μm for a typical 5 μm bubble (Figure 1B).[55] Since surface tension at the bubble interface is a function of temperature, the strong surface tension gradient results in Marangoni convection (also called thermocapillary convection), a swirling motion in the fluid above the surface.[58] This convective flow leads to the attraction of inks in the solution within a range of 2–20 times the bubble size.[59,60] Microbubble generation is nearly instantaneous, with a timescale in the order of microseconds. The increase of laser exposure time results in the continuous growth of the bubble until it is saturated. The steady bubble size is a function of solvent properties, laser power, and laser exposure duration.[55] On-demand patterning on the substrate via BPL is achieved by the relative translation between the sample and the laser beam. As the heating spot shifts, the bubble continuously moves from one position to the other, which enables the continuous attraction of ink materials to the three-phase contact line for the immobilization and patterning on the substrate (Figure 1C).[56] An example of experimental realization of continuous patterning via BPL is shown in Figure 1D.[57] In addition to continuous printing, discrete printing is also possible by altering the “ON” and “OFF” state of the laser beam (Figure 1E).[52] By using a high-speed shutter to block the laser beam, discrete patterning can be achieved. Figure 1F shows the discrete patterning of microparticle clusters on the substrate. Typically, continuous printing is extensively applied in conductometric sensing in biology and chemistry,[4,61–63] while discrete patterning is commonly used for displays, interconnects, and multiplexed optical sensor arrays.[64–66] Versatile patterning of diverse structures can be effectively achieved by altering the continuous and discrete patterning. The implementation of BPL for practical applications necessitates an in-depth understanding of the underlying physical mechanisms and a coordinated tuning of multiple parameters to control the printing resolution. In the following sub-sections, we present the experimental control and theoretical understanding of BPL, which have led to the enhanced capability of BPL. 2.1 | Experimental demonstration of BPL Figure 2A shows the typical optical setup to focus a laser beam onto the substrate for BPL. In this setup, optically generated microbubbles are created and moved through the digital control of the sample stage to fabricate designed patterns. Several factors in BPL, such as heating source, solvent, ink, and substrate, need to be considered to achieve the desired printing performance. 2.1.1 | Heating source BPL is initiated with optothermally induced microbubbles, which require an efficient light-to-heat conversion in the system for solvent boiling and bubble formation.[67] This process can be achieved with the use of a combination of a light-absorbing substrate or a light-absorbing solution/dispersion with a relevant laser source. Plasmonic substrates are commonly used light-absorbing substrates in BPL.[68,69] Plasmonic substrates can be obtained by the deposition of a thin layer (<10 nm) of metals on other surfaces, such as glass and polyethylene terephthalate (PET). A thermal annealing process is usually followed to convert the continuous thin film into discrete nanoislands to enhance the localized heating.[70,71] Alternatively, plasmonic substrates can be prepared by the spin-coating or drop-casting of metal nanoparticle solution to deposit plasmonic nanoparticles onto the substrate. Both methods can lead to a relatively uniform distribution of metal nanoparticles and hotspots, and efficient light-to-heat conversion. Other light-absorbing substrates, such as indium-tin-oxide[72–74] and silicon-based substrates,[75,76] are also commonly used to convert light into heat. In addition, light-absorbing metal particles suspended in solvent can act as localized heat sources for plasmonic bubble generation.[77] A high concentration of particles is usually required for efficient bubble-controlled patterning of such light-absorbing particles. In instances where the ink particles are not light-absorbing, a combination of ink particles and metal particles is utilized for effective bubble generation and continuous lithography.[78] The other important parameters to control the optical heating include laser wavelength, laser power, and bubble exposure time. For example, plasmonic heating depends strongly on the laser wavelength relative to the plasmon resonance wavelength of the plasmonic substrate or dispersed plasmonic particle. Therefore, choosing the relevant laser wavelength is essential for a particular substrate or particle to generate bubbles at a lower laser power. As laser power increases, the bubble size and three-phase contact line expand, which increases the printing linewidth (Figure 2B).[56] The printing linewidth is also dependent on the bubble exposure time, which can be controlled by the laser/stage scanning speed (continuous patterning) or the laser duration (discrete patterning). As the laser/stage scanning speed increases, the exposure time reduces, which decreases the accumulation time of the dispersed materials and reduces the patterning linewidth (Figure 2C).[79] Similarly, Roy et al. demonstrated that the rings formed during the discrete patterning of composites had an increasing size along with the increasing laser duration (Figure 2D).[78] 2.1.2 | Substrate In addition to light absorption, the contact angle of bubbles on the substrate should also be optimized for the efficient immobilization of ink materials. For instance, Liu et al. demonstrated that the dynamics of generation and dissolution of plasmonic bubbles depend strongly on the wettability of the substrate.[80] The light-to-heat conversion is weakened when the substrate is coated with hydrophobic polymers, increasing the threshold laser power for bubble generation. Microbubbles generated on hydrophobic substrates also grow slower than those on hydrophilic substrates. The size of the microbubble determines the temperature gradient across the bubble that alters the printing forces for the colloidal particles. 2.1.3 | Dispersion Bubble generation and its stability strongly depend on the viscosity, boiling point, solute concentration, and stabilizing surfactants of the ink dispersions, which should be controlled for the better performance of BPL. Viscosity As the solution viscosity increases, the Marangoni convection decreases due to the greater fluidic friction, which weakens the particle accumulation. Moreover, for a given size of the bubble, bubble manipulation becomes increasingly difficult due to an increase in the drag force on the bubble during laser/stage movement in higher viscosity solutions. Thus, low-viscosity solutions (<10 cP) such as water and alcohols are typically utilized for BPL. Boiling point The higher the boiling point of the dispersion, the higher the laser flux required to generate the bubble. If all other parameters are the same, it is easier for bubble formation in volatile liquids with lower boiling points than water. Entrapped air in the solution The bubble formation and growth dynamics strongly depend on the contents of the bubble – is it a vapor bubble or a gaseous bubble? Wang et al. presented a thorough understanding of bubble dynamics in the presence and absence of air in the dispersion (Figure 2E).[81] During the instantaneous formation of a bubble (bubble nucleation), solvent around the heating spot is immediately vaporized to form a vapor bubble. Later, as laser heating continues, the air dispersed in the medium is attracted to the bubble, thus forming a mixture of air and solvent vapor inside the bubble. As the air continues to enter the bubble, the size of the bubble increases drastically, and an asymptotic size is achieved at longer time scales. If the dispersion is degassed, the effects of the entrapped air in the dispersion will be minimum, and the change rate of bubble size is much lower compared to that of bubble size in air-entrapped dispersions.[81] Solute concentration An increasing solute concentration can increase the particle accumulation at the three-phase contact line. Figure 2F shows the dependence of accumulation at varying particle concentrations (five orders of difference) at the same laser power and laser duration.[82] In addition, if the heating sources are light-absorbing particles in the solution, a high concentration of solute particles is essential for efficient light-to-heat conversion. Stabilizing surfactants Surfactants are often utilized to synthesize and stabilize colloidal dispersions to prevent aggregate formation by inducing electrostatic repulsions among particles.[83,84] In general, surfactants in the dispersion will be attracted to the bubble interface and reduce the surface tension magnitude and gradient, resulting in a decreased convection and range of particle attraction. In the absence of any counteracting chemical reactions or particle-surfactant interactions, the accumulation of particles thus reduces with the surfactant addition, resulting in smaller feature sizes of the printed structures. However, in some instances, additional surfactants can enhance the particle accumulation. Yamamoto et al. showed such a contrary response in the presence of a nonionic surfactant. Since the adsorption of nonionic surfactants is exothermic, the surfactant molecules are densely accumulated on the top portion of the microbubble (where the temperature is lower), thereby increasing the surface tension gradient near the three-phase contact line. This result caused an additional trapping force and led to a larger accumulation of the colloidal particles.[82] We have briefly described the effects of these dispersion parameters on bubble formation and particle accumulation. Detailed theoretical analysis on the printing dynamics of BPL is still essential in promoting its practical applications, which will be discussed in the following sub-section. 2.2 | Theoretical analysis of BPL Optothermal microbubbles are generated due to the laser heating of light-absorbing substrates or light-absorbing particles. In some instances, microbubbles are also formed due to two-photon/multi-photon absorption-mediated heating of liquid at the substrate-liquid interface. We first consider bubble nucleation on light-absorbing plasmonic substrates. For light-absorbing plasmonic substrates such as gold nanoislands, discrete nucleation sites are densely located on the substrate to facilitate bubble formation during the printing. Figure 3A shows the temperature distribution around a laser-generated microbubble using computational fluid dynamics simulations. The temperature of the plasmonic substrates can reach beyond 120°C prior to bubble nucleation,[49] whereas, for the light-absorbing metal particles, the temperature of particle surfaces can reach beyond 700°C.[85] The vapor close to the thermal hotspot is superheated, and the temperature within the bubble decreases with increasing distance from the hot surface. The resultant temperature gradient causes a surface tension gradient along the bubble interface, which results in Marangoni convection with a spatially varying velocity distribution, as shown in Figure 3B. Rajeeva et al. determined the trajectories of metal ions in the solution under the influence of plasmonic bubbles using a combination of computational fluid dynamics and random walk simulations (Figure 3C).[53] The results revealed that only a small fraction (~5%) of metal ions in the solution reached the three-phase contact line, which could be used for predetermination of the precursor concentration for a required linewidth of a printed structure. “What forces print the particle on the substrate?” is one of the fundamental questions in BPL. Research has revealed that a combination of several distinct forces act on a particle as shown in Figure 3D. First, a particle away from the microbubble is dragged toward the three-phase contact line due to the Marangoni convection. When approaching the microbubble, the particle is then dragged toward the three-phase contact line by an evaporative force induced by continuous evaporation of the superheated liquid between the bubble and the substrate (middle panel in Figure 3D).[86] As the particle touches the bubble surface, evaporative force, capillary force (from bubble surface distortion), electrostatic force, and Van der Waals force act on the particle.[87] Since the bubble surface is almost a rigid interface due to the pressure difference between inside and outside the bubble, the downward capillary force on the particle is extremely high, which can overcome the electrostatic and Van der Waals forces. The net printing force enables the immobilization of particle on the substrate. It should be noted that the direction of the capillary force is strongly dependent on the contact angle of the bubble. Thus, there are variations in the printing capability for different solute-solvent-substrate systems. In the case of light-absorbing particles as the heating sources, the force analysis for particle immobilization remains the same, while the mechanism of particle trapping and accumulation is different. For well-dispersed light-absorbing particle dispersion, the bubble nucleation sites are the particles themselves, which are mobile and homogenously distributed throughout the volume above the substrate. Zhang et al. performed a detailed analysis of bubble generation due to volumetric heating of the dispersion and the resultant accumulation of particles due to both stationary and moving bubbles.[88] Briefly, the metal particles are excited with a resonant femtosecond laser beam to generate and grow the bubbles to a certain size.[89] Because of the existence of metal nanoparticles in the region above the bubble surface, the transmitted laser beam excites the metal particles, and the temperature increases on the top of the bubble (Figure 3E). The surface tension gradient is pointed downward along the bubble interface, which slides the particles toward the substrate and immobilizes them on the substrate (Figure 3F). As the laser beam moves for continuous printing, the laser beam is refracted at the bubble interface, which creates a temperature hotspot in the direction of laser movement and causes an additional drag force on the particles. The accumulated particles on the substrate near the three-phase contact line act as local heat sources and cause the bubble to move forward. The significant difference between light-absorbing substrates and light-absorbing particles is the effect of bubble motion. The bubble motion in the former case causes particle immobilization at a new position,[49] while in the latter case, particle immobilization happens at the existing position to move the bubble to a new equilibrium position due to asymmetric heating.[88] 3 | APPLICATIONS OF BPL The accumulation and patterning of particles, ions, and molecules are essential in many emerging applications.[90–92] Diverse nanoparticles and molecules patterned in selected configurations have been implemented for electronic, biological, and chemical applications.[93–96] Accumulated nanoparticles have also been implemented in optical metamaterials and chiral metamolecules.[100,97–99] Compared to other lithography techniques, BPL has a unique advantage in concentrating particles locally before their immobilization on the substrates. This concentrating capability enables the use of diluted inks with a low amount of materials for less wastage and easier processing rather than commercial viscous nanoparticle inks. Moreover, the localized high temperature and pressure at the three-phase contact line can be exploited as a local chemical nanoreactor to intensify chemical reactions and drive unconventional reaction pathways, which facilitate the in situ synthesis of complex inks and structured novel nanomaterials. In this section, we highlight the prominent applications of BPL, including quantum dot (QD) display, clinical diabetes detection, temperature-sensitive protein sensing, gas sensing, and microrobotics. 3.1 | BPL for QD patterning QDs are known for their tunable optical properties with narrow emission bandwidth and high quantum efficiency.[101,102] They have been used in many applications, such as displays,[103] nanolasers,[104] photodetectors,[105] and electrocatalysis.[106] The structured patterning of QDs on solid substrates is essential for many of these applications. Currently, the widely used inkjet printing technique for QD patterning suffers from several limitations such as nonuniform luminescence due to the coffee ring effect, poor printing resolution leading to large display pixels, and long postprocessing time required for the solvent evaporation from the patterned QD droplets. BPL is capable of overcoming these limitations. As an initial demonstration, Rajeeva et al. patterned a butterfly using red QDs without spreading or distortion of the printed lines (Figure 4A).[56] They also achieved multicolor QD printing, where QDs of different emission wavelengths were utilized to pattern the United States map (Figure 4B). The authors further demonstrated the printing of QDs on flexible PET substrate using BPL[56] for such applications as flexible and wearable displays.[107] BPL has also been exploited to modify the fluorescence characteristics of QDs. Tuning the laser power and exposure time was used to control the photon-induced oxidation of QDs[108] during the patterning process. The formation of the oxide layer on the QDs during BPL reduces the effective diameter of the QDs, which causes an increase in the quantum confinement and bandgap energy. This controllability is revealed by the tunable emission of the patterned QDs (Figure 4C). To explore the versatility of BPL and enhance the user experiences, Zheng and coworkers developed smartphone-controlled BPL, which takes a manually drawn pattern on a smartphone as the input and prints the corresponding QD patterns onto the substrate (Figure 4D).[79] Figure 4E shows the example of a printed QD spiral pattern that was drawn on the smartphone. Given the digital nature of this technique, the printed QD patterns can be easily altered or scaled on-demand via programmable control (Figure 4F). In addition, by tuning the handwriting speed on the smartphone, the printing speed of BPL can be altered to tune the fluorescence emission as a function of human physical motion. BPL with such versatile operation can serve as a highly accessible tool to manipulate matter at the nanoscale. With the further development along this line, one will expect smart BPL where internet-of-things, robotics, and cloud computing, and artificial intelligence are synergized to achieve digital micro/nanomanufacturing with full automation, high speed, mass customization, and global collaboration. The implementation of BPL for QD patterning has immense potential in many applications because QDs are not only used for displays but also in a variety of sensors.[109–111] Through the versatile control of BPL parameters, the patterned QDs can be tuned for optimal device performance. Furthermore, the plasmonic substrates used in BPL can also be utilized to enhance the performance of QD-based devices via plasmon-enhanced QD emission.[112] 3.2 | BPL for clinical detection of diabetes The capability of concentrating and printing molecules makes BPL a desirable tool that can be integrated into various sensors for the detection of low-concentrated biomolecules for disease diagnosis. Liu et al. implemented BPL on plasmonic chiral metamaterials for the diagnosis of diabetes through the ultrasensitive label-free detection of abnormal chiral metabolites in urine.[113] First, the urine samples were purified through centrifugation to remove all macromolecules and cells (Figure 5A). A plasmonic moire chiral metamaterial (MCM)[114–116,51] was used to generate optothermal microbubbles and enable the high-sensitive chiral measurement of immobilized molecules with plasmon-enhanced superchiral fields. The total molecular chirality was measured by recording the circular dichroism (CD) spectra.[117] Bubble-induced accumulation and printing of molecules on the plasmonic substrate causes different shifts in the CD peaks or dips of right-handed and left-handed MCMs (Figure 5B). By measuring the CD spectral shifts and the corresponding dissymmetry factors (the difference between the CD spectral shifts of left-handed and right-handed MCMs), abnormalities in molecular chirality of the lower-concentrated metabolites associated with diabetes are revealed. Figure 5C shows the measured dissymmetry factors due to the accumulation and printing of the molecules. The dissymmetry factor for BPL-assisted accumulation at the lower concentrations (100 pM–100 μM) of molecules is comparable to the dissymmetry factors for the non-BPL-assisted accumulation at a much higher concentration (10–100 mM). This result shows that the detection limit for BPL-assisted accumulation is enhanced by seven orders of magnitude, which is promising for the earlier disease diagnosis. By comparing the normalized dissymmetric factors for normal and diabetic urine samples, one can see that the diabetes-positive clinical sample has a larger normalized dissymmetry factor (Figure 5D). The accuracy of detection is validated with receiver operating curve analysis (Figure 5E), confirming a higher diagnostic accuracy of 84% than that of conventional enzyme tests at ~72%. Many chiral molecules in human body exhibit the phenomenon of homochirality, where one of the enantiomers is dominant in concentration than the other.[118,119] Progress in life sciences has revealed that more diseases can affect the relative concentrations of these enantiomers, which can be exploited for disease detection.[120] Compared with other chiral sensing techniques,[121] bubble accumulation-assisted sensors have the unique advantages of (1) ultrahigh sensitivity (~100 pM concentration), (2) low cost, and (3) high detection efficiency without sophisticated derivation and labeling, which are required with other methods, such as liquid chromatography coupled with mass spectrometry. 3.3 | BPL for enhanced analyte sensing in liquids Recent years have witnessed rapid progress in exploiting BPL to enhance the sensing of various analytes in liquid environments. For instance, Kotnala et al. demonstrated microbubble-assisted nanoaperture-based particle detection with fluorescence microscopy.[122] The bubble-induced analyte accumulation at the three-phase contact line and the plasmonic enhancement of optical signals from the accumulated analytes at the nanoapertures can overcome the diffusion-limited analyte delivery and increase the analyte fluorescence efficiency. As a result, the detection time was reduced from several minutes to less than 1 min, and the detection sensitivity was increased by one order of magnitude. Although BPL can be readily implemented for the accumulation of a variety of samples on the substrates, the temperature increase at the three-phase contact could deteriorate the analytes. Two strategies have been proposed to overcome the high-temperature issue. Tokonami et al. deposited a 50-nm gold (Au) layer to absorb the laser beam and generate a microbubble.[123] The continuous metal layer could effectively dissipate heat to reduce the effect of high temperature on the immobilized bacteria. A high concentration (107 cm−2) of bacteria assembly with a high survival rate of 80–90% was demonstrated. Although the temperature is reduced by a certain amount, it might not be sufficient to pattern more fragile biomolecules or cells without destroying their viability. For instance, using microbubbles for the detection of protein molecules would denature the proteins, as the denaturation temperature ranges between 29 and 99°C.[124] To considerably reduce the temperature near the three-phase contact line and enable protein-protein interaction studies, Kim et al. proposed a biphasic system with perfluoropentane (PFP) droplets dispersed in water.[125] PFP has a boiling point of 30°C, which makes it an ideal candidate for low-temperature BPL of proteins without denaturation. In addition, to enable the study of protein-protein interactions, the plasmonic substrate was modified with zwitterions to prevent unwanted immobilization of proteins during the BPL (Figure 6A). Fluorescence-labeled protein was used as a target sample to visualize the BPL-assisted surface capture event. Figure 6B shows the conjugated events in the event of microbubble accumulation for 1 min at different concentrations of fluorescent protein (0, 10, 20, 50, and 75 nM). This proof-of-concept work demonstrated the use of BPL to capture targeted proteins on functionalized substrates at low temperatures without any damage, which is essential for the broader implementation of BPL in biology. Biocompatible deposition of biomolecule-functionalized nanoparticles to optically transparent surfaces was demonstrated using shrinking surface plasmonic bubbles.[126] Another approach to implementing BPL in analyte sensing is based on surface-enhanced Raman spectroscopy (SERS). Researchers have implemented BPL to pattern silver-ring-based Raman substrates for the optical detection of analytes at sub-micro molar concentrations.[52] Upon the optothermal generation of bubbles, silver ions of the precursor solution were reduced at the three-phase contact lines to produce the silver rings due to the high ion accumulation and temperature. Karim et al. applied BPL to pattern plasmonic nanoparticles into nanogap-rich structures on the substrates and then to concentrate analytes on the nanogap-rich structures for SERS (Figure 6C).[127] Using Rhodamine 6G molecules as analytes, the authors demonstrated the bubble-enhanced SERS intensity for detecting low-concentrated molecules. The SERS-detectable molecular concentration is 10 pM without bubble (Figure 6D) and 1 pM with bubble (Figure 6E). Although this work involved BPL of nanoparticles and analytes in sequential patterning steps, the nanoparticles could also be mixed with analytes for single-step patterning. 3.4 | BPL for gas sensing Hydrogen presents as an attractive, clean energy alternative to replace fossil fuels due to its high energy density, and cost reduction of its production enables the realization of energy production in diverse applications at all scales.[128–130] Since hydrogen is a colorless, odorless, and highly flammable gas, the detection of production and leakage of hydrogen is essential to facilitate its applications as a safe energy source.[131,132] Palladium (Pd) is widely used to detect hydrogen via dissociative absorption. Current fabrication methods for Pd-based hydrogen sensors involve a two-step process: synthesis of Pd crystals and deposition of the crystals into prescribed patterns. BPL can achieve single-step fabrication of Pd sensors from the palladium precursor solution (Figure 7A).[133] A hydrogen gas sensor was prepared by printing colloidal Pd/Ni alloy nanoparticles in a predetermined path between Au pads (Figure 7B). As hydrogen gas was absorbed, the printed palladium expanded and created new pathways for electrons to flow, decreasing the resistivity between the Au pads. The resultant hydrogen sensor had a detection limit of 100 ppm H2 in air at room temperature (Figure 7C).[133] 3.5 | BPL for catalysis Metal nanoalloys have demonstrated enhanced catalytic properties compared with their individual counterparts.[134] However, synthesis of such nanoalloys, is difficult due to the complexity of synthetic techniques and chemical precursors.[135] This complexity arises from the preparation and containment of reagents in nanoscale spaces such as micelles.[136] Laser-induced microbubbles create an environment where high temperature and high precursor accumulation lead to the fast and micelle-free synthesis of nanoalloys.[133] To explore the formation dynamics of nanoalloys and their enhanced catalytic property, Zheng and coworkers have achieved a microbubble-mediated accumulation of Au and rhodium (Rh) ions for Au/Rh nanoalloy formation. Figure 8A shows the bubble printing of microscale Au/Rh nanoalloy lines from their precursor mixtures. To test the catalytic performance of the alloys, the authors carried out a reduction reaction of p-nitrophenol in the presence of pure Au, pure Rh, and the alloys. Figure 8B shows the ultraviolet (UV)-visible (vis) absorption spectra revealing the reduction reaction catalyzed by the different catalysts. A higher conversion percentage of the reactants catalyzed by the alloys demonstrates a better catalytic performance of the alloys than pure metals (Figure 8C). In another example, BPL was implemented to fabricate a catalytic chip based on a soft oxometalate (SOM) and a porous organic framework (POF) material.[137] Figure 8D shows the printed architecture of the SOM-POF composite. The chip was designed with various SOM-POF composites and tested against benchmark molecular catalysts, displaying high catalytic activity due to the high accessibility of mesospheres. The BPL-based trail had a resolution of 50 μm (Figure 8E). Figure 8F quantifies the catalytic ability through Raman peaks corresponding to the oxidation of benzaldehyde to benzoic acid. The peaks corresponding to the reaction products increase with time only on the printed trail, demonstrating the site-specific nature of the catalysis. In conclusion, the high-temperature zone at an optothermally generated bubble with highly concentrated ions creates a favorable environment for intensified nanoscale alloying.[138–141] With the capability of in situ synthesis and structuring of nanoalloys, BPL is instrumental in creating architected nanoalloys that exhibit superior catalytic activities without postprocessing. 3.6 | BPL for microrobotics Manipulation of micro-objects and biological materials is instrumental in diverse applications such as drug delivery, cell-cell interactions, and colloidal metamaterials.[142–145,70] Many micro-/nano-manipulation techniques exploiting electrical, thermal, optical, magnetic, and acoustic fields have been developed.[146–152] Magnetic manipulation is well-known for its noninvasive nature, high penetration depth, and high biocompatibility.[153,154] However, magnetic manipulation is limited to magnetic materials. Wang et al. overcame this limitation by implementing BPL for the fabrication of magnetic nanoparticles on various microscale and milliscale objects as substrates.[87] Specifically, they used a femtosecond laser for microbubble formation at the interface of the substrate and magnetic nanoparticle solution through two-photon absorption. Compared to the plasmonic-heating-mediated bubble generation, two-photon absorption requires higher laser power. The two-photon method is effective for BPL because the generated bubble can still drive Marangoni convection that attracts colloidal particles in the solution toward the substrate (Figure 9A). For instance, a 20-μm circular area of magnetic particles was deposited on a 100-μm nonmagnetic microstructure (Figure 9B). When exposed to an external magnetic field (e.g., a magnet), the magnetic nanoparticle aggregate on the microstructure responded to the external magnetic field, driving the microstructure to rotate at a high speed of ~1900 revolutions per second (Figure 9C). To demonstrate the wide applicability of BPL, printing of magnetic particles on different substrates such as glass, polymer, and biological material (living daphnia) was implemented. The on-demand BPL of magnetic particles and other functional materials on diverse substrates facilitate the fabrication of a wide range of hybrid structures for robotic and biological applications. 4 | PERSPECTIVE BPL has demonstrated its viability in many distinct fields. BPL is compatible with a wide range of substrates, including transparent substrates (glass), biological substrates (living daphnia), and flexible substrates (PET). In addition, the high integrability of the remotely controlled BPL with microfluidic systems is paving a way toward continuous cycling of multiple precursor solutions for automatic multi-material printing. Table 1 summarizes the recent progress of BPL in terms of applicable substrate, printed material, printing mode, printing resolution, laser wavelength and power, and printing speed. Despite the many advantages of BPL, one needs to address several issues for BPL to be implemented for a broader range of applications. 4.1 | Thickness and morphology control Optical heating strongly depends on the light-absorbing materials in the BPL system, which can be substrates or nanoparticle dispersions. In both scenarios, the deposited materials on the substrates can alter the local absorption of the laser beam and distort the bubble size during the patterning process.[155] The effect of the deposited materials becomes more prominent while performing multi-pass BPL for greater thickness. Although the thickness of the patterned structures increases during multi-pass, the rate at which thickness increases might not be a constant due to the influence of the deposited materials. A comprehensive relationship among material delivery, concentration, immobilization, and other physiochemical processes must be established for any targeted BPL system to precisely control the thickness of printed patterns. Similarly, the morphology of the printed patterns is currently not well controlled. Understanding the aggregate formation or chemical reaction near the three-phase contact line is essential to the printing of ink particles or reaction products into patterned structures of desired morphology. While smooth patterns are usually favorable for electronics, other applications such as catalysis benefit from the high surface roughness of the printed structures. Thus, one should achieve on-demand control of the morphology of patterned structures to enable the widespread applicability of BPL to different applications. 4.2 | Bubble stability and motion stability Two questions remain unanswered for high-speed laser scanning in the high-throughput BPL. One is whether the same bubble is stably manipulated for uniform patterning. The other is whether the Marangoni convection turns turbulent. For continuous patterning using BPL, the microbubble must exist throughout the patterning time. In discrete patterning, the microbubble is directed away from the substrate due to the sudden absence of laser heating and high buoyancy force on the bubble. However, new bubbles are formed immediately (in several milliseconds) at the new laser spots. This cyclic process of bubble generation, bubble movement, and bubble emigration from the substrate must be well understood and controlled to achieve any desired patterning. For example, small deficits in the patterns could occur during the bubble emigration stage, which result in electrical breakage of the printed conductors. To overcome this limitation, a better adhesion between the bubble and the substrate is necessary. In order to push toward the high-throughput BPL with high-speed laser scanning, one should employ an iterative process of substrate functionality design and study of the adhesion and drag forces on the bubble to develop a better understanding and control of bubble dynamics. A quantitative description of the bubble stability in terms of average bubble generation time and average bubble lifetime is required for better comparison across future works. An alternative method of understanding the efficiency of BPL as a lithographic approach is to quantify printing defects for benchmarking this technology for practical applications. During the extremely fast BPL, the high-speed bubble motion might result in a large Reynolds number, pushing Marangoni convection into the turbulent regime. Thus, another future research direction in high-throughput BPL is to understand the transition of convection into turbulent flows at high laser speeds and its impacts on patterning and alloying at the nanoscale. Alternatively, to overcome the throughput limitation in BPL due to a lack of understanding and control of bubble dynamics at high speeds, digital micromirror devices and spatial light modulators can be used to split one laser beam into multiple laser beams with independent control to opothermally generate multiple bubbles for parallel BPL. This multiple-beam flexibility can significantly increase the throughput of BPL, featuring advantages over other printing techniques like inkjet printing with a limited number of nozzle heads. However, a full understanding of interactions among different flows arising from individual bubbles in the arrays is needed to ensure the high-throughput BPL of desired structures in parallel. 5 | CONCLUSION We have reviewed progress in both experimental and theoretical fronts of BPL. Compared to other lithographical approaches, BPL has the unique advantages of concentrating ions, molecules, nanoparticles, and biomaterials before patterning them on the substrates. Such an accumulation process enables the use of low-concentrated solution-based inks or analyte samples without extensive pre-processing. Moreover, the small ink or sample volume (~100 μL) along with the low material wastage enables BPL to be a table-top prototyping apparatus or a point-of-care device. With the continuous efforts in advancing fundamental understanding of light-matter interactions and opto-thermo-fluidic multiphysics, colloidal chemistry, and instrumentation from multidisciplinary communities, we envision that BPL will be further improved to positively impact diverse industrial domains such as energy, information technology, and health care. ACKNOWLEDGMENTS The authors acknowledge the financial support of the National Institute of General Medical Sciences of the National Institutes of Health (grant number: DP2GM128446) and the National Science Foundation (grant numbers: NSF-ECCS-2001650 and NSF-CMMI-1761743). Jingang Li also acknowledges the financial support of University Graduate Continuing Fellowship from The University of Texas at Austin. Pavana Siddhartha Kollipara is a graduate student, pursuing his doctoral studies under the supervision of Dr. Yuebing Zheng in the Walker Department of Mechanical Engineering at the University of Texas at Austin. He received his B.Tech and M.Tech degrees in mechanical engineering at the Indian Institute of Technology, Madras, in 2017. His current research interests are thermoelectricity, optical trapping, and manipulation of colloids, optical lithography, and light–matter interactions. Ritvik Mahendra is an undergraduate student majoring in Electrical and Computer Engineering and Math at the University of Texas at Austin. His current research interests include optical lithography, with a scientific interest in the catalysis and sensing applications of printed lab-on-chip devices. Yuebing Zheng is an associate professor of mechanical engineering and materials science and engineering at The University of Texas at Austin. He is holding the Temple Foundation Endowed Teaching Fellowship in Engineering #2. He received his PhD in engineering science and mechanics from The Pennsylvania State University in 2010. He was a postdoctoral researcher at The University of California, Los Angeles, from 2010 to 2013. His research group (http://zheng.engr.utexas.edu) innovates optical manipulation and measurement for nanoscale, biological, and extraterrestrial world. DATA AVAILABILITY STATEMENT None. FIGURE 1 An overview of bubble-pen lithography (BPL): (A) Schematic of BPL. (B) Working principle of BPL. The surface tension (γ) decreases with increasing temperature (T). (C) Schematic of continuous printing via BPL. (D) Continuous printing of lines via BPL. (E) Schematic of discrete printing via BPL. (F) Discrete patterning of microparticle clusters via BPL. Scale bar: 5 μm; (A, B, and F) Reproduced with permission: 2016, American Chemical Society.[49] (C) Reproduced with permission: 2017, American Chemical Society.[56] (D) Reproduced with permission: 2020, American Chemical Society.[57] (E) Reproduced with permission: 2018, WILEY-VCH[52] FIGURE 2 Experimental tunability of bubble-pen lithography (BPL): (A) A typical optical setup of BPL. Reproduced with permission: 2016, American Chemical Society.[49] (B) Tunable printing linewidth by changing the laser intensity. Reproduced with permission: 2017, American Chemical Society.[56] (C) Tunable linewidth and fluorescence by changing the laser scanning speed. Reproduced with permission: 2017, The Royal Society of Chemistry.[79] (D) Tunable diameter of the printed rings by changing the laser exposure time. Reproduced with permission: 2013, American Chemical Society.[78] (E) Tunable bubble size and bubble growth rate by changing the gaseous content of the solution. Reproduced with permission: 2017, Americal Chemical Society.[81] (F) Tunable aggregation behavior by changing the particle concentration in the solvent from 4.55 × 108 particles/ml (left panel) to 4.55 × 104 particles/ml (right panel). The particle concentration decreased 10-fold for each panel from left to right. Reproduced with permission: 2019, American Chemical Society[82] FIGURE 3 Fundamental understanding of bubble-pen lithography (BPL): (A) Temperature distribution around an optothermal microbubble generated on a light-absorbing substrate. (B) Resultant Marangoni convection due to the surface tension gradient at the bubble, spanning several tens of micrometers. (A and B) Reproduced with permission: 2016, American Chemical Society.[49] (C) Random walk simulations demonstrating particle trajectories under the influence of Marangoni convection. Reproduced with permission: 2019, Elsevier Inc.[53] (D) Force analysis on the particles: Drag force (Fconv) attracts the particle toward the bubble because of Marangoni convection. In the vicinity of the bubble, the particle moves toward the three-phase contact line by the evaporation of the solution (Fevap). Once the particle touches the bubble surface, capillary force (Fcap), electrostatic force (Fe), Van der Walls force (Fvdw), and evaporative force (Fevap) determine the printability of the particle on the substrate surface. Reproduced with permission: 2021, American Chemical Society.[87] (E) For bubbles generated by light-absorbing particles, the direction of Marangoni flow on the bubble surface is opposite to that in the light-absorbing substrate. Inset: The heating source is the interaction of refracted laser beam and the light-absorbing particles above the substrate, which causes the local temperature to increase on the top of the bubble. (F) The particles are attracted toward the three-phase contact line due to the asymmetry in the refracted beam during laser motion. (E and F) Reproduced with permission: 2019, American Chemical Society[88] FIGURE 4 Bubble-pen lithography (BPL) for quantum dot (QD) patterning. (A) Fluorescence image of a butterfly pattern composed of red QDs. (B) Multistep BPL leading to a multicolor US map with the states of Texas, California, and Pennsylvania printed with different QDs. (C) Merged fluorescence images of yellow QDs printed in 20 μm × 20 μm squares with different parameters. Optical power, stage-translation speed, waiting time between neighboring lines, and line spacing for squares 1–4: (1) 0.52 mW/μm2, 1000 μm/s, 500 ms, and 1 μm; (2) 0.54 mW/μm2, 500 μm/s, 600 ms, and 1 μm; (3) 0.56 mW/μm2, 100 μm/s, 800 ms, and 1 μm; and (4) 0.58 mW/μm2, 100 μm/s, 1 s, and 0.5 μm. (A–C) Reproduced with permission: 2017, American Chemical Society.[56] (D) Illustration of smartphone-controlled BPL. QDs are immobilized on a substrate to create patterns with different emission wavelengths based on the variable hand movement on the phone screen. (E) QD printing of a hand-drawn spiral pattern on a smartphone screen. (F) QD printing of hand-drawn patterns combined with scaling factors. (D–F) Reproduced with permission: 2017, The Royal Society of Chemistry[79] FIGURE 5 Bubble-pen lithography (BPL) for clinical detection of diabetes. (A) Schematic of the collection, purification, accumulation, and immobilization of metabolites in urine for clinical detection of diabetes based on the molecular-chirality-dependent circular dichroism (CD) spectral shifts. (B) Evolution of CD spectra of left-handed and right-handed moire chiral metamaterial (MCM) through the successive microbubble-assisted accumulation of L-glucose and D-glucose. (C) CD spectral shifts (Δλ) and dissymmetry factors (ΔΔλ) due to adsorption of D-glucose and L-glucose with and without the use of the microbubble to accumulate the molecules. BPL enhanced the detection limit by eight-fold. (D) Normalized dissymmetry factors (ΔΔλ/λsum) for urine samples from normal and diabetic humans. (E) Receiver operating curves of normalized dissymmetry factor and glucose concentration to measure the accuracy of the testing process with the bubble concentration time of 5 s. (A–E) Reproduced with permission: 2021, American Chemical Society[113] FIGURE 6 Bubble-pen lithography (BPL) for enhanced analyte sensing in liquids. (A) Schematic showing the immobilization and capture of protein with and without the use of microbubble in a bi-phasic system. (B) Fluorescence images after concentration of fluorescein isothiocyanate (FITC)-protein A/G with the use of microbubble at varying concentrations. Scale bar: 5 μm. (A and B) Reproduced with permission: 2020, American Chemical Society.[125] (C) Schematic of active sensing of analytes using BPL with nanogap-rich architectures. (D) Surface-enhanced Raman spectra of R6G analyte without bubble. (E) Surface-enhanced Raman spectra of R6G analyte assisted by BPL. (C–E) Reproduced with permission: 2019, The Royal Society of Chemistry[127] FIGURE 7 Bubble-pen lithography (BPL) for gas sensing. (A) Flowchart outlining the fabrication of the sensor: (i) synthesis of organometallic-based nanoparticles; (ii) construction of conductive pads; (iii) deposition of the nanoparticles using BPL. (B) Scanning electron micrograph of the PdNi nanoparticle deposition between two Au electrodes. (C) The sensor response to various H2 concentrations with a constant voltage. (A–C) Reproduced with permission: 2019, WILEY-VCH[133] FIGURE 8 Bubble-pen lithography (BPL) for catalysis. (A) Scanning electron micrograph of bubble-printed nanoalloys composed of gold and rhodium with a 2-μm line spacing. (B) UV-vis absorption spectra demonstrate the reduction of p-nitrophenol by the printed Au-Rh nanoalloys as catalysts. (C) Conversion percentage of reactants into products shows that printed Au-Rh lines are more efficient for catalytic applications compared to SiO2, Au, and Rh control samples. (A–C) Reproduced with permission: 2019, Elsevier Inc.[53] (D) Optical image of bubble-printed soft oxometalate–porous organic framework (SOM-POF) composites for catalytic oxidation of benzaldehyde into benzoic acid. (E) The resolution of BPL of SOM-POF trails. (F) Raman spectroscopy depicting catalysis at trail site (on trail) versus away from the trial site (off trail). An increase in the intensity of the product benzoic acid (1010 nm peak) indicates the site-specific nature of catalysis. (D–F) Reproduced with permission: The Royal Society of Chemistry[137] FIGURE 9 Bubble-pen lithography (BPL) for micromachines. (A) Schematic outlining the micromachine skeleton and the printing of γ-Fe2O3 nanoparticles on a micromachine via BPL using a femtosecond laser. The printed magnetic nanoparticles act as a magnetic label, enabling the motion of the micromachine under an external magnetic field. (B) Optical image of a micromachine after BPL of nanomagnets. (C) Time-lapse images of micromachine rotation demonstrate the conversion of a passive structure into an active machine. (A–C) Reproduced with permission: 2021, American Chemical Society[87] TABLE 1 A summary of recent progress of bubble-pen lithography (BPL) Substrate Material Printing mode Laser wavelength Laser power/intensity Feature size Printing speed Reference Au nanoislands on glass or PET PS microparticles Discrete 532 nm 0.56 mW/μm2 (0.43 mW) ~2 μm – [49] Quantum dots Continuous 532 nm 0.3 mW/μm2 (0.23 mW) 500 nm 10 mm/s [56, 79] Au and Rh precursors Continuous 532 nm 0.6 mW/μm2 (0.47 mW) 550 nm 40 μm/s [53] Silver precursor Discrete 532 nm 0.2 mW/μm2 (0.16 mW) <1 μm (Ring: 3 μm) – [52] Au on Glass Pd and Ni precursors Continuous 532 nm 150 mW 4 μm 30 μm/s [133] Glass Polyaniline nanoparticles Continuous 532 nm 2.8 mW 2 μm 0.8 mm/s [57] Soft/poly oxometallates Continuous 1064 nm 25 mW 5 μm – [137] Soft/poly oxometallates Continuous 1064 nm 35 mW 5 μm 1 mm/s [78] Discrete 3 μm Iron oxide/Silver precursors 532 nm 13 mW  (Ring: 10 μm) 1 μm/s [156] Continuous 5 μm Glass/Polymer/Living substrates Magnetic microparticles Continuous 800 nm (fs laser) 10 mW 1 μm – [87] Abbreviations: BPL, bubble-pen lithography; PET, polyethylene terephthalate; PS, polystyrene. CONFLICT OF INTEREST The authors declare no conflict of interest. ETHICS STATEMENT All the authors have read and approved the manuscript, and given consent for submission and subsequent publication of the manuscript. ==== Refs REFERENCES 1. Zhang W , Wang R , Luo F , Wang P , Lin Z , Chin. Chem. Lett. 2020, 31 , 589. 2. Srisomwat C , Teengam P , Chuaypen N , Tangkijvanich P , Vilaivan T , Chailapakul O , Sens. Actuators B Chem. 2020, 316 , 128077. 3. Zhao C , Liu Q , Cheung KM , Liu W , Yang Q , Xu X , Man T , Weiss PS , Zhou C , Andrews AM , ACS Nano 2021, 15 , 904.33337135 4. Liu Q , Zhao C , Chen M , Liu Y , Zhao Z , Wu F , Li Z , Weiss PS , Andrews AM , Zhou C , iScience 2020, 23 , 101469.33083757 5. Moonen PF , Yakimets I , Huskens J , Adv. Mater. 2012, 24 , 5526.22887056 6. Peng Z , Ye R , Mann JA , Zakhidov D , Li Y , Smalley PR , Lin J , Tour JM , ACS Nano 2015, 9 , 5868.25978090 7. Malinowski PE , Ke TH , Nakamura A , Liu YH , Vander Velpen D , Vandenplas E , Papadopoulos N , Kronemeijer AJ , van der Steen JL , Steudel S , Kuo CC , Huang YY , Chen YH , Yeh MH , Gelinck G , Heremans P , J. Soc. Inf. Disp. 2018, 26 , 128. 8. Xiong J , Hsiang EL , He Z , Zhan T , Wu ST , Light Sci. Appl. 2021, 10 , 216.34697292 9. Quack N , Sattari H , Takabayashi AY , Zhang Y , Verheyen P , Bogaerts W , Edinger P , Errando-Herranz C , Gylfason KB , IEEE J. Quantum Electron. 2020, 56 , 8400210. 10. Ma J , Jiang L , Pan X , Ma H , Lin B , Qin J , Microfluid. Nanofluidics 2010, 9 , 1247. 11. Li N , Xu Z , Dong Y , Hu T , Zhong Q , Fu YH , Zhu S , Singh N , Nanophotonics 2020, 9 , 3071. 12. Wang X , Guo X , Ye J , Zheng N , Kohli P , Choi D , Zhang Y , Xie Z , Zhang Q , Luan H , Nan K , Kim BH , Xu Y , Shan X , Bai W , Sun R , Wang Z , Jang H , Zhang F , Ma Y , Xu Z , Feng X , Xie T , Huang Y , Zhang Y , Rogers JA , Adv. Mater. 2019, 31 , 1805615. 13. Sreenivasan SV , Microsyst. Nanoeng. 2017, 3 , 17075.31057889 14. Traub MC , Longsine W , Truskett VN , Annu. Rev. Chem. Biomol. Eng. 2016, 7 , 583.27070763 15. Suresh V , Ding L , Chew AB , Yap FL , ACS Appl. Nano Mater 2018, 1 , 886. 16. Ravi Kiran N , Chauhan M , Sharma SK , Ghosh S , Gonsalves KE , ACS Appl. Electron. Mater. 2020, 2 , 3805. 17. Pfeiffer HC , presented at in SPIE Photomask Technol, Monterey, CA, 2010. 18. Shimomura M , Sawadaishi T , Curr. Opin. Colloid Interface Sci. 2001, 6 , 11. 19. Ma ZC , Zhang YL , Han B , Chen QD , Sun HB , Small Methods 2018, 2 , 1700413. 20. Matsuhisa N , Kaltenbrunner M , Yokota T , Jinno H , Kuribara K , Sekitani T , Someya T , Nat. Commun. 2015, 6 , 7461.26109453 21. Zhang CJ , McKeon L , Kremer MP , Park SH , Ronan O , Seral-Ascaso A , Barwich S , Coileáin C , McEvoy N , Nerl HC , Anasori B , Coleman JN , Gogotsi Y , Nicolosi V , Nat. Commun. 2019, 10 , 1795.30996224 22. Wang Y , Liu Z , Huo N , Li F , Gu M , Ling X , Zhang Y , Lu K , Han L , Fang H , Shulga AG , Xue Y , Zhou S , Yang F , Tang X , Zheng J , Antonietta Loi M , Konstantatos G , Ma W , Nat. Commun. 2019, 10 , 5136.31723126 23. Yang Y , Liu R , Wu J , Jiang X , Cao P , Hu X , Pan T , Qiu C , Yang J , Song Y , Wu D , Su Y , Sci. Rep. 2015, 5 , 13480.26311022 24. Crista DMA , da Silva JCGE , da Silva LP , Nanomaterials 2020, 10 , 1316.32635483 25. Ma Z , Zhang Y , Han B , Hu X , Li C , Chen Q , Sun H , Nat. Commun. 2020, 11 , 4536.32913189 26. Sun Y , Dong W , Niu L , Jiang T , Liu D , Zhang L , Wang Y , Chen Q , Kim D , Sun H , Light Sci. Appl. 2014, 3 , e129. 27. Yang G , Park SJ , Materials (Basel). 2019, 12 , 1177.30978917 28. Danks AE , Hall SR , Schnepp Z , Mater. Horizons 2016, 3 , 91. 29. Sumida K , Liang K , Reboul J , Ibarra IA , Furukawa S , Falcaro P , Chem. Mater. 2017, 29 , 2626. 30. Grammatikopoulos P , Steinhauer S , Vernieres J , Singh V , Sowwan M , Adv. Phys. X 2016, 1 , 81. 31. Guan BY , Lu Y , Wang Y , Wu M , Lou XWD , Adv. Funct. Mater. 2018, 28 , 1706738. 32. Huang H , Zhou S , Yu C , Huang H , Zhao J , Dai L , Qiu J , Energy Environ. Sci. 2020, 13 , 545. 33. Levin A , Hakala TA , Schnaider L , Bernardes GJL , Gazit E , Knowles TPJ , Nat. Rev. Chem. 2020, 4 , 615. 34. Ikkala O , Ten Brinke G , Sci. 2002, 295 , 2407. 35. Ikkala O , Ten Brinke G , Chem. Commun. 2004, 19 , 2131. 36. Zhang H , Moon SK , Ngo TH , Int. J. Precis. Eng. Manuf. - Green Technol. 2020, 7 , 511. 37. Nayak L , Mohanty S , Nayak SK , Ramadoss A , J. Mater. Chem. C 2019, 7 , 8771. 38. Lohse D , Annu. Rev. Fluid Mech. 2022, 54 , 349. 39. Um SH , Chung YW , Seo Y , Seo H , Ok MR , Kim YC , Han HS , Chung JJ , Edwards JR , Jeon H , Adv. Funct. Mater. 2020, 30 , 2005233. 40. Yeo J , Hong S , Manorotkul W , Suh YD , Lee J , Kwon J , Ko SH , J. Phys. Chem. C 2014, 118 , 15448. 41. Yeo J , Hong S , Kim G , Lee H , Suh YD , Park I , Grigoropoulos CP , Ko SH , ACS Nano 2015, 9 , 6059.26035452 42. Yeo J , Hong S , Wanit M , Kang HW , Lee D , Grigoropoulos CP , Sung HJ , Ko SH , Adv. Funct. Mater. 2013, 23 , 3316. 43. Goodridge RD , Tuck CJ , Hague RJM , Prog. Mater. Sci. 2012, 57 , 229. 44. Pan H , Hwang DJ , Ko SH , Clem TA , Fréchet JMJ , Bäuerle D , Grigoropoulos CP , Small 2010, 6 , 1812.20665755 45. Kang B , Ko S , Kim J , Yang M , Opt. Express 2011, 19 , 2573.21369077 46. Fina F , Madla CM , Goyanes A , Zhang J , Gaisford S , Basit AW , Int. J. Pharm. 2018, 541 , 101.29454028 47. Liu S , Yuen MC , White EL , Boley JW , Deng B , Cheng GJ , Kramer-Bottiglio R , ACS Appl. Mater. Interfaces 2018, 10 , 28232.30045618 48. Dementyeva JN , Kashapov RN , Kashapov NF , Kashapov LN , IOP Conf. Ser. Mater. Sci. Eng. 2019, 570 , 012015. 49. Lin L , Peng X , Mao Z , Li W , Yogeesh MN , Rajeeva BB , Perillo EP , Dunn AK , Akinwande D , Zheng Y , Nano Lett. 2016, 16 , 701.26678845 50. Ghosh S , Ranjan AD , Das S , Sen R , Roy B , Roy S , Banerjee A , Nano Lett. 2021, 21 , 10.33296219 51. Wu Z , Kelp G , Yogeesh MN , Li W , McNicholas KM , Briggs A , Rajeeva BB , Akinwande D , Bank SR , Shvets G , Zheng Y , Nanoscale 2016, 8 , 18461.27778012 52. Rajeeva BB , Wu Z , Briggs A , Acharya PV , Walker SB , Peng X , Bahadur V , Bank SR , Zheng Y , Adv. Opt. Mater. 2018, 6 , 1701213. 53. Rajeeva BB , Kunal P , Kollipara PS , V Acharya P , Joe M , Ide MS , Jarvis K , Liu Y , Bahadur V , Humphrey SM , Zheng Y , Matter 2019, 1 , 1606. 54. Fang Z , Zhen Y-R , Neumann O , Polman A , García de Abajo FJ , Nordlander P , Halas NJ , Nano Lett. 2013, 13 , 1736.23517407 55. Baffou G , Polleux J , Rigneault H , Monneret S , J. Phys. Chem. C 2014, 118 , 4890. 56. Bangalore Rajeeva B , Lin L , Perillo EP , Peng X , Yu WW , Dunn AK , Zheng Y , ACS Appl. Mater. Interfaces 2017, 9 , 16725.28452214 57. Edri E , Armon N , Greenberg E , Hadad E , Bockstaller MR , Shpaisman H , ACS Appl. Mater. Interfaces 2020, 12 , 22278.32297505 58. Xu X , Luo J , Appl. Phys. Lett. 2007, 91 , 124102. 59. Chikazawa JI , Uwada T , Furube A , Hashimoto S , J. Phys. Chem. C 2019, 123 , 4512. 60. Zhao C , Xie Y , Mao Z , Zhao Y , Rufo J , Yang S , Guo F , Mai JD , Huang TJ , Lab Chip 2014, 14 , 384.24276624 61. Xu X , Liu W , Ji Z , Hao D , Yan W , Ye Z , Hu Y , Fang M , Wang C , Ma L , Huang J , Xu X , Weiss PS , ACS Mater. Lett. 2021, 3 , 1189. 62. Cheung KM , Yang KA , Nakatsuka N , Zhao C , Ye M , Jung ME , Yang H , Weiss PS , Stojanović MN , Andrews AM , ACS Sensors 2019, 4 , 3308.31631652 63. Nakatsuka N , Yang KA , Abendroth JM , Cheung KM , Xu X , Yang H , Zhao C , Zhu B , Rim YS , Yang Y , Weiss PS , Stojanović MN , Andrews AM , Sci. 2018, 362 , 319. 64. Kuang M , Wang J , Bao B , Li F , Wang L , Jiang L , Song Y , Adv. Opt. Mater. 2014, 2 , 34. 65. De Gans BJ , Schubert US , Langmuir 2004, 20 , 7789.15323532 66. Yang Q , Deng M , Li H , Li M , Zhang C , Shen W , Li Y , Guo D , Song Y , Nanoscale 2015, 7 , 421.25308163 67. Xie Y , Zhao C , Nanoscale 2017, 9 , 6622.28485456 68. Chen J , Kang Z , Wang G , Loo JFC , Kong SK , Ho HP , Lab Chip 2015, 15 , 2504.25963226 69. Lee Y , Kang BH , Kang M , Chung DR , Yi GS , Lee LP , Jeong KH , ACS Appl. Mater. Interfaces 2020, 12 , 12533.32101396 70. Liu Y , Lin L , Bangalore Rajeeva B , Jarrett JW , Li X , Peng X , Kollipara P , Yao K , Akinwande D , Dunn AK , Zheng Y , ACS Nano 2018, 12 , 10383.30226980 71. Lin L , Li J , Li W , Yogeesh MN , Shi J , Peng X , Liu Y , Rajeeva BB , Becker MF , Liu YYY , Akinwande D , Zheng Y , Adv. Funct. Mater. 2018, 28 , 1803990. 72. Yavas O , Ochiai C , Takai M , Appl. Phys. A Mater. Sci. Process. 1999, 69 , 875. 73. Serkov AA , Snelling HV , Heusing S , Amaral TM , Sci. Rep. 2019, 9 , 1773.30741970 74. Fujii S , Fukano R , Hayami Y , Ozawa H , Muneyuki E , Kitamura N , aki Haga M , ACS Appl. Mater. Interfaces 2017, 9 , 8413.28217991 75. Tsoulos TV , Tagliabue G , Nanophotonics 2020, 9 , 3849. 76. Zhang Q , Li R , Lee E , Luo T , Nano Lett. 2021, 21 , 5485.33939430 77. Yusupov VI , Tsypina SI , Bagratashvili VN , Laser Phys. Lett. 2014, 11 , 116001. 78. Roy B , Arya M , Thomas P , Jürgschat JK , Venkata Rao K , Banerjee A , Malla Reddy C , Roy S , Langmuir 2013, 29 , 14733.24171640 79. Rajeeva BB , Alabandi MA , Lin L , Perillo EP , Dunn AK , Zheng Y , J. Mater. Chem. C 2017, 5 , 5693. 80. Liu X , Bao L , Dipalo M , De Angelis F , Zhang X , Sci. Rep. 2015, 5 , 18515.26687143 81. Wang Y , Zaytsev ME , Le The H , Eijkel JCT , Zandvliet HJW , Zhang X , Lohse D , ACS Nano 2017, 11 , 2045.28088847 82. Yamamoto Y , Tokonami S , Iida T , ACS Appl. Bio Mater. 2019, 2 , 1561. 83. Zhan W , Shu Y , Sheng Y , Zhu H , Guo Y , Wang L , Guo Y , Zhang J , Lu G , Dai S , Angew. Chem. - Int. Ed. 2017, 56 , 4494. 84. Vinnacombe-Willson GA , Chiang N , Weiss PS , Tolbert SH , Scarabelli L , J. Chem. Educ. 2021, 98 , 546.34024937 85. Lombard J , Biben T , Merabia S , J. Phys. Chem. C 2017, 121 , 15402. 86. Takahashi K , Weng JG , Tien CL , Microscale Thermophys. Eng. 1999, 3 , 169. 87. Wang H , Xu B , Zhang Y , Kollipara PS , Liu S , Lin L , Chen Q , Zheng Y , Sun H , Nano Lett. 2021, 21 , 1628.33555185 88. Zhang Q , Pang Y , Schiffbauer J , Jemcov A , Chang HC , Lee E , Luo T , ACS Appl. Mater. Interfaces 2019, 11 , 48525.31794181 89. Armon N , Greenberg E , Layani M , Rosen YS , Magdassi S , Shpaisman H , ACS Appl. Mater. Interfaces 2017, 9 , 44214.29172418 90. Li J , Zheng Y , Accounts Mater. Res. 2021, 2 , 352. 91. Li J , Hill EH , Lin L , Zheng Y , ACS Nano 2019, 13 , 3783.30875190 92. Chen Z , Li J , Zheng Y , Chem. Rev. 2022, 122 , 3122.34797041 93. Kotov NA , Liz-Marzán LM , Weiss PS , ACS Nano 2021, 15 , 12457. 94. Nakatsuka N , Hasani-Sadrabadi MM , Cheung KM , Young TD ,Bahlakeh G , Moshaverinia A , Weiss PS , Andrews AM , ACS Nano 2018, 12 , 4761.29664607 95. Wang DH , Kim DY , Choi KW , Seo JH , Im SH , Park JH , Park OO , Heeger AJ , Angew. Chem. 2011, 123 , 5633. 96. Lewis JA , Ahn BY , Nature 2015, 518 , 42.25652992 97. Lin L , Zhang J , Peng X , Wu Z , Coughlan ACH , Mao Z , Bevan MA , Zheng Y , Sci. Adv. 2017, 3 , e1700458.28913423 98. Lin L , Lepeshov S , Krasnok A , Jiang T , Peng X , Korgel BA , Alù A , Zheng Y , Mater. Today 2019, 25 , 10. 99. Li J , Liu Y , Lin L , Wang M , Jiang T , Guo J , Ding H , Kollipara PS , Inoue Y , Fan D , Korgel BA , Zheng Y , Nat. Commun. 2019, 10 , 5672.31831746 100. Li J , Wang M , Wu Z , Li H , Hu G , Jiang T , Guo J , Liu Y , Yao K , Chen Z , Fang J , Fan D , Korgel BA , Alù A , Zheng Y , Nano Lett. 2020, 21 , 973.33372805 101. Biju V , Itoh T , Anas A , Sujith A , Ishikawa M , Anal. Bioanal. Chem. 2008, 391 , 2469.18548237 102. Mintz KJ , Zhou Y , Leblanc RM , Nanoscale 2019, 11 , 4634.30834912 103. Moon H , Lee C , Lee W , Kim J , Chae H , Adv. Mater. 2019, 31 , 1804294. 104. Hsieh YH , Hsu BW , Peng KN , Lee KW , Chu CW , Chang SW , Lin HW , Yen TJ , Lu YJ , ACS Nano 2020, 14 , 11670.32701270 105. Biondi M , Choi MJ , Wang Z , Wei M , Lee S , Choubisa H , Sagar LK , Sun B , Baek SW , Chen B , Todorović P , Najarian AM , Sedighian Rasouli A , Nam DH , Vafaie M , Li YC , Bertens K , Hoogland S , Voznyy O , García de Arquer FP , Sargent EH , Adv. Mater. 2021, 33 , 2101056. 106. Hoang VC , Dave K , Gomes VG , Nano Energy 2019, 66 , 104093. 107. Choi MK , Yang J , Hyeon T , Kim D-H , npj Flex. Electron. 2018, 2 , 10. 108. Hines DA , Becker MA , Kamat PV , J. Phys. Chem. C 2012, 116 , 13452. 109. Zhang J , Jin J , Wan J , Jiang S , Wu Y , Wang W , Gong X , Wang H , Chem. Eng. J. 2021, 408 , 127351. 110. Tajik S , Dourandish Z , Zhang K , Beitollahi H , Van Le Q , Jang HW , Shokouhimehr M , RSC Adv. 2020, 10 , 15406.35495425 111. Cui L , He XP , Chen GR , RSC Adv. 2015, 5 , 26644. 112. Wu J , Lee S , Reddy VR , Manasreh MO , Weaver BD , Yakes MK , Furrow CS , Kunets VSP , Benamara M , Salamo GJ , Mater. Lett. 2011, 65 , 3605. 113. Liu Y , Wu Z , Kollipara PS , Montellano R , Sharma K , Zheng Y , ACS Nano 2021, 15 , 6448.33760602 114. Wu Z , Liu Y , Hill EH , Zheng Y , Nanoscale 2018, 10 , 18096.30004551 115. Wu Z , Chen X , Wang M , Dong J , Zheng Y , ACS Nano 2018, 12 , 5030.29708728 116. Wu Z , Zheng Y , Adv. Opt. Mater. 2018, 6 , 1701057. 117. Neubrech F , Hentschel M , Liu N , Adv. Mater. 2020, 32 , 1905640. 118. Liu J , Mukherjee S , Wang F , Fischer RA , Zhang J , Chem. Soc. Rev. 2021, 50 , 5706.33972960 119. Chang KL , Pee HN , Yang S , Ho PC , Sci. Rep. 2015, 5 , 9000.25760794 120. Kimura T , Hamase K , Miyoshi Y , Yamamoto R , Yasuda K , Mita M , Rakugi H , Hayashi T , Isaka Y , Sci. Rep. 2016, 6 , 26137.27188851 121. Yu RB , Quirino JP , TrAC - Trends Anal. Chem. 2019, 118 , 779. 122. Kotnala A , Kollipara PS , Li J , Zheng Y , Nano Lett. 2020, 20 , 768.31834809 123. Tokonami S , Kurita S , Yoshikawa R , Sakurai K , Suehiro T , Yamamoto Y , Tamura M , Karthaus O , Iida T , Sci. Adv. 2020, 6 , eaaz5757.32158951 124. Ku T , Lu P , Chan C , Wang T , Lai S , Lyu P , Hsiao N , Comput. Biol. Chem. 2009, 33 , 445.19896904 125. Kim Y , Ding H , Zheng Y , Nano Lett. 2020, 20 , 7020.32667815 126. Moon S , Zhang Q , Huang D , Senapati S , Chang H , Lee E , Luo T , Adv. Mater. Interfaces 2020, 7 , 2000597. 127. Karim F , Vasquez ES , Sun Y , Zhao C , Nanoscale 2019, 11 , 20589.31638631 128. Hosseini SE , Wahid MA , Int. J. Energy Res. 2020, 44 , 4110. 129. Stern AG , Int. J. Hydrogen Energy 2018, 43 , 4244. 130. Poompavai T , Kowsalya M , Renew. Sustain. Energy Rev. 2019, 107 , 108. 131. Sulaiman A , Inambao F , Bright G , in Robot Intelligence Technology and Applications. Robot Intelligence Technology and Applications 3 (Eds.: Kim J-H , Yang W , Jo J , Sincak P , Myung H ), Advances in Intelligent Systems and Computing, Vol. 345 , Springer International Publishing, Cham, 2015, 819. 132. Hübert T , Boon-Brett L , Black G , Banach U , Sens. Actuators, B Chem. 2011, 157 , 329. 133. Rahamim G , Greenberg E , Rajouâ K , Favier F , Shpaisman H , Zitoun D , Adv. Mater. Interfaces 2019, 6 , 1900768. 134. Rabilloud F , Cottancin E , Pellarin M , Nanoalloys, Elsevier Inc., Amsterdam, Netherlands 2020. 135. Muñoz-Flores BM , Kharisov BI , Jiménez-Pérez VM , Elizondo Martínez P , López ST , Ind. Eng. Chem. Res. 2011, 50 , 7705. 136. Zhang S , Zhao Y , ACS Nano 2011, 5 , 2637.21366348 137. Thomas P , Pei C , Roy B , Ghosh S , Das S , Banerjee A , Ben T , Qiu S , Roy S , J. Mater. Chem. A 2015, 3 , 1431. 138. Yao Y , Dong Q , Hu L , Matter 2019, 1 , 1451. 139. Xie H , Chen F , Nie A , Pu T , Rehwoldt M , Yu D , Zachariah MR , Wang C , Shahbazian-Yassar R , Ju L , Hu L , Sci. 2018, 359 , 1489. 140. Xie P , Yao Y , Huang Z , Liu Z , Zhang J , Li T , Wang G , Shahbazian-yassar R , Hu L , Wang C , Nat. Commun. 2019, 10 , 4011.31488814 141. Wang C , Xie H , Zhang L , Gong Y , Pastel G , Dai J , Liu B , Wachsman ED , Hu L , Adv. Energy Mater. 2018, 8 , 1701963. 142. Li J , Chen Z , Liu Y , Kollipara PS , Feng Y , Zhang Z , Zheng Y , Sci. Adv. 2021, 7 , eabh1101.34172454 143. Lin L , Kollipara PS , Kotnala A , Jiang T , Liu Y , Peng X , Korgel BA , Zheng Y , Light Sci. Appl. 2020, 9 , 34.32194948 144. Kotnala A , Kollipara PS , Zheng Y , Nanophotonics 2020, 9 , 927.34290954 145. Peng X , Chen Z , Kollipara PS , Liu Y , Fang J , Lin L , Zheng Y , Light Sci. Appl. 2020, 9 , 141.32864116 146. Chiou PY , Ohta AT , Wu MC , Nature 2005, 436 , 370.16034413 147. Anhäuser A , Wunenburger R , Brasselet E , Phys. Rev. Lett. 2012, 109 , 034301.22861858 148. Tian Z , Yang S , Huang PH , Wang Z , Zhang P , Gu Y , Bachman H , Chen C , Wu M , Xie Y , Huang TJ , Sci. Adv. 2019, 5 , eaau6062.31172021 149. Ozcelik A , Rufo J , Guo F , Gu Y , Li P , Lata J , Huang TJ , Nat. Methods 2018, 15 , 1021.30478321 150. Jiang HR , Wada H , Yoshinaga N , Sano M , Phys. Rev. Lett. 2009, 102 , 100. 151. Nedev S , Carretero-Palacios S , Kühler P , Lohmüller T , Urban AS , Anderson LJE , Feldmann J , ACS Photonics 2015, 2 , 491.25950013 152. Lehmuskero A , Johansson P , Rubinsztein-Dunlop H , Tong L , Käll M , ACS Nano 2015, 9 , 3453.25808609 153. Lin Z , Fan X , Sun M , Gao C , He Q , Xie H , ACS Nano 2018, 12 , 2539.29443501 154. Huang TY , Sakar MS , Mao A , Petruska AJ , Qiu F , Chen XB , Kennedy S , Mooney D , Nelson BJ , Adv. Mater. 2015, 27 , 6644.26415002 155. Nishimura Y , Nishida K , Yamamoto Y , Ito S , Tokonami S , Iida T , J. Phys. Chem. C 2014, 118 , 18799. 156. Greenberg E , Armon N , Kapon O , Ben-Ishai M , Shpaisman H , Adv. Mater. Interfaces 2019, 6 , 1900541.

PMC010xxxxxx/PMC10350028.txt

==== Front bioRxiv BIORXIV bioRxiv Cold Spring Harbor Laboratory 10.1101/2023.07.04.547724 preprint 1 Article Modeling the Correlation between Z and B in an X-ray Crystal Structure Refinement Buscagan Trixia M. http://orcid.org/0000-0001-8242-9203 Rees Douglas C. http://orcid.org/0000-0003-4073-1185 04 7 2023 2023.07.04.547724http://biorxiv.org/lookup/doi/10.1101/2023.07.04.547724 nihpp-2023.07.04.547724.pdf Abstract We have examined how the refined B -factor changes as a function of Z (the atomic number of a scatterer) at the sulfur site of the [4Fe:4S] cluster of the nitrogenase iron protein by refinement. A simple model is developed that quantitatively captures the observed relationship between Z and B , based on a Gaussian electron density distribution with a constant electron density at the position of the scatterer. From this analysis, the fractional changes in B and Z are found to be similar. The utility of B -factor refinement to potentially distinguish atom types reflects the Z dependence of X-ray atomic scattering factors; the weaker dependence of electron atomic scattering factors on Z implies that distinctions between refined values of B in an electron scattering structure will be less sensitive to the atomic identity of a scatterer than for the case with X-ray-diffraction. This behavior provides an example of the complementary information that can be extracted from different types of scattering studies. ==== Body pmc

PMC010xxxxxx/PMC10350081.txt

==== Front bioRxiv BIORXIV bioRxiv Cold Spring Harbor Laboratory 10.1101/2023.07.07.548143 preprint 1 Article Sea lamprey enlightens the origin of the coupling of retinoic acid signaling to vertebrate hindbrain segmentation Bedois Alice M. H. http://orcid.org/0000-0002-6350-4017 Parker Hugo J. http://orcid.org/0000-0001-7646-2007 Bronner Marianne E. http://orcid.org/0000-0003-4274-1862 Krumlauf Robb http://orcid.org/0000-0001-9102-7927 07 7 2023 2023.07.07.548143http://biorxiv.org/lookup/doi/10.1101/2023.07.07.548143 nihpp-2023.07.07.548143.pdf Abstract Retinoic acid (RA) is involved in antero-posterior patterning of the chordate body axis and, in jawed vertebrates, has been shown to play a major role at multiple levels of the gene regulatory network (GRN) regulating hindbrain segmentation. Knowing when and how RA became coupled to the core hindbrain GRN is important for understanding how ancient signaling pathways and patterning genes can evolve and generate diversity. Hence, we investigated the link between RA signaling and hindbrain segmentation in the sea lamprey Petromyzon marinus , an important jawless vertebrate model providing clues to decipher ancestral vertebrate features. Combining genomics, gene expression, and functional analyses of major components involved in RA synthesis (Aldh1as) and degradation (Cyp26s), we demonstrate that RA signaling is coupled to hindbrain segmentation in lamprey. Thus, the link between RA signaling and hindbrain segmentation is a pan vertebrate feature of the hindbrain and likely evolved at the base of vertebrates. ==== Body pmc

PMC005xxxxxx/PMC5000010.txt

==== Front Sci RepSci RepScientific Reports2045-2322Nature Publishing Group srep3204010.1038/srep32040ArticleIce VII from aqueous salt solutions: From a glass to a crystal with broken H-bonds Klotz S. a1Komatsu K. 2Pietrucci F. 1Kagi H. 2Ludl A.-A. 1Machida S. 3Hattori T. 4Sano-Furukawa A. 4Bove L. E. 151 Institut de Minéralogie, de Physique des Matériaux et de Cosmochimie, CNRS UMR 7590, Université Pierre-et-Marie-Curie, F-75252 Paris, France2 Geochemical Research Center, Graduate School of Science, The University of Tokyo, Tokyo 113-0033, Japan3 CROSS-Tokai, Research Centre for Neutron Science and Technology, 162-1 Shirakata, Tokai, Ibaraki 319-1106, Japan4 J-PARC Center, Japan Atomic Energy Agency, Tokai, Naka, Ibaraki 319-1195, Japan5 Institute of Condensed Matter Physics, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerlanda Stefan.Klotz@impmc.upmc.fr26 08 2016 2016 6 3204014 05 2016 26 07 2016 Copyright © 2016, The Author(s)2016The Author(s)This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/It has been known for decades that certain aqueous salt solutions of LiCl and LiBr readily form glasses when cooled to below ≈160 K. This fact has recently been exploited to produce a « salty » high-pressure ice form: When the glass is compressed at low temperatures to pressures higher than 4 GPa and subsequently warmed, it crystallizes into ice VII with the ionic species trapped inside the ice lattice. Here we report the extreme limit of salt incorporation into ice VII, using high pressure neutron diffraction and molecular dynamics simulations. We show that high-pressure crystallisation of aqueous solutions of LiCl∙RH2O and LiBr∙RH2O with R = 5.6 leads to solids with strongly expanded volume, a destruction of the hydrogen-bond network with an isotropic distribution of water-dipole moments, as well as a crystal-to-amorphous transition on decompression. This highly unusual behaviour constitutes an interesting pathway from a glass to a crystal where translational periodicity is restored but the rotational degrees of freedom remaining completely random. ==== Body All forms of ice are made of hydrogen-bonded networks of H2O molecules with local tetrahedral coordination, see refs 1, 2, 3, 4 and references therein. Such geometry is thought to be incompatible with the presence of ionic species since the electric field of an ion interacts strongly with the electric dipole moment of the water molecule and perturbs the directional H-bonding to its neighbours5. Indeed, when salt water freezes it tends to expel the salt, as observed in the shelf ice of freezing sea water6. A related effect is the ‘salting-out’ of proteins by concentrated salt solutions5. In general, the interaction of water with dissolved ionic species is of fundamental interest for various fields of research ranging from life sciences to planetology. High-pressure ice in contact with salt is a major ingredient in the geology of satellites in the outer solar system78 and the understanding of the interior of these bodies rely critically on our knowledge of properties of aqueous salt solutions under strong compression. Recently, it has been shown that the most stable high-pressure ice phase, ice VII, can accommodate significant amounts of Li+ and Cl− ions, if prepared by pressure-induced crystallisation of the glassy solution9. In fact, in a narrow concentration range between R = 5.5 and R = 6.5, aqueous LiCl and LiBr solutions LiCl∙RH2O and LiBr∙RH2O easily form glasses, simply by cooling the liquid, see Fig. 1 1011. When such a solid solution with R = 6 is compressed below 100 K and heated, it crystallises into ‘salty ice VII’ where the Li-ions are trapped interstitially in the octahedral voids of the bcc-type ice VII structure, and the Cl− ion are located at water sites9. Such ‘doping’ of ice VII leads to an increased unit cell volume and a local perturbation of the hydrogen bond network which is visible in the diffraction patterns by an increased Debye-Waller factor. The very narrow range wherein such solid amorphous aqueous solutions can be prepared (Fig. 1) suggests that the crystallisation behaviour into ice VII depends sensitively on the salt concentration. In the following we show that this is indeed the case, i.e. that the structure of such crystallized salty ice VII is highly sensitive to the amount of salt inclusions, and that they give rise to salty ices with strongly expanded volumes, random water-dipole orientation and crystal-amorphous transitions under decompression at low temperatures. For this purpose, deuterated aqueous solutions with R = 5.6 of LiCl and LiBr (LiCl∙5.6D2O, and LiBr∙5.6D2O) were prepared. They present the limit of salt concentration where a homogeneous glass can easily form without applying extreme cooling speeds. We used deuterated samples to avoid the very large incoherent scattering of hydrogen, a standard method applied in powder neutron diffraction, including in almost all previous structural work on ice2349 (the isotope effect on the structure and phase diagram of ice is known to be negligible). The vitrified samples were then compressed at 80–85 K to pressures between 4 and 6 GPa and warmed to 300 K while keeping the pressure approximately constant. Neutron diffraction patterns were collected during the warm-up with accumulation times between 5 and 20 minutes. Figure 2 shows the behaviour of both the LiCl- and the LiBr-samples close to room temperature. Crystallisation of the glass is observed to start at ≈270 K and reaches a stable state slightly above 300 K. The diffraction patterns are characterized by a single, large peak at d ≈ 2.5 Å corresponding to the 110 Bragg reflection of ice VII. All other (sharp) reflections are either from a small amount of lead which was added as a pressure marker, or tungsten carbide which is the anvil material. Obviously, the high-pressure crystallized solids are far from being perfect crystals, but they certainly cannot be qualified as a glass either. In fact, the patterns are sufficiently crystalline to be able to analyze them by Rietveld methods, as shown in Fig. 3 for the two systems. Without surprise the fits are not perfect since they cannot reproduce the significant diffuse scattering which leads to the asymmetric wings around the 110 reflection. The large width of the reflection is most likely a particle size effect (typical size: 20 Å), and to a minor extent due to strain, as deduced from the refined peak shape parameters. The most surprising feature of these patterns is, however, the total absence of the 111 reflection which should occur at ≈2.0 Å (red arrows in Fig. 3), and which is clearly present in samples with lower salt concentrations9. This cannot be an effect of the strong Debye-Waller factor since the 211 reflection at lower d-spacing (≈1.35 Å, blue arrows in Fig. 3) is clearly visible, though weak, as expected. In the ice VII structure, the 111 reflection is entirely due to hydrogen (deuterium), i.e. it has no contribution from oxygen. Therefore, its absence is a clear and unambiguous proof of hydrogen being strongly displaced from its normal position along <x,x,x>, see Fig. 5 for a sketch of the ice VII structure and inset of Fig. 3 for a comparison with a pattern of pure ice VII. In initial fits this effect was simulated by varying displacement factors for hydrogen which gave <x2>1/2 ≈ 0.6 Å, which is indeed large. For comparison, the displacements in pure ice VII at 2.6 GPa and 295 K are <x2>1/2 ≈ 0.15–0.20 Å1213, i.e. more than three times smaller. A more realistic model with 8 positions was hence adopted, 4 positions along <x,x,x>, and 4 others at the anti-tetrahedral (‘interstitial’) positions, a geometry which describes a more isotropic distribution of H around O. Refinements in this model gave better fits with reduced H-displacements of <x2>1/2 ≈ 0.30 Å. Fits based on this model correspond to the lines through the data shown in Fig. 3, and results of the refinements are given in Table 1. In conclusion, whatever the detail of the model adopted, the evidence is that hydrogen is more or less isotropically disordered. This necessarily entails – given the fact that the molecular geometry is intact - that the electric dipole moment has no preferred orientation with respect to the crystallographic axes. The second remarkable feature is the strongly expanded lattice parameter: at the pressures measured by the lead marker the volume expansion relative to pure ice VII is 14% and 18% for the LiCl- and LiBr-system, respectively. These values scale exactly with the difference in size r of the anions involved, i.e. rCl = 1.81 Å and rBr = 1.96 Å, i.e. it is an unambiguous proof that the ions are indeed included into the structure, on specific lattice sites. Surprisingly, the volume expansion is not proportional to the salt content: For a 7% lower LiCl concentration (R = 6), the volume expansion was observed to be 43% smaller (i.e. only 8%)9. This sensitivity to the salt concentration must be related to the fact that – contrary to the low pressure ice forms - phase VII achieves its high density through interpenetration, i.e. by forming two interpenetrating but non-connected sublattices. Such a configuration permits, for example, to have the shortest O-O distances larger than in ordinary ice Ih, despite its considerably larger density. The destruction of the interpenetrating sublattices by the presence of ionic species must necessarily lead to an increase of volume which is expected to be highly sensitive to the ion concentration. The structural properties of these highly salt-loaded ice phases were then investigated by molecular dynamics simulations which give access to physical properties which cannot be extracted by diffraction measurements alone. For this purpose simulations were carried out with 15 × 15 × 15 ice VII supercells corresponding to 19227 atoms. Note that the structural results are insensitive to the choice of the isotope (hydrogen or deuterium) since the potentials used in the simulations are the same. Details of the calculations are described in the Methods section. Figure 4 shows a snapshot of typical molecular configurations in LiCl∙5.6H2O compared to pure ice VII, viewed along the cubic [100] direction. Results for the LiBr-system are qualitatively similar and are not shown here. It is seen that inclusion of Li+ and Cl− leads to molecular displacements of several tenths of angstroms, but on average the H2O molecules are still at the bcc sites of ice VII. It is also seen that the Li+-ions prefer to locate themselves in the octahedral interstitials, whereas the much larger Cl-ions are on average closer to sites where H2O is expected. Radial distribution functions were calculated for the two systems and can be found as Supplementary Information. These indicate an average of three H-atoms up to a distance of 2.1 Å compared to four in pure ice VII, suggesting that a sizeable number of H-bonds are broken. The most important detail in this context concerns however the spatial distribution of dipole moments. For this purpose the orientation of each water molecule was extracted from the last part of the 12 ns trajectory (1 ns) and employed to build the spherical probability maps shown in Fig. 5. The probability density is normalized so as to give an integral over all orientations equal to one. For comparison, a similar-size model of ice VII was also built and equilibrated at constant pressure P = 10 GPa and T = 300 K (see Methods for the choice of the pressure). As shown in the Fig. 5, pure ice VII is characterized by molecular dipoles sharply aligned along [100] and equivalent directions, as expected. In LiCl∙5.6H2O and LiBr∙5.6H2O, however, the orientational probability is almost exactly isotropic. Close inspection of the dipole distribution shows a small residual preference for <100> directions in the case of LiCl∙5.6H2O which is hardly discernible in Fig. 5, but an almost isotropic distribution for the Br-system. This is confirmed by calculating the standard deviation σ of the probability over all dipole orientations. For the Cl-system we find σ(LiCl) = 0.012, for the Br-system σ(LiBr) = 0.0090, with average probabilities of 0.075 (LiCl) and 0.078 (LiBr), i.e. relative variations of only 16% and 11%, respectively. The simulations therefore come to the same conclusion as the diffraction data: The ice VII-like structures emerging from crystallisation of highly concentrated LiCl and LiBr-solutions under pressure have completely disordered molecular orientations which means that a large fraction of hydrogen-bonds are broken and that the ice-rules are violated. This distinguishes it sharply from the orientational disorder present in most of the (pure) ice phases where the ice-rules are conserved. The crystallisation observed here could be classified as ‘fractional’ since it concerns mainly the three translational degrees of freedoms whereas the three rotational freedoms remain uniformly random. Breaking of hydrogen bond networks is a phenomenon which has been observed in liquid salt solutions at ambient pressure1014, and also in pure liquid water above the critical point15. For solid water, however, including amorphous ice phases, this seems to be a new observation which is only possible by crystallisation of the solutions under pressure. In this context the question arises if the inclusion of ionic species into ice VII promotes “plasticity”, i.e. the formation of an ice VII phase with freely rotating molecules, as predicted by several simulations161718 to occur close to the melting line. A diffraction experiment can – by principle – not give information on the dynamics of molecular orientations, but our simulations can safely exclude this scenario. In fact, the calculated orientational correlation times of water dipoles in LiCl∙5.6H2O and LiBr∙5.6H2O are typically 4–5 times longer compared to pure ice VII at the same pressure. Therefore, although the ions severely perturb the H-bond network, their presence seems to hamper their orientational dynamics rather than to promote it. A further interesting aspect concerns the behaviour under de-compression at low temperatures: The two salty ice-forms are found to be not quenchable at 90 K, and probably even not at any temperature below, contrary to all forms of pure ice. Figure 6 shows diffraction patterns of the two systems at ambient pressure and the same temperature, before compression and after the compression/crystallisation cycle, after decompression at ≈90 K. The surprising finding is that both ice VII forms of LiCl∙5.6D2O and LiBr∙5.6D2O return to an amorphous solid, a disordered system which is significantly different from the original glass. A comparison with published structure factors from ref. 19 shows a striking resemblance with high density amorphous ice (HDA), except that the main diffraction feature is at ≈0.2 Å shorter d-spacing, approximately at a position where it is found in very high density amorphous ice (VHDA)20. From this we suspect that upon decompression, the hydrogen-bond network might be partially restored with a slightly higher molecular density than found in HDA. Hence, we observe a strong memory effect: the vitrified glass has a memory of the liquid with its structure governed by hydration of the ionic species, whereas the de-compressed amorphous ice has kept a memory of the crystalline ice VII, where the cohesion is governed by the need for maximizing the packing density. In conclusion, we have presented neutron diffraction and molecular simulation results on highly concentrated LiCl- and LiBr aqueous solutions which were crystallized under high pressure. They indicate that the ionic species can be trapped inside a strongly expanded ice VII – like structure with random electric dipole orientation. These structures cannot be quenched (‘recovered’) to ambient pressure and low temperatures but transform back to a glass with a structure different to the initial one. The trapping of ionic species by crystallisation from a glass (devitrification) might be found in other network forming glasses, for example in silica. In fact, “stuffed” SiO2 polymorphs incorporating small ions such as lithium are well known and can be formed at ambient pressure21. The high-pressure route used in this study might enable to synthesize “stuffed” high pressure silica phases which are expected to have remarkable properties, for example improved hardness, and therefore might be of technological interest. Methods Sample preparation Aqueous solutions of LiCl and LiBr with R = 5.6 were prepared from the salts purchased from Aldrich, and 99.8% enriched heavy water (D2O) from Eurisotop, France. The solutions were filtered to remove particles larger than 3 μm to minimize the possibility of crystallisation during cooling, and loaded into TiZr gaskets together with a small piece of lead (10 mg). After sealing the gasket with a load of 10 kN onto the anvils the cell was cooled at a speed of 10 K/min to 80 K which transformed the liquids into pure glasses. The samples were then compressed at 80–110 K in steps up to 4–5 GPa and then heated at an average speed of typically 1 K/min up to the temperature where crystallisation occurred. After crystallisation, the samples were stable over at least the timescale of the diffraction experiments, i.e. 1–3 hours. Samples at lower pressures (below ca. 3.3 GPa for the LiCl compound and ca. 4.3 GPa for the LiBr system) crystallized into solids with unknown structure. Neutron diffraction High pressure neutron diffraction measurements were carried out at the Japanese high pressure beamline PLANET22 at MLF, the Japan Accelerator Research Complex (J-PARC), Tokai, Ibaraki, Japan, using a variable P-T hydraulic press (‘Mito system’)23, and tungsten carbide anvils with a profile as described in ref. 24 and 9/6 mm outer/inner diameters. Encapsulating gaskets were made of null-scattering TiZr, surrounded by a supporting ring of a high-tensile aluminium alloy24. Normalisation of the data was achieved in a separate run using a vanadium pellet and an anvil configuration which was strictly similar. Molecular dynamics simulations Salty ices of stoichiometry Li(Cl/Br)∙5.6H2O have been simulated employing the TIP4P/Ew interatomic potential for water25 and the potentials of Joung & Cheatham for Li+, Cl−, and Br− ions26. The potentials for hydrogen and deuterium are strictly identical which allows us to make meaningful comparison with the neutron diffraction data which used deuterated samples. Molecular dynamics simulations were performed employing the software Gromacs27. The temperature was controlled with the stochastic velocity rescaling thermostat28, and the pressure with a Berendsen barostat29. For each composition (Cl or Br) samples were generated in the following way: Ice VII 15 × 15 × 15 supercells with 19227 atoms were constructed employing the experimental lattice parameter (a = 3.42835 Å with Cl and a = 3.45795 Å with Br) and positioning 6750 water molecules at ideal bcc lattice sites with a random orientation of dipoles, imposing the ice rules. 1023 water molecules, selected at random, were removed and replaced with either Cl− or Br− ions. Finally, 1023 Li+ ions were introduced choosing randomly their positions among the centers of octahedra defined by water molecules and halide ions. These initial geometries were constructed avoiding to place two identical ions closer than the lattice parameter a. The composition corresponds to a stoichiometry Li(Cl/Br)∙5.63H2O. Each structure was initially relaxed at T = 10 K and constant volume for 30 ps (with a timestep of 0.1 fs). Then, the system was slowly heated up to T = 300 K employing a time constant of 50 ps for the thermostat (with a timestep of 1 fs) and equilibrated for 12 ns. The dipole moment orientation of each water molecule was extracted from the last part of the trajectory (1 ns) and employed to build the spherical probability maps. The orientational correlation time of water dipoles was estimated performing, for each system, additional 50 ns-long simulations using deuterium instead of hydrogen masses. In a second set of simulations a smaller 4 × 5 × 7 cell was used to generate 100 configurations which served to calculate neutron diffraction patterns and verify the consistency with experiment. Note that constraining the system at the experimental volume corresponds to a pressure of about 10 GPa: this is consistent with the domain of stability of ice VII with the interatomic potential employed30, whereas imposing the experimental pressure (4–5 GPa) to the theoretical samples would correspond to the stability domain of the liquid resulting in much more disordered structures. Additional Information How to cite this article: Klotz, S. et al. Ice VII from aqueous salt solutions: From a glass to a crystal with broken H-bonds. Sci. Rep. 6, 32040; doi: 10.1038/srep32040 (2016). Supplementary Material Supplementary Information This work was supported by the French-Japanese PHC grant 29717X between IMPMC and the University of Tokyo (P.I.: S.K. and H.K.), the JSPS KAKENHI grant no. 26246039, the cluster of excellence MATériaux Interfaces Surfaces Environnement (MATISSE) led by Sorbonne Universités, through a PhD fellowship for A.-A.L. and by the ANR within the Blanc International program PACS under reference no. ANR-13-IS04-0006-01. The experiment at the Materials and Life Science Experimental Facility of the J-PARC was performed under a user program (Proposal No. 2014A0048). The authors wish to thank Dr. J. Abe for much assistance during the experiments at the PLANET beamline and D. Bowron, L. Cormier, Ch. Sanloup and H. Fischer for helpful discussions. Author Contributions S.K. and L.E.B. proposed the experiment, S.K., A.-A. L., K.K., H.K., S.M., T.H. and A.S.-F. carried out the experiments. S.K. did the neutron data analysis and wrote the paper. F.P. did the molecular dynamics simulations. Figure 1 Left panel: Schematic metastable phase diagram of LiCl∙RH2O and LiBr∙RH2O, redrawn after refs 1010,11. In a small region of concentration R ≈ 5.5–6, the supercooled liquid transforms directly into a glass. Regions to the left and the right are stability domains of ice and various hydrates. Right panel: Pressure-temperature path applied to form LiCl- and LiBr containing ice VII together with the phase diagram of pure water. Red dots indicate initial sample conditions at ambient pressure. Figure 2 Neutron diffraction patterns taken during crystallisation of compressed LiCl∙5.6D2O (left, accumulation time 5 min) and LiBr∙5.6D2O (right, accumulation time 10 min) upon warming. Sharp reflections are of a small quantity of lead (Pb) mixed with the sample which serves as pressure marker, the asterisk is a reflection of the tungsten carbide anvils. The high temperature patterns were shifted vertically to avoid overlap. Figure 3 Neutron diffraction patterns (circles) and Rietveld fits (lines) to the data using the model described in the text. Sharp peaks are from lead, the asterisk is a reflection from the tungsten carbide anvils. Expected positions of 111 and 211 reflections are indicated by red and blue arrows, respectively. Upper and lower tick marks correspond to reflections of ice VII and lead, respectively, the difference curves are shown below. Accumulation time is 15 (left) and 30 minutes (right). For comparison, a pattern of well-crystallised pure ice VII at 4.7 GPa is shown in the inset including Miller indices of the three strongest reflections. Crosses indicate reflections of the lead pressure marker. Figure 4 Snapshots of molecular configuration in pure ice VII (a) and LiCl∙5.6H2O (b), viewed along the cubic [100] direction. Li-ions are drawn in yellow, Cl-ions in green. The dashed lines are guides to the eye. Figure 5 Probability density distribution of the electric dipole moment of H2O in pure ice VII (upper left), LiCl∙5.6H2O (upper right) and LiBr∙5.6H2O (lower right), derived from molecular dynamics simulations. The distribution was normalized to give an integrated probability equal to one. Note a very weak preference of dipoles along <100> in the LiCl-system, and a completely flat distribution for the LiBr-system. A sketch of a typical instantaneous configuration in pure ice VII is given (lower left) with hydrogen bonds indicated as dashed lines and the non-bonding (“anti-tetrahedral”) directions as dotted lines. Figure 6 Neutron diffraction patterns at ambient pressure, before compression of the glass, and after decompression of the crystallized phases (see Fig. 3). For comparison, the solid line represents neutron data for high density amorphous ice (HDA) from ref. 19. Table 1 Structural parameters derived from Rietveld refinements to neutron diffraction patterns shown in Fig. 3.   LiCl∙5.6D2O LiBr∙5.6D2O Pressure (GPa) 4.8(1) 5.2(1) a (Å) 3.4284(17) 3.4580(20) ΔV/V 14% 18% <x2>1/2 (Å) D/O 0.31/0.18 0.50/0.25 Data were refined in pace group of ice VII1213 with oxygen at (0.25, 0.25, 0.25), hydrogen (deuterium) at the 4 tetrahedral (0.41, 0.41, 0.41) positions as well as 4 further “anti-tetrahedral” sites at (0.08, 0.08, 0.08). Li was placed at interstitial octahedral sites at (0.75, 0.75, 0.25) and Cl (Br) at O-sites. Pressure values were determined from the refined lattice parameters of lead which was added to the sample31. a is the refined lattice parameter of the samples, ΔV/V the volume expansion compared the pure ice VII, using the equation of state of ref. 32, <x2>1/2 is the mean square displacement of deuterium/oxygen. ==== Refs Petrenko V. F. & Whitworth R. W. Physics of Ice , Oxford University Press (Oxford, 2002 ). Bartels-Rausch Th. . Ice structures, patterns and processes: A view across the icefield . Review of Modern Physics 84 , 885 –944 (2012 ). Salzmann Ch. G. , Radaelli P. G. , Hallbrucker A. , Mayer E. & Finney J. L. The preparation and structures of hydrogen ordered phases of ice . Science 311 , 1758 –1761 (2006 ).16556840 Falenty A. , Hansen Th. C. & Kuhs W. F. Formation and properties of ice XVI obtained by emptying a type sII clathrate hydrate . Nature 516 , 231 –233 (2014 ).25503235 Marcus Y. Effect of ions on the structure of water: Structure making and breaking . Chem. Rev. 109 , 1346 –1370 (2009 ).19236019 Lake R. A. & Lewis E. L. Salt rejection by sea ice during growth . Journal of Geophys. Research 75 , 583 –597 (1970 ). Frank M. R. , Runge C. E. , Scott H. P. , Maglio S. J. , Olson J. , Prakapenka V. B. & Shen G. Experimental study of the NaCl-H2O system up to 28 GPa: Implications for the ice-rich planetary bodies . Phys. Earth Planet. Interiors 155 , 152 –162 (2008 ). Journaux B. , Daniel I. , Caracas R. , Montagnac G. & Cardon H. Influcence of NaCl on ice VI and ice VII melting curves up to 6 GPa, implications for large icy moons . Icarus 226 , 355 –363 (2013 ). Klotz S. , Bove L. E. B. , Strässle Th. , Hansen Th. C. & Saitta A. M. The preparation and structure of salty ice VII under pressure . Nature Materials 8 , 405 –409 (2009 ).19349969 Prével B. , Jal J. F. , Dupuy-Philon J. & Soper A. K. Structural characterization of an electrolytic aqueous solution, LiCl∙6H2O, in the glass, supercooled, and liquid states . The Journal of Chemical Physics 103 , 1886 –1896 (1995 ). Kessis J. J. Le système eau-bromure de lithium . Bulletin de la société chimique de France , 48 –52 (1965 ). Kuhs W. F. , Finney J. L. , Vettier C. & Bliss D. V. Structure and hydrogen ordering in ices VI, VII, and VIII by neutron powder diffraction . The Journal of Chemical Physics 81 , 3612 –3623 (1984 ). Jorgensen J. D. & Worlton T. G. Disordered structure of D2O ice VII from in situ neutron powder diffraction . The Journal of Chemical Physics 83 , 329 –333 (1985 ). Tromp R. H. H. , Neilson G. W. & Soper A. K. Water structure on concentrated lithium chloride solutions . The Journal of Chemical Physics 96 , 8460 –8469 (1992 ). Postorino P. , Tromp R. H. , Ricci M. A. , Soper A. K. & Neilson G. W. The interatomic structure of water at supercritical temperatures . Nature 366 , 668 –670 (1993 ). Takii Y. , Koga K. & Tanaka H. A plastic phase of water from computer simulation . The Journal of Chemical Physics 128 , 204501 –8 (2008 ).18513026 Aragones J. L. & Vega C. Plastic crystal phases of simple water models . The Journal of Chemical Physics 130 , 244504 –14 (2009 ).19566163 Matsumoto M. , Himoto K. & Tanaka H. Spin-one Ising model for ice VII-plastic ice phase transformations . Journal of Physical Chemistry B 118 , 13387 –13392 (2014 ).25144575 Bellissent-Funel M. C. , Teixeira J. & Bosio L. Structure of high-density amorphous water. II. Neutron scattering study . The Journal of Chemical Physics 87 , 2231 –2235 (1987 ). Bowron D. T. , Finney J. L. , Hallbrucker A. , Kohl I. , Loerting T. Mayer E. & Soper A. K. The local and intermediate range structures of the five amorphous ices at 80 K and ambient pressure: A Faber-Ziman and Bhatia-Thornton analysis . The Journal of Chemical Physics 125 , 194502 –14 (2006 ).17129118 Palmer D. C. Stuffed derivatives of the silica polymorphs . In: Silica – Physical behavior, geochemistry and materials applications (editors Herney P. J. , Prewitt C. T. & Gibbs G. V. ) 83 –122 , Review in Mineralogy, vol. 29 (Mineral Soc. of America, Washington D.C., 1994 ). Hattori T. . Design and performance of high-pressure PLANET beamline at pulsed neutron source at J-PARC . Nucl. Instrum. Methods in Phys. Res. A 780 , 55 –67 (2015 ). Komatsu K. , Moriyama M. , Koizumi T. , Nakayama K. , Kagi H. , Abe J. & Hario S. Development of a new P-T controlling system for neutron-scattering experiments , High Pressure Res. 33 , 208 –213 (2013 ). Komatsu K. , Klotz S. , Shinozaki A. , Iizuka R. , Bove L. E. & Kagi H. Performance of ceramic anvils for high pressure neutron scattering , High Pressure Res. 34 , 494 –499 (2014 ). Horn H. W. , Swope W. C. , Pitera J. W. , Madura J. D. , Dick T. J. , Hura G. L. & Head-Gordon T. Development of an improved four-site water model for biomolecular simulations: TIP4P-Ew . The Journal of Chemical Physics 120 , 9665 –9678 (2004 ).15267980 Joung I. S. & Cheatham T. E. III Determination of alkali and halide monovalent ion parameters for use in explicitly solvated biomolecular simulations . The Journal of Phys. Chem. B 112 , 9020 –9041 (2008 ).18593145 Hess B. , Kutzner C. , Van Der Spoel D. & Lindahl E. GROMACS 4: algorithms for highly efficient, load-balanced, and scalable molecular simulation . J. Chem. Theory Comput. 4 , 435 –447 (2008 ).26620784 Bussi G. , Donadio D. & Parrinello M. Canonical sampling through velocity rescaling . The Journal of Chemical Physics 126 , 014101 (2007 )17212484 Berendsen H. J. C. , Postma J. P. M. , van Gunsteren W. F. , DiNola A. & Haak J. R. Molecular dynamics with coupling to an external bath . The Journal of Chemical Physics 81 , 3684 –3690 (1984 ). Vega C. , Abascal J. L. F. , Sanz E. , MacDowell L. G. & McBride C. Can simple models describe the phase diagram of water? Journal of Physics: Condens . Matter 17 , S3283 –S3288 (2005 ). Strässle Th ., Klotz S. , Kunc K. , Pomjakushin V. & White J. S. Equation of state of lead from high-pressure neutron diffraction up to 8.9 GPa and its implication for the NaCl pressure scale . Phys. Rev. B 90 , 014101 –8 (2014 ). Besson J. M. , Pruzan Ph. , Klotz S. , Hamel G. , Silvi B. , Nelmes R. J. , Loveday J. , Wilson R. M. & Hull S. Variation of interatomic distances in ice VIII to 10 GPa . Phys. Rev. B 49 , 12540 –12551 (1994 ).

PMC005xxxxxx/PMC5000011.txt

==== Front Sci RepSci RepScientific Reports2045-2322Nature Publishing Group srep3207510.1038/srep32075ArticleSharp Contradiction for Local-Hidden-State Model in Quantum Steering Chen Jing-Ling a12Su Hong-Yi b13Xu Zhen-Peng 1Pati Arun Kumar c41 Theoretical Physics Division, Chern Institute of Mathematics, Nankai University, Tianjin 300071, People’s Republic of China2 Centre for Quantum Technologies, National University of Singapore, 3 Science Drive 2, Singapore 117543, Singapore3 Department of Physics Education, Chonnam National University, Gwangju 500-757, Republic of Korea4 Quantum Information and Computation Group, Harish-Chandra Research Institute, Chhatnag Road, Jhunsi, Allahabad 211 019, Indiaa chenjl@nankai.edu.cnb hysu@mail.nankai.edu.cnc akpati@hri.res.in26 08 2016 2016 6 3207518 05 2016 02 08 2016 Copyright © 2016, The Author(s)2016The Author(s)This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/In quantum theory, no-go theorems are important as they rule out the existence of a particular physical model under consideration. For instance, the Greenberger-Horne-Zeilinger (GHZ) theorem serves as a no-go theorem for the nonexistence of local hidden variable models by presenting a full contradiction for the multipartite GHZ states. However, the elegant GHZ argument for Bell’s nonlocality does not go through for bipartite Einstein-Podolsky-Rosen (EPR) state. Recent study on quantum nonlocality has shown that the more precise description of EPR’s original scenario is “steering”, i.e., the nonexistence of local hidden state models. Here, we present a simple GHZ-like contradiction for any bipartite pure entangled state, thus proving a no-go theorem for the nonexistence of local hidden state models in the EPR paradox. This also indicates that the very simple steering paradox presented here is indeed the closest form to the original spirit of the EPR paradox. ==== Body In 1935, Einstein, Podolsky and Rosen (EPR) questioned the completeness of quantum mechanics under the assumption of locality and reality1 that underlie the classical world view. By considering continuous-variable entangled state, EPR proposed a famous thought experiment that involves a dilemma concerning local realism against quantum mechanics. This dilemma is nowadays well-known as the EPR paradox. For a long time, the EPR argument remained a philosophical problem at the foundation of quantum mechanics. In 1964, Bell made an important step forward2 by considering a version based on the entanglement of spin-1/2 particles introduced by Bohm. The EPR paradox, according to Bell’s reasoning, could, supposedly, be resolved by supplementing the theory with local hidden variables (LHV), which nevertheless show an incompatibility with quantum predictions via violation of Bell’s inequality. Later, the violation of the so-called Clause-Horne-Shimony-Holt (CHSH) inequality, was verified experimentally3. As for the violation of Bell’s inequality, the incompatibility between the LHV models and quantum mechanics was essentially demonstrated in a statistical manner. If instead one aims to achieve a more sharper conflict, one can have the Greenberger-Horne-Zeilinger (GHZ) theorem, an “all-versus-nothing” proof of Bell’s nonlocality that applies to three or more parties45. The elegant GHZ argument involved the three-qubit GHZ state5 where |0〉 and |1〉 are the eigenstates of the Pauli matrix σz with the eigenvalues +1 and −1. respectively. It is easy to verify that the GHZ state is the common eigenstate of the following four mutually commutative operators: σ1xσ2xσ3x, σ1xσ2yσ3y, σ1yσ2xσ3y, and σ1yσ2yσ3x (here σ1x denotes the Pauli matrix σx measured on the 1st qubit, similarly for the others), with the eigenvalues being +1, −1, −1, −1, respectively. However, a contradiction arises if one tries to interpret the quantum result with LHV models. Specifically, we denote the supposedly definite values of σ1x, σ2y, … as v1x, v2y, … (with v’s being 1 or −1), then a product of the last three operators, according to LHV models, yields , in sharp contradiction to the first operator v1xv2xv3x = +1. Such a full contradiction “1 = −1” indicates that the GHZ theorem is a no-go theorem for quantum nonlocality, i.e., there is no room for the LHV model to completely describe quantum predictions of the GHZ state. The GHZ theorem has already been verified by photon-based experiment6, and recently a fault-tolerant test of the GHZ theorem has also been proposed based on nonabelian anyons7. In the original formulation of the EPR paradox1, a bipartite entangled state is considered which is a common eigenstate of the relative position and the total linear momentum and can be expressed as with the Planck constant. Experimentally one can generate the two-mode squeezed vacuum state in the nondegenerate optical parametric amplifier (NOPA)8 as where r > 0 is the squeezing parameter, , are respectively the annihilation and creation operators, |m〉 ≡ |Ψm(x)〉 are the Fock states of the Harmonic oscillator. In the infinite squeezing limit, , thus the original EPR state is a maximally entangled state for the bipartite continuous-variable system. Since the discovery of the EPR paradox, the question of whether the original EPR state possesses the LHV models has pushed many researchers to achieve intriguing and thought provoking results91011121314. Bell first showed that the Wigner function of the EPR state, due to its positive definiteness, can directly be used to construct the LHV models9. However, attempt has also been made to reveal its nonlocality in phase space by considering displaced parity operators upon the NOPA state in the large r limit10. Moreover, maximal violations of the EPR state by multicomponent Bell’s inequalities have also been investigated in refs 15,16. Very recently the notion of “steering”1718 has stimulated people to reconsider the exact implication of the EPR argument. For instance, Werner has remarked on why Einstein did not go all the way to discover Bell’s inequality19 Steering is indeed a quite old concept. In response to the EPR paper20, Schrödinger, who believed the validity of quantum mechanical descriptions of Nature, introduced in the same year of EPR’s paper a term “steering” to depict the “spooky action at a distance” which was mentioned in the EPR paper. Specifically, steering in a bipartite scenario describes an ability of one party, say Alice, to prepare the other party’s (say Bob’s) particle with different quantum states by simply measuring her own particle with different settings. This is also at the heart of remote state preparation protocol using EPR state21. However, steering lacked operational meanings, until in the year 2007 Wiseman et al.1718 gave a rigorous definition of it through the quantum information task. It then turns out that the EPR paradox concerns more precisely the existence of local hidden state (LHS) models, rather than that of LHV models leading to Bell’s inequality. That is, the exact type of quantum nonlocality in the EPR paradox is EPR steering, rather than Bell nonlocality. After that, there has been rapid development in EPR steering both theoretically and experimentally2223242526, such as in the test of steering inequalities27282930 and the experimental observation of one-way EPR steering31. Thus, a natural question arises: since there exist a simple GHZ paradox, i.e., “1 = −1”, which rule out the LHV models more uncompromisingly than Bell inequalities, one may ask whether a similar contradiction can be found so as to completely rule out the LHS models, especially for the EPR state. The merits of confirmatively answering this question include not only finding out the aforementioned missing piece of proofs of steering in analogy to proofs of Bell nonlocality, but also accomplishing the demonstration of the EPR paradox in its most original sense. The aim of this paper is to present a very simple steering paradox, i.e., “2 = 1”, which intuitively demonstrates the steerability for the EPR state, directly confirming that EPR steering is exactly the type of quantum nonlocality inherited in the EPR paradox, henceforth proving a no-go theorem for nonexistence of LHS models in EPR’s original sense. Results Simple steering paradox in two qubits We shall show that in the original EPR’s scenario, there exists a simple steering paradox that leads to “2 = 1”. A two-setting EPR steering scenario together with a bipartite entangled state are sufficient to demonstrate this full contradiction. To illustrate the central idea, let us first consider the two-qubit case. In a two-setting steering protocol of (with ), Alice prepares a two-qubit state ρAB, she keeps one and sends the other to Bob. Bob asks Alice to perform his choice of either one of two possible projective measurements (i.e. two-setting) and on her qubit and tell him the measurement results of a. Here is the projector, with the measurement direction, a (with a = 0, 1) the Alice’s measurement result, the 2 × 2 identity matrix, and the vector of the Pauli matrices. After Alice’s measurements, Bob obtains four conditional states as with j = 1, 2 and a = 0, 1. Suppose Bob’s state has a LHS description, then there exists an ensemble and a stochastic map satisfying where (with ) and are probabilities satisfying , and for a fixed ξ, and ρB = trA(ρAB) is Bob’s reduced density matrix (or Bob’s unconditioned state)1718. Then, Bob will check the following set of four equations: If these four equations have a contradiction (or say they cannot have a common solution of and ), then Bob is convinced that a LHS model does not exist and Alice can steer the state of his qubit. Now, let the state ρAB be an arbitrary two-qubit pure entangled state, which is given in its Schmidt form as where θ ∈ (0, π/2). The pure entangled state ρAB = |Ψ(θ)〉〈Ψ(θ)| has a remarkable property: Bob’s normalized conditional states are always pure, and for (Here are four different pure states when ρAB is a pure entangled state). It is well-known that a pure state cannot be obtained by a convex sum of other different states, namely, a density matrix of pure state can only be expanded by itself. Therefore without loss of generality, from Eq. (7) one has with the probabilities , and other terms are zeros (see Methods for more detail of derivation). By summing them up and taking trace, due to , the left-hand side gives 2trρB = 2. But the right-hand side, by definition, gives , this leads to a full contradiction of “2 = 1”. The above simple paradox “2 = 1” offers a transparent argument of nonexistence of LHS models (or existence of EPR steering) for a two-qubit pure entangled state. The subtlety of the paradox lies in the fact the wavefunction |Ψ(θ)〉 can have different decompositions, such as with and . In practice, the two-setting protocol can be chosen as . Namely, Bob asks Alice to measure her qubit along the -direction and the -direction, respectively. Suppose Alice performs her measurement in the -direction (or the -direction), for convenient, one may denote the set of her projectors as (or ), then she can project Bob’s system into one of the pure states {|0〉, |1〉} (or {|χ+〉, |χ−〉}). It is easy to verify that are locally orthogonal and complete bases. Namely, 〈0|1〉 = 〈+|−〉 = 0, , and the basis can be obtained from the diagonal basis through a unitary transformation. Generalization to bipartite high-dimensional systems Suppose in the steering scenario, the quantum state that Alice prepares is a pure entangled state of two d-dimensional systems (two-qudit), then one can have the same simple paradox “2 = 1”. Let us consider the two-qudit pure entangled state in its Schmidt form where |m〉 is the state in the diagonal basis, λm’s are the Schmidt coefficients, and . In the two-setting steering protocol of , Alice prepares a two-qudit pure state ρAB = |Φ〉〈Φ|, she keeps one and sends the other to Bob. To verify the steerablity of Alice, Bob asks Alice to perform his choice of either one of two possible projective measurements |m〉〈m| and |m′〉〈m′| on her qubit and tell him the measurement results of m and m′. Similarly, the sets of projectors for Alice are as follows In principle, the choice of and is rather arbitrary, as long as any element in does not fully overlap with that in . For simplicity and here can be taken as two of the mutually unbiased bases for a d-dimensional system, such that |〈m|m′〉|2 = 1/d for any pair of m and m′. After Alice’s measurements, Bob obtains 2d conditional states as and . Similarly, Bob can check the following set of 2d equations: with m, m′ = 0, 1, 2, …, d − 1. If these 2d equations have a contradiction, then there is no a LHS model description and Bob has to be convinced that Alice can steer the state of his qubit. Because ρAB = |Φ〉〈Φ| is a pure entangled state, it can be directly verified that Bob’s normalized conditional states are always pure, for instance one has . Due to the fact that a density matrix of pure state can only be expanded by itself, therefore, from equation (13) one has with m, m′ = 0, 1, 2, …, d − 1. By summing them up and taking the trace, we have From (15), one sees that the left-hand side gives 2trρB = 2 and the right-hand side gives trρB = 1, leading to a full contradiction of “2 = 1”. The above analysis is also valid when d tends to infinity. By chosing and let d → ∞, then one can have a similar paradox “2 = 1” for the continuous-variable state |NOPA〉, which includes the original EPR state by taking the infinite squeezing limit. Thus, we complete the demonstration of the simple steering paradox for the original EPR scenario, which is a no-go theorem for nonexistence of LHS models in the EPR paradox. In other words, the sharp contradiction “2 = 1” indicates that there is no room for the LHS description of any bipartite pure entangled state, including the original EPR state. Remark 1.—The original EPR state has the following elegant decompositions where in the last step we have operated a translation transformation on |Ψ〉EPR that does not change the state |Ψ〉EPR, is a real number, and . Thus the two-setting steering protocol can be chosen as . Discussions The EPR paradox has resulted in search for local hidden variable models with locality and reality as starting points, but Bell’s inequaliy rules out such mdels as the predictions of LHV models do not match quantum theory. The GHZ paradox demonstrates sharp contradiction between the predictions of local hidden variable theory and quantum mechanics without using any inequality. However, the GHZ paradox is not applicable to bipartite systems. Hardy did attempt to extend the all-versus-nothing argument to a two-qubit system to reveal Bell’s nonlocality3233, and this proof is usually considered as “the best version of Bell’s theorem”34. However, Hardy’s proof works for only 9% of the runs of a specially constructed experiment, and moreover, it is not valid for two-qubit maximally entangled state. Thus, in this sense, Hardy’s proof may not be considered appropriately as the closest form to the spirit of EPR’s original scenario. In summary, we have presented a simple steering paradox that shows the incompatibility of the local hidden state model with quantum theory for any bipartite pure entangled state, including the original EPR state. The full contradiction that results in “2 = 1”; not only intuitively demonstrates the steerability for the EPR state, directly confirming that EPR steering is exactly the type of quantum nonlocality inherited in the EPR paradox, but also indicates that the very simple steering paradox is the closest in its form to the spirit of the EPR paradox. Furthermore, if one considers the EPR steering scenario in k-setting, then following the similar derivation one can arrive at a full contradiction, i.e., “k = 1”. We expect that the simple steering paradox can be demonstrated in both two-qubit system and continuous-variable system by photon entangled based experiments in the near future. Methods Detail derivation of the steering paradox for two qubits It can be directly verified that, if the state ρAB = |Ψ(θ)〉〈Ψ(θ)| is a pure entangled state, then are four different pure states. For example and for convenient, let us take Then in the two-setting steering protocol of , Bob asks Alice to perform his choice of either one of two possible projective measurements along the z-direction (with the projector ) and the x-direction (with the projector ) on her qubit and tell him the measurement results of a (with a = 0, 1). More precisely, one has the projectors as with . Then Bob’s four unnormalized conditional states become with . Thus, Bob’s four normalized conditional states are which are obviously four different pure states. Now, if Bob’s four unnormalized conditional states can have a LHS description, then they must satisfy Since the four states in the left-hand-side of Eqs (21)–(24), , , are all proportional to pure states, thus it is sufficient for ξ to run from 1 to 4, namely, one can take the ensemble as with (if , it implies that the corresponding state ρξ′ is not the hidden state considered in the ensemble ), and ρi (i = 1, 2, 3, 4) are the hidden states. Then, Eqs (21)–(24), , , become In the following, we come to show a simple steering paradox “2 = 1” based on Eqs (26)–(29), , , under the constraints of Eq. (6), and It is well-known that a pure state cannot be obtained by a convex sum of other different states, namely, a density matrix of pure state can only be expanded by itself. Let us look at Eq. (26), because the left-hand side is proportional to a pure state, without loss of generality, one has Similarly, one has With the help of Eq. (31), one has This directly yields which is just the set of equations given in (9). It can be verified that For Eq. (37), by summing them up and taking trace, the left-hand side gives 2trρB = 2. But the right-hand side, by definition in Eq. (6), gives , this leads to the sharp contradiction “2 = 1,” as shown in the main text. Existence of LHS model for the pure separable state Consider now, however, a pure separable state of two qubits For this state, we shall show that a local hidden state model does exist. Without loss of generality, let Alice’s two choices of projective measurements be with By acting these projectors on the separable state (39), Bob’s four conditional states are found to be It then turns out that there exists a local hidden state model, with Alice’s strategy based on a single hidden state, that could simulate the above Bob’s four conditional states: Thus, local hidden state model is possible for pure separable states. Additional Information How to cite this article: Chen, J.-L. et al. Sharp Contradiction for Local-Hidden-State Model in Quantum Steering. Sci. Rep. 6, 32075; doi: 10.1038/srep32075 (2016). J.L.C. is supported by the National Basic Research Program (973 Program) of China under Grant No. 2012CB921900 and the National Natural Science Foundation of China (Grant Nos 11175089 and 11475089). H.Y.S. is supported by Institute for Information and Communications Technology Promotion (IITP) grant funded by the Korea Government (MSIP) (No. R0190-16-2028, Practical and Secure Quantum Key Distribution). A.K.P is supported by the Special Project of University of Ministry of Education of China and the Project of K. P. Chair Professor of Zhejiang University of China. Author Contributions J.-L.C. initiated the idea. J.-L.C., H.-Y.S., Z.-P.X. and A.K.P. derived the results and wrote the manuscript. All authors reviewed the manuscript. ==== Refs Einstein A. , Podolsky B. & Rosen N. Can quantum-mechanical description of physical reality be considered complete? Phys. Rev. 47 , 777 (1935 ). Bell J. S. On the Einstein-Podolsky-Rosen paradox . Physics (NY) 1 , 195 –200 (1964 ). Clauser J. F. , Horne M. A. , Shimony A. & Holt R. A. Proposed Experiment to Test Local Hidden-Variable Theories . Phys. Rev. Lett. 23 , 880 (1969 ). Greenberger D. M. , Horne M. A. & Zeilinger A. In Bell’s Theorem, Quantum Theory, and Conceptions of the Universe , edited by Kafatos M. , p. 69 (Kluwer, Dordrecht, 1989 ). Greenberger D. M. , Horne M. A. , Shimony A. & Zeilinger A. Bell’s theorem without inequalities . Am. J. Phys. 58 , 1131 (1990 ). Pan J. W. , Bouwmeester D. , Daniell M. , Weinfurter H. & Zeilinger A. Experimental test of quantum nonlocality in three-photon Greenberger-Horne-Zeilinger entanglement . Nature (London) 403 , 515 (2000 ).10676953 Deng D. L. , Wu C. , Chen J. L. & Oh C. H. Fault-Tolerant Greenberger-Horne-Zeilinger Paradox Based on Non-Abelian Anyons . Phys. Rev. Lett. 105 , 060402 (2010 ).20867962 Reid M. D. & Drummond P. D. Quantum Correlations of Phase in Nondegenerate Parametric Oscillation . Phys. Rev. Lett. 60 , 2731 (1988 ).10038437 Bell J. S. Speakable and Unspeakable in Quantum Mechanics . Chap. 21 (Cambridge University Press, Cambridge, England, 1987 ). Banaszek K. & Wódkiewicz K. Testing Quantum Nonlocality in Phase Space . Phys. Rev. Lett. 82 , 2009 (1999 ). Banaszek K. & Wódkiewicz K. Nonlocality of the Einstein-Podolsky-Rosen state in the Wigner representation . Phys. Rev. A 58 , 4345 (1998 ). Ou Z. Y. , Pereira S. F. , Kimble H. J. & Peng K. C. Realization of the Einstein-Podolsky-Rosen Paradox for Continuous Variables . Phys. Rev. Lett. 68 , 3663 (1992 ).10045765 Cetto A. M. , De La Pena L. & Santos E. A Bell inequality involving position, momentum and energy . Phys. Lett. A 113 , 304 (1985 ). Cohen O. Nonlocality of the original Einstein-Podolsky-Rosen state . Phys. Rev. A 56 , 3484 (1997 ). Chen Z. B. , Pan J. W. , Hou G. & Zhang Y. D. Maximal Violation of Bell’s Inequalities for Continuous Variable Systems . Phys. Rev. Lett. 88 , 040406 (2002 ).11801099 Chen J. L. et al. Multicomponent Bell inequality and its violation for continuous-variable systems . Phys. Rev. A 71 , 032107 (2005 ). Wiseman H. M. , Jones S. J. & Doherty A. C. Steering, Entanglement, Nonlocality, and the Einstein-Podolsky-Rosen Paradox . Phys. Rev. Lett. 98 , 140402 (2007 ).17501251 Jones S. J. , Wiseman H. M. & Doherty A. C. Entanglement, Einstein-Podolsky-Rosen correlations, Bell nonlocality, and steering . Phys. Rev. A 76 , 052116 (2007 ). Werner R. F. Steering, or maybe why Einstein did not go all the way to Bell’s argument . J. Phys. A: Math. Theor. 47 , 424008 (2014 ). Schrödinger E. Discussion of probability relations between separated systems . Naturwiss . 23 , 807 (1935 ). Pati A. K. Minimum cbits for remote preperation and measurement of a qubit . Phys.Rev. A 63 014320 (2000 ). Oppenheim J. & Wehner S. The Uncertainty Principle Determines the Nonlocality of Quantum Mechanics . Science 330 , 1072 (2010 ).21097930 Branciard C. , Cavalcanti E. G. , Walborn S. P. , Scarani V. & Wiseman H. M. One-sided device-independent quantum key distribution: Security, feasibility, and the connection with steering . Phys. Rev. A 85 , 010301 (R) (2012 ). Chen J. L. et al. All-Versus-Nothing Proof of Einstein-Podolsky-Rosen Steering . Sci. Rep. 3 , 2143 (2013 ).23828242 Chen J. L. et al. Beyond Gisin’s Theorem and its Applications: Violation of Local Realism by Two-Party Einstein-Podolsky-Rosen Steering . Scientific Reports 5 , 11624 (2015 ).26108704 Sun K. et al. Experimental Demonstration of the Einstein-Podolsky-Rosen Steering Game Based on the All-Versus-Nothing Proof . Phys. Rev. Lett. 113 , 140402 (2014 ).25325621 Saunders D. J. , Jones S. J. , Wiseman H. M. & Pryde G. J. Experimental EPR-steering using Bell-local states . Nature Phys . 6 , 845 (2010 ). Smith D. H. et al. Conclusive quantum steering with superconducting transition edge sensors . Nature Comm . 3 , 625 (2012 ). Bennet A. J. et al. Arbitrarily loss-tolerant Einstein-Podolsky-Rosen steering allowing a demonstration over 1 km of optical fiber with no detection loophole . Phys. Rev. X 2 , 031003 (2012 ). Wittmann B. et al. Loophole-free quantum steering . New J. Phys. 14 , 053030 (2012 ). Händchen V. et al. Observation of one-way Einstein-Podolsky-Rosen steering . Nature Photonics 6 , 596 (2012 ). Hardy L. Nonlocality for Two Particles without Inequalities for Almost All Entangled States . Phys. Rev. Lett. 71 , 1665 (1993 ).10054467 Rabelo R. , Zhi L. Y. & Scarani V. Device-Independent Bounds for Hardy’s Experiment . Phys. Rev. Lett. 109 , 180401 (2012 ).23215258 Mermin N. D. The best version of Bell’s theorem . Ann. (NY) Acad. Sci. 755 , 616 –623 (1995 ).

PMC005xxxxxx/PMC5000012.txt

==== Front Sci RepSci RepScientific Reports2045-2322Nature Publishing Group srep3201610.1038/srep32016ArticleA novel mass spectrometric strategy “BEMAP” reveals Extensive O-linked protein glycosylation in Enterotoxigenic Escherichia coli Boysen Anders 12Palmisano Giuseppe a13Krogh Thøger Jensen 1Duggin Iain G. 2Larsen Martin R. 1Møller-Jensen Jakob b11 Department of Biochemistry and Molecular Biology, University of Southern Denmark, Campusvej 55, 5230 Odense M, Denmark2 The ithree institute, University of Technology Sydney, PO Box 123, Broadway NSW 2007, Australia3 GlycoProteomics laboratory, Department of Parasitology, University of Sao Paulo, Brazila palmisano.gp@usp.brb jakobm@bmb.sdu.dk26 08 2016 2016 6 3201601 06 2016 01 08 2016 Copyright © 2016, The Author(s)2016The Author(s)This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/The attachment of sugars to proteins via side-chain oxygen atoms (O-linked glycosylation) is seen in all three domains of life. However, a lack of widely-applicable analytical tools has restricted the study of this process, particularly in bacteria. In E. coli, only four O-linked glycoproteins have previously been characterized. Here we present a glycoproteomics technique, termed BEMAP, which is based on the beta-elimination of O-linked glycans followed by Michael-addition of a phosphonic acid derivative, and subsequent titanium dioxide enrichment. This strategy allows site-specific mass-spectrometric identification of proteins with O-linked glycan modifications in a complex biological sample. Using BEMAP we identified cell surface-associated and membrane vesicle glycoproteins from Enterotoxigenic E. coli (ETEC) and non-pathogenic E. coli K-12. We identified 618 glycosylated Serine and Threonine residues mapping to 140 proteins in ETEC, including several known virulence factors, and 34 in E. coli K-12. The two strains had 32 glycoproteins in common. Remarkably, the majority of the ETEC glycoproteins were conserved in both strains but nevertheless were only glycosylated in the pathogen. Therefore, bacterial O-linked glycosylation is much more extensive than previously thought, and is especially important to the pathogen. ==== Body Enterotoxigenic Escherichia coli (ETEC) is the major source of E. coli mediated diarrhoea in humans and livestock1. ETEC infections cause more than 280 million annual episodes of diarrhoea resulting in mortality numbers exceeding 550,000 deaths of children under the age of five years2. The significant negative health- and socio-economic impact of ETEC infection manifests itself mainly in the undeveloped nations with poor sanitation and inadequate supplies of clean water3. ETEC strains are a diverse group of pathogens defined by their ability to colonize the small intestine and secrete heat-labile and/or heat stable enterotoxins4. Their pathogenicity is further attributed to the presence of virulence genes on mobile genetic elements, including a number of plasmids and chromosomal pathogenicity islands5. Much attention has been devoted to the understanding of how ETEC and other mucosa-associated pathogens interact with host tissue during infection6. A number of studies have revealed that bacterial protein glycosylation plays an important role in mediating adhesion, colonization and invasion of host tissue789101112. However, up until now, the known protein glycosylation repertoire of E. coli was limited to just four proteins, all of which are surface-exposed adhesins with functions in bacterial pathogenesis13141516. While the intimate coupling between protein glycosylation and bacterial pathophysiology has become apparent, the discovery of novel glycoproteins implicated in virulence is advancing slowly61718. This is attributed to the inherent challenges associated with glycoproteomics. The analytical tools developed for enrichment of eukaryotic O- and N-linked glycopeptides rely on a limited set of defined physiochemical properties, e.g. glycan hydrophilicity or specific lectin recognition, which are relatively rare in bacteria1920. Discovery and characterization of glycoproteins is further complicated by heterogeneous glycosylation, low abundance, poor ionization of peptides modified with carbohydrates compared to the non-modified counterpart21 and lack of specific enzymes to remove the heterogeneous glycan structure prior to mass spectrometric (MS) analysis22. Mapping of O-linked glycan moieties has also proven to be challenging owing to the diverse nature of carbohydrate structures available for protein modification in bacteria23. Although methods such as periodic acid/hydrazide glycan labelling and metabolic oligosaccharide engineering (MOE) have identified glycoproteins in a range of bacteria, these techniques present limitations in the form of low specificity for glycosylated proteins and dependence on sugar uptake and integration into bacterial glycoproteins, respectively1724. Recently, the diversity within the O-linked protein glycosylation systems of Acinetobacter species were described using Hydrophilic Interaction LIquid Chromatography (HILIC) glycopeptide enrichment in combination with ETD HCD and CID fragmentation in the MS instrument25. Here we describe a novel mass spectrometry-based technique, termed BEMAP, which can be employed to map O-linked glycoproteins from theoretically any biological source. BEMAP is an extension of a method, which was originally described for phosphorylated peptides enrichment analysis of phosphorylated proteins as a tool for probing the phosphoproteome26 and later for O-GlcNAcylated peptides27. In this strategy the phosphate or O-GlcNAc moieties are replaced by an affinity tag, commonly a biotin group, using base-catalyzed β-elimination followed by Michael-addition of the affinity tag. The BEMAP reaction efficiently substitutes O-linked carbohydrate moieties with a 2-Aminoethyl phosphonic acid (AEP) group, which can be selectively isolated based on its affinity for titanium dioxide, as previously shown for phosphorylated peptides28. Here we employed the BEMAP strategy to map 618 novel protein O-glycosylation sites in the ETEC strain H10407 and the non-pathogenic E. coli K-12. The far majority of the sites were identified in the pathogenic E. coli strain. These results highlight protein O-glycosylation in bacteria as an abundant, yet largely unexplored, post-translational protein modification associated with cellular core processes and pathophysiology. Bacterial O-glycoproteins potentially constitute an important reservoir of novel therapeutic targets, biomarkers and vaccine candidates. Results and Discussion BEMAP enables selective and efficient enrichment of O-linked glycopeptides The intimate coupling between bacterial protein glycosylation and pathophysiology found in several species6 prompted us to identify novel glycosylated proteins in ETEC strain H10407. ETEC modifies proteins with heptose and potentially N-acetylglucosamine (GlcNAc) monosaccharides1516. Due to the lack of suitable methods to investigate the O-linked glycoproteome in bacteria, we developed a mass spectrometry-based technique for unbiased identification of O-linked protein glycosylation sites on a proteome scale. Selective enrichment of phosphopeptides and O-GlcNAc modified glycopeptides was previously demonstrated by use of phosphate tag, which has a high affinity for TiO2282930. Our method, termed BEMAP, relies on β-elimination of O-linked carbohydrate modifications, Michael addition of 2-Aminoethyl phosphonic acid (AEP), and then subsequent titanum dioxide (TiO2) enrichment of the resulting phosphonate peptides. Thus, BEMAP combines an established in vitro chemical modification with a highly selective enrichment protocol31. The reactions take place in a single sample without the need for intermediate purification steps (see Experimental procedures section). The BEMAP method was first established using a synthetic O-linked mannosylated peptide as a model compound. As shown in Fig. 1A,B, MALDI MS demonstrated that BEMAP efficiently replaced the carbohydrate moiety of the synthetic peptide (m/z = 1181.59 Da) with the AEP group and thus introduced a phosphonate peptide (m/z = 1126.64 Da). The overall efficiency of substitution exceeded 95% (Fig. 1B) without the formation of any visual degradation products. The AEP-modified peptide was then efficiently enriched using TiO2 affinity chromatography; both the intact glycopeptide and the β-eliminated peptide (1001.62 Da) were absent in the MALDI MS spectrum after enrichment (Fig. 1C). We found that BEMAP converted O-linked glycopeptides into phosphonate peptides in a time dependent fashion and that maximum conversion had occurred after 195 minutes (Supplementary Fig. S1A–E). We analysed the gas-phase fragmentation properties of the converted O-linked glycopeptide. As shown in Supplementary Fig. S2, the exchange of a carbohydrate moiety with AEP has several advantages. The AEP addition substitutes a labile glycoside bond with a stronger covalent C-N bond, which greatly improved mapping of O-glycosylated amino acid residues by high-energy C-trap dissociation (HCD) fragmentation (Supplementary Fig. S2). Moreover, the AEP group yielded two characteristic reporter ions during HCD fragmentation (m/z = 126.03 Da and m/z = 138.03 Da), which are very useful for glycopeptide identification and validation in complex MS/MS spectra. It should be noted that the AEP molecule has a phosphonate functional group, which is stable under CID and HCD fragmentation conditions in contrast to a phosphate group, which is labile under these conditions. This allows unambiguous assignment of the modified amino acid residues and avoids false positives in site localization assignment. In addition, HCD fragmentation yielded higher peptide sequence coverage and allows one to monitor the two reporter ions (Supplementary Fig. S2). Due to that, HCD fragmentation was used in other experiments. Of importance, even though the synthetic glycopeptide contains three threonine residues, only the O-linked serine amino acid was converted into a phosphonate peptide by the BEMAP chemistry. This implies a specific chemical reactivity towards O-linked glycans compared to unmodified serine/threonine residues. Next, we applied BEMAP to a purified heptosylated protein: Ag43 from E. coli32. As seen in Fig. 1D, in-gel trypsin digestion of the glycosylated protein yielded heptosylated and unmodified peptides. Heptosylated peptides marked by an asterisk are suppressed by the non-modified peptides. However, the BEMAP strategy enriched the three heptosylated peptides present in Fig. 1D as well as four additional glycopeptides initially undetectable by MALDI MS (Fig. 1E,F). It is concluded that BEMAP is a specific and sensitive method for identification of formerly O-linked glycosylated peptides. Enzymatic dephosphorylation treatment required for large-scale glycoproteome analyses Both pathogenic and commensal E. coli phosphorylates a subset of its proteome33343536. As the BEMAP chemistry relies on β-elimination and Michael addition, which previously has been shown to react with serine and threonine phosphorylation26 we therefore considered phosphoprotein as a potentially confounding factor in identification of the genuine O-glycoproteome. Moreover, the TiO2 enrichment step included in the workflow could potentially affinity-purify phosphopeptides not converted into 2-AEP peptides and suppress the ionization of 2-AEP modified O-glycopeptides. In order to evaluate the extent of such false identification, we first assessed how efficiently BEMAP converts phosphorylated serine and threonine residues into 2-AEP modified peptides. Using the experimental conditions described above, we studied the multiply phosphorylated protein Fetuin using MALDI MS. TiO2 enriched phosphopeptides from in-solution digested Fetuin were isolated and then analysed by MALDI MS (Supplementary Fig. S3A)28. Assuming similar ionization efficiency of the phospho- and AEP-modified peptides by MALDI, we observe that treating the phosphopeptides with the BEMAP chemistry, results in less than 38% conversion of the mono-phosphopeptide to the corresponding phosphonate peptide. The di-phosphorylated peptide is converted to an even lesser extent. Of peptides containing two possible substitution sites, only 36% were converted on a single site. 83% of the tri-phosphorylated peptide were substituted with a single AEP modification; double- and triple substitutions were not observed (Supplementary Fig. S3B). The BEMAP conversion rate is similar to the one observed by Wells et al.27 but indicates that the chemistry could exchange serine and threonine phosphorylations with the 2-AEP tag. We then performed an experiment in order to estimate the number of false positive identifications which could be expected in an outer membrane O-glycoproteome BEMAP analysis. In three biological experiments, proteins associated with the outer membrane were sequentially isolated37 and digested with trypsin. The peptides were treated either with or without alkaline phosphatase (AP) prior to phosphopeptide enrichment and LC-MS/MS analysis28. When analysing the samples not treated with AP prior to phosphopeptide enrichment and mass spectrometry, we identified 227 phosphorylated S/T/Y residues which could be assigned to 99 proteins (Supplementary Fig. S4A,B and Supplementary Table S1). In the parallel experiment, which included an AP sample treatment, we mapped 51 phosphorylation sites to 31 proteins. When comparing the two datasets, 23 phosphorylation sites were shared between 13 proteins, see Supplementary Fig. S4A,B. These results show that at the protein level, AP sample treatment significantly reduces the number of identified phosphoproteins as expected. Taking into account, that only proteins associated with the outer membrane were sampled, we have identified a high number of phosphopeptides when comparing to the published in-depth phosphoproteome studies carried out in commensal MG1655 E. coli333536. In the MG1655 analyses, the serine and threonine phosphorylation represented approximately 70% and 20% of all modifications, respectively. Less than 10% were phosphorylated tyrosine residues. We observe a similar ratio between pS, pT and pY. Although tyrosine phosphorylation is relatively rare, a phosphotyrosine immunoaffinity based pulldown experiment identified 512 modified residues in both commensal and pathogenic enterohemorrhagic E. coli (EHEC)34. In our analysis, S/T phosphopeptides are potentially a source for overestimating the O-glycoproteome. However, a low BEMAP chemistry conversion rate of phosphorylated S/T residues in combination with AP treatment prior to a BEMAP O-glycoproteome analysis removes the majority of the false positive identifications. Characterization of the E. coli outer membrane O-glycoproteome The four known E. coli glycoproteins AIDA-1, Antigen 43 (Ag43), TibA and EtpA are all associated with the outer membrane, and we reasoned that additional glycoproteins involved in interacting with host cells could also be associated with the surface. We therefore sought to identify additional O-glycoproteins from the outer membrane of the ETEC strain H10407 using BEMAP. Interestingly, the closely related commensal E. coli K-12 strain MG1655 carries the potential for protein glycosylation at the genetic level14 and were thus included in the analysis as a non-pathogenic reference. If protein O-glycosylation could be identified in both E. coli strains, a direct comparison of the two related organisms sampled under identical conditions would potentially highlight important differences and similarities useful for predicting novel O-glycosylated therapeutic targets in ETEC. The outer membrane protein fractions of H10407 and MG1655 were isolated using exactly the same outer membrane- and TiO2 enrichment protocol as for phosphopeptide purification and subjected to BEMAP analysis for site specific identification of O-linked glycosylation sites. This identified a total of 547 glycosylated residues (Fig. 2A) which could be assigned to 127 proteins (Fig. 2B). A total of 125 glycosylated proteins were identified in the ETEC outer membrane, whereas 34 glycoproteins were found in the commensal strain (Fig. 2B). When using a sequence alignment with a 60% identity and 80% similarity cut-off threshold at the protein level, 32 proteins were shared between the two strains; leaving just 2 proteins only found modified by the non-pathogen, see Supplementary Table S2. Out of the 93 glycoproteins specifically identified in the ETEC sample only seven proteins were uniquely expressed in the pathogen (Fig. 2B), implying that ETEC O-glycosylates its protein to a much higher extent than to the non-pathogen, just as tyrosine phosphorylation has been proposed to be in EHEC34. This indicates that post-translational modifications (PTMs) could play a role in the pathophysiology of E. coli. Interestingly, some of these proteins, e.g. adhesin TibA autotransporter, colonization factor CfaB, pesticin/yersiniabactin TonB-dependent receptor and Yersiniabactin siderophore biosynthetic protein are associated with various aspects of pathogenesis such as host cell adhesion and iron acquisition (bold highlighted proteins in Supplementary Table S2)438. The adhesin TibA autotransporter has previously been described to be associated with the outer membrane39. The identification of TibA in our analysis validates the BEMAP enrichment (Supplementary Fig. S5A,B). Knowing that a sub-fraction of the identified O-glycopeptides could potentially be phosphopeptides (see Supplementary Table S1 and Table S2) we compared the list of O-linked glycopeptides to the list of phosphorylated serine- and threonine-phosphopeptides previously identified in the sample (Supplementary Table S3). As shown in Fig. 3, omitting AP treatment prior to phosphopeptide enrichment and LC-MS/MS analysis revealed that only 24 peptides of the 546 identified O-glycopeptides were shared between the two data sets. These 24 peptides could be assigned to nine proteins (Supplementary Table S3). Only 5 phosphorylated sites overlapped between the BEMAP enriched- and the phosphopeptide enriched samples following AP treatment (Fig. 3). We propose that these five particular sites are AP treatment resistant. We conclude that AP treatment can reduce the number of false positive identifications and that our BEMAP workflow identifies O-glycosylated peptides with high selectivity. Of note, less than 5% of the modified sites (29 out of 715 sites; Fig. 3) are overlapping, suggesting that protein phosphorylation and O-glycosylation occurs with different specificity. Furthermore, glycosylation appears to be a more abundant than phosphorylation, at least in the outer membrane fraction. In order to determine if protein O-glycosylation could be associated with specific cellular functions, the outer membrane associated glycoproteins were clustered into four groups according to the number of O-linked glycosylation sites. We additionally categorized the glycoproteins according to their gene ontology (GO) annotation40. As shown in Fig. 4, the clustering revealed that 101 of the 127 proteins had a limited number of O-glycosylated residues (1–4 sites).Twenty-six glycoproteins were more extensively modified, showing more than four distinct O-glycosylation sites each. Roughly 20% of the proteins in all four groups were of unknown function. Approximately half of all the proteins that carried up to four glycosylation sites were predicted to be metabolic enzymes whereas more than 70% of all proteins with more than 10 glycosylation sites were categorized as outer membrane transporters. Nine of theses transporters, CirA, FepA, FhuA, OmpC, OmpF, OmpA, Peptidoglycan-associated lipoprotein, OmpT and TolC are involved in acquisition of iron, hydrophilic solutes, ions as well as peptides. Since the majority of these have previously been crystallized (see Supplementary Table S2) we next investigated if the O-glycosylation sites were randomly distributed within the protein structure or if they clustered into particular spatial regions, which could be of biological importance. Therefore, we visualized the structural position of the O-glycosylation sites by highlighting the sugar-modified residues within crystal structures (Supplementary Fig. S6A-5I). Generally, the O-glycosylation sites were located in either unstructured regions on the protein exterior or positioned within the barrel pore. The spatial arrangement of the glycosylated residues suggest that the hydrophilic environment created by the monosaccharide is an integral function of the E. coli outer membrane transporters. It is striking that the carbohydrate modifications have gone by unnoticed in all published crystal structure we examined. We speculate that crystals used for X-ray diffraction require large amounts protein, which is usually obtained by overexpression in a non-native host, which may not necessarily express the cognate glycosyltranferases required for glycosylation or other PTMs. It is also possible that ectopic protein expression might deplete the host cell for nucleotide-activated sugar that is funneled into the glycosylation pathway thus rendering the protein un-modified. Identification of glycoproteins in ETEC membrane vesicles Both pathogenic and commensal E. coli produce membrane vesicles (MVs) containing proteins41. However, when comparing the relative protein composition of MVs released by ETEC and MG1655, it has been shown that vesicles produced by the pathogen functions as a vessel for export of LT toxin and bacterial virulence factors targeting the host mucosal layer4243. We isolated ETEC MVs and performed a BEMAP analysis in order to examine whether any of these proteins could be glycosylated. Using BEMAP, we identified 133 glycosylated residues which could be assigned to 22 proteins (Supplementary Table S4). Out of the 22 proteins, the known glycoproteins EtpA and TibA as well as Flagellin, EatA, YghJ, CfaB and CexE are characterized as ETEC virulence factors. The MVs also contained a number of proteins with putative unknown functions. The exact role of glycoproteins in pathogenesis remains to be determined but approximately 1/6 of all proteins identified in ETEC vesicles appear to be glycosylated43. We observed that a number of different glycopeptides identified in both the outer membrane fraction and in MVs could be assigned to the same protein. Thus, we combined the two glycopeptide lists into a non-redundant dataset showing the full extent of protein O-glycosylation of proteins associated with the outer membrane in ETEC and commensal MG1655 (Supplementary Table S5). Quite remarkably, the combined list revealed that 84 H10407 FliC Ser/Thr residues out of 100 possible were modified. This extent of modification surpasses any reported number of O-linked glycosylation sites on a single protein in bacteria644. FliC (flagellin), the major structural component of the flagellum, is conserved in the two related E. coli strains. However, the number of glycosylated FliC residues in the two organisms is strikingly different. FliC is only glycosylated in ETEC (see Supplementary Table S2). Previous studies have shown that the flagellum is required for efficient adherence to the intestinal epithelium45 and our findings group ETEC with the mucosal-associated pathogens P. aeruginosa, H. pylori and C. jejuni, which extensively glycosylate their flagella. In these species, flagellar glycosylation is absolutely essential for the biogenesis of functional flagella and hence for virulence11454647). To verify the modified sites identified by BEMAP, we used an orthogonal experimental setup similar to the methods applied to obtain the current knowledge about Flagellin glycosylation46474849. In this direct approach, an isolated Flagellin protein fraction was separated by SDS PAGE. The FliC protein was in-gel digested using trypsin and the resulting peptides were HILIC fractionated before analysis by nanoLC-MS/MS50. The acquired data were searched with either heptose or GlcNAc as variable modification. This experimental workflow identified 14 heptosylated Ser and Thr residues (Fig. 5A and Supplementary Table S6). A total of 12 of these sites were also found in the BEMAP analysis. The relatively low number of observed sites when applying an orthogonal experimental approach highlights the advantage of the BEMAP strategy. Our data also indicate that heptosylation is one of the most abundant sugar modifications in ETEC as no GlcNAc sites were identified on the most extensively modified protein FliC. To further characterize the FliC O-glycosylation, we mapped the modified FliC residues onto the primary sequence (see Fig. 5B). The majority of the O-glycosylation sites could be assigned to the N- and C-terminal domains, which are conserved amongst bacteria51. Interestingly, a sequence alignment of MG1655 and H10407 flagellin showed that the majority of the modified residues in the pathogen were conserved, but unmodified in the non-pathogen (see Fig. 5B). We mapped the O-glycosylation sites onto the Salmonella FliC crystal structure (Fig. 5C). This analysis illustrated that the sugar modifications were positioned on the interior face of the flagellum. The ETEC modification distribution is different from that of C. jejuni and H. pylori FliC, which are previously proposed to be glycosylated exclusively in the variable surface exposed domain4647. In light of the close evolutionary relationship between the strains, a BEMAP analysis could be used to explore whether or not O-glycosylation is conserved feature in these pathogens. The glycosylated adhesin EtpA is mounted on the tip of the flagellum and interacts with the conserved domains of FliC45. It is plausible that the EtpA-FliC protein interactions could be augmented by glycosylation. Considering that a single flagellum may consist of as many as 30000 flagellin subunits the very large number of modifications make this extracellular appendage an substantial metabolic investment even if only a fraction of the sites are O-glycosylated52. In addition, the combined dataset showed that ETEC Colonization Factor Antigen 1 (CFA/I) is O-glycosylated at two residues. Such modifications have been found to play an important role for the pathogenesis of other mucosa-associated pathogens5354. Our analysis shows that the amino acid residues T74 and T78 are glycosylated. When mapping the glycosylated sites onto the CfaB crystal structure the glycans were found in the interface between CfaB subunits in the asymmetric unit (Supplementary Fig. S7). It is conceivable that the hydrophilic environment created by the glycans could facilitate fimbrial assembly. Searching for an ETEC O-linked protein glycosylation sequence motifs The scarcity of identified E. coli O-linked glycosylation sites has led to the assumption that residue selectivity may rely on recognition through a structural arrangement spanning 19 amino acids rather than a linear sequence motif3255. The obtained outer membrane glycosylation dataset allowed us to search for a sequence motif using bioinformatics prediction. Thus, we explored the nature of O-linked glycosylation in E.coli using the MotifX algorithm56. We searched for overrepresented motifs in a nine amino acid sequence window surrounding 389 and 239 unique Ser and Thr sites, respectively. An ETEC H10407 database was used as background and only motifs with high significance (p < 10−4) were considered. Seven serine and four threonine glycosylation motifs were significantly enriched in our data set comprising 37% of all identified sites (Fig. 6). In four out of five serine sequence outputs, glycosylation correlated with preference for asparagine in position −7, +1, +2 and +6 relative to the central character whereas the fifth motif contained a threonine in position +1. The threonine motifs were identical to the serine motifs. The glycosylation motif analysis extracted an overrepresentation of threonine in position +1 and asparagine in +2. Protein glycosylation in ETEC is proposed to be catalysed by the predicted N-acetylglucosamine (GlcNAc) transferase EtpC and the heptosyltransferase TibC which attaches heptoses to its target1639. The unambiguous enrichment of asparagine flanking both the serine and threonine suggests that this amino acid is functionally important for glycosylation. It remains to be determined which glycosyltransferase recognises the individual motifs. Conclusion We set out to devise a new method for accelerating the discovery of novel O-linked glycosylated proteins in E. coli. Our method, termed BEMAP, represents a refinement of a previous β-elimination/Michael addition experimental strategy originally described for phosphorylated peptides and O-GlcNAcylated peptides26. The novelty of BEMAP lies in the use of 2-Aminoethyl phosphonic acid (AEP) for nucleophilic peptide tagging. The selectivity of BEMAP is achieved by the glycan-to-phosphonate molecule exchange combined with a highly specific enrichment protocol for enrichment of the phosphonate peptide28. Importantly, the BEMAP chemistry can be applied in principle to any organism on a large-scale proteomics level irrespective of the chemical properties and nature of the O-linked sugar moiety. As demonstrated in Fig. 1, BEMAP replaces the carbohydrate moiety of a synthetic glycosylated peptide with a phosphonate tag in a chemical reaction exceeding 95% efficiency. We have shown that phosphorylated serine and threonine residues can be converted by the BEMAP chemistry, as previously published by Oda et al.26, and hence represent a source of false identification, see Fig. S2A,B. This can however be mitigated by including a simple Alkaline Phosphatase treatment of the sample prior to the BEMAP workflow (Fig. 3) or by including proper controls of enriched phosphopeptides from the sample. Another limitation of BEMAP relates to establishing the identity of the eliminated sugar modification. Building on prior knowledge, ETEC modifies proteins with heptose and potentially GlcNAc monosaccharides1516. However, when working with organisms for which the nature of O-glycosylation is limited or unknown, different approaches can be considered. In one workflow, monosaccharides from a purified protein can be identified using a direct chemical analysis1557. Alternatively, the O-glycans can be released from individual proteins or whole cells, isolated and finally identified using advanced nanoLC-ESI-MS/MS58. We highlight that BEMAP is compatible with experimental strategies which investigate the relative abundance of proteins such as stable isotope labeling with amino acids in cell culture (SILAC) and isobaric tags for relative and absolute quantitation (iTRAQ)5960. We propose that BEMAP can improve our understanding of the intricacies of bacterial protein glycosylation amassed during pathogenesis which should lead to new opportunities to manipulate these pathways. In this regard, it should be noted that BEMAP constitutes a powerful qualitative glycoprotein discovery tool; no information about the modification frequency of a given amino acid residue is provided. It is highly conceivable that bacterial pathogens may exploit variable multisite glycosylation patterns to scramble their surface structure in order to evade recognition by the immune system. Indeed, a recent study61 demonstrated extensive antigenic variation in Neisseria meningitidis type IV pili. We propose the use of top-down mass spectrometric analysis subsequent to BEMAP glycosite discovery in order to determine the frequency of modification at specific sites of selected glycoproteins. To identify specific pathogenic E. coli associated glycoproteins we compared the outer membrane protein complement to non-pathogenic reference strain MG1655 sampled under identical conditions. By applying our BEMAP workflow we identified a total of 618 glycosylated residues which could be assigned to 149 proteins, see Supplementary Table S5. By categorizing the identified ETEC outer membrane associated glycoproteins according to their gene ontology (GO) annotation40, our data indicate that protein glycosylation in E. coli plays a role in ETEC virulence as well as normal cellular physiology, see Figs 4 and 5 and Supplementary Fig. S6. Our findings parallel recently published data revealing an abundance of post translational modifications (PTMs) in E. coli including tyrosine phosphorylation, lysine succinylation and lysine acetylation as well as glycosylation in e.g. H. pylori and A. baumannii242534626364. Using the motifX algorithm, our findings reinforce the notion that O-glycosylation residue selection in ETEC may rely on a sequence motif (Fig. 6). Previous studies have demonstrated that the E. coli heptosyltransferases are rather promiscuous in target selection1314 and our data cannot discriminate whether EtpC and TibC compete for monosaccharide accepting Ser or Thr residues and whether they are capable of accepting both heptose and GlcNAc as substrate. The presented work has exposed a significant amount of previously unexplored glycoproteins in E. coli that warrant deeper characterization. Our data supersedes any reported number of glycoproteins originating from a single microorganism23. Experimental Procedures Strains Pathogenic E. coli ETEC H10407 and E. coli K-12 MG1655 were used in this study. Bacterial strains are listed in Table 1. General methods Cells were grown in M9 minimal medium containing 0.4% glucose65. Cultures were inoculated at an OD600 of 0.01 from an overnight (O/N) culture. Purification of outer membrane associated proteins Outer membrane proteins were isolated from cells grown to OD600 = 0.6 in 1.5 L M9 minimal medium supplemented with 0.4% glucose. Bacteria were harvested at 4500 × g for 30 min at 4 °C. The cell pellet was resuspended in 15 ml lysis buffer (50 mM sodium phosphate pH 7.0, 150 mM NaCl supplemented with EDTA-free protease inhibitor tablet from Roche) and lysed three times in a French press at 10,000 psi. Outer membrane associated proteins were obtained essentially as described by37. Outer membrane associated proteins were precipitated using the improved Wessel/Flügge method for large sample volumes66. Briefly described, the 20 ml outer membrane extract was vortexed together with 15 ml ice cold Methanol. 5 ml ice cold Chloroform was added and the sample was vortexed once again before centrifuged at 10000 × g for 45 min at 4 °C. The aqueous phase was carefully removed and 16.66 ml ice cold Methanol was added to the remaining protein sheet floating on top of the organic phase. The sample was vortexed and the protein precipitated by centrifugation at 10000 × g for 45 min at 4 °C. Organic solvent was removed and the protein pellet was allowed to air dry. Purification of membrane vesicles Membrane vesicles were isolated from a cell culture grown to OD600 = 0.6 in 1.5 L M9 minimal medium supplemented with 0.4% glucose. A cell free culture supernatant was obtained by pelleting bacteria at 4500 × g for 30 min at 4 °C and sterile filtration (Millipore 0.22 μm filter cup). Supernatant was concentrated 21 fold using a stirred ultrafiltration cell model 8400 (Millipore) fitted with a PLCC ultrafiltration disc with a 5 kDa NMWL cut-off (Millipore). Vesicles were obtained from concentrated culture supernatant by centrifugation at 125000 × g for 3 hours at 4 °C. Proteins were isolated from the collected vesicles using the improved Wessel/Flügge method described above, by keeping the ratios between aqueous and organic solvents fixed. Outer membrane and MV in solution reduction, alkylation and proteolytic digestion Outer membrane as well as MV associated protein samples were solubilized in 100 μl buffer containing 6M Urea, 2M thiourea, 100mM TEA bicarbonate pH 8.0 before reduced in 10 mm DTT for 1 hour at 25 °C and alkylated in 50 mM iodoacetamide for 40 minutes at 25 °C in the dark. Each sample was diluted 1:10 with 50 mm TEAB pH 8.0 and digested with 2% (w/w) trypsin 16 h at 25 °C. The supernatant was desalted using Oasis HLB Plus short cartridges (Waters) as recommended by manufacturer and finally dried by vacuum centrifugation and stored at −20 °C. Fetuin Alkaline Phosphatase treatment analysis 1nmol Fetuin was reduced in 10 mM DTT for 1 hour at 25 °C and alkylated in 50 mm iodoacetamide for 40 minutes at 25 °C in the dark in total reaction volume of 50 μl containing 10 mM TEAB pH 8.0. Fetuin sample was digested with 2% (w/w) trypsin 16 h at 25 °C before micro-tip desalted67 and finally dried by vacuum centrifugation and stored at −20 °C. Phosphopeptides were isolated using TiO2 enrichment protocol as described in ref. 28 and finally lyophilized. When required, Alkaline Phosphatase treatment was performed as recommended by manufacturer (Thermo Scientific. FastAP Thermosensitive Alkaline Phosphatase; EF0654). Samples were micro-tip desalted as described above before analysed by MALDI TOF MS (UltraFlex II, Bruker Daltonics, Bremen). Isolation of Flagellum for orthogonal ESI MS/MS verification Flagellum was isolated from H10407 as described by68 with a few modifications. Briefly described, 200 ml of bacteria were grown into mid-exponential phase. One hundred OD600 units were harvested at 6000 × g for 15 min at 4 °C. The cell pellet was resuspended in 2 ml of 10 mM Tris-HCl pH 8, 75 mM NaCl and incubated at 60 °C for 20 min. A PRO200 homogenizer (PRO Scientific) set in position 2 (~10000 r.p.m.) was used to detach the fimbriae in two pulses lasting two minutes at 4 °C. A cell free supernatant was obtained by centrifugation at 14000 × g for 10 min at 4 °C and sterile filtration (0.22 μm). Flagellae and other proteins in supernatant were isolated using the improved Wessel/Flügge method as described above, by keeping the ratios between aqueous and organic solvents fixed. 1D SDS-PAGE and In-gel digestion Crude FliC extract was re-suspended in 1x SDS loading buffer (60 mM Tris-HCl pH 6.8, 2% SDS, 10% glycerol, 0.005% bromophenol blue, 5 mM EDTA, 0.1 M DTT) to a final conc. of 0.5 OD600 unit/μl. Protein fractions were boiled for 5 min. Crude FliC fraction was separated on a 4–12% NuPage novex Bis-Tris mini gel (Invitrogen). FliC protein band was visualized using colloidal coomassie blue staining and subsequently excised69. Protein bands were in-gel digested70 and micro-tip desalted67 and finally dried by vacuum centrifugation and stored at −20 °C. BEMAP BEMAP protocol was carried out with either 1.5 mg outer membrane or 0.3 mg vesicle derived peptides as input. Lyophilized peptide sample was resuspended in 25 μl Alkaline Phosphatase (AP) solution containing 10U of thermosensitive Alkaline Phosphatase (ThermoFischer scientific; EF0654). Phosphatase reaction continued for 45min at 37 °C. 75 μl BEMAP reaction mixture was added to the AP solution. The final concentration of BEMAP chemicals in the 100 μl reaction volume was 0.4 M 2-AEP (Sigma; 268674), 0.75 M NaOH (Sigma; S8045), 20 mM Ba(OH)2 (Sigma; 433373). The BEMAP reaction was incubate at 37 °C in a heating block for 3.15 hours shaking at 1300 r.p.m. The reaction was stopped by acidification (1% TFA final concentration). Sample volume was increased to 1 ml and the peptides were purified on an Oasis HLB Plus short cartridge (Waters) as recommend by manufacturer and subsequently lyophilized. TiO2 enrichment was performed as described by28. Mass spectrometric analysis of BEMAP samples The BEMAP-enriched peptides were dissolved in 0.1% formic acid and separated by nano-LC-MS/MS on an in-house packed 17 cm × 100 μm Reprosil-Pur C18-AQ column (3 μm; Dr. Maisch GmbH, Germany) using an Easy-LC nano-HPLC (Thermo Scientific, Germany). The HPLC gradient was 0–34% solvent B (A = 0.1% formic acid; B = 90% ACN, 0.1% formic acid) in 180 mins at a flow of 250 nL/min. Mass spectrometric analysis was performed using an LTQ Orbitrap Velos (Thermo Scientific, Bremen, Germany). An MS scan (400–2000 m/z) was recorded in the Orbitrap at a resolution of 30,000 at 400 m/z for a target of 1e6 ions. The top seven most intense ions were fragmented by HCD MS/MS using the following parameters: activation time = 0.1 ms, normalized energy = 48, dynamic exclusion enabled with repeat count 1, exclusion duration = 30 s, intensity threshold = 5000, target ions = 2e5. HILIC fractionation Tryptic peptides isolated from flagellin were resuspended in 90% ACN, 0.1% TFA and injected onto an in-house packed TSKgel Amide-80 HILIC (Tosoh, 5 μm) 320 μm × 170 mm μHPLC column using an Agilent 1200 HPLC system50. The peptides were eluted using a gradient from 90% ACN, 0.1% TFA to 60% ACN, 0.1% TFA over 35 mins at a flow rate of 6 μl/min. Fractions were automatically collected at 1 min intervals after UV detection at 210 nm and the fractions were combined to a total of 12–14 fractions according to UV detection. All fractions were dried by vacuum centrifugation. Orthogonal mass spectrometric analysis of FliC Peptides were HILIC fractionated as described above before analyzed by an Easy-nLC and nanospray source (Thermo Fisher Scientific) coupled with a Q-Exactive Plus mass spectrometer (Thermo Fisher Scientific, Bremen, Germany). Approximately 1 μg of in-gel digested FliC peptide (5 μl) was reconstituted in 0.1% formic acid and loaded onto a trap column at 250 bar (2 cm length, 100 μm inner diameter, ReproSil, C18 AQ 5 μm 120 Å pore (Dr. Maisch, Ammerbuch, Germany)) vented to waste via a micro-tee and eluted across a fritless analytical in-house packed resolving column (17 cm length, 75 μm inner diameter, ReproSil, C18 AQ 3 μm 120 Å pore) with a 107 min gradient of 0–30% LC-MS buffer B (LC-MS buffer A: 0.1% formic acid; LC-MS buffer B: 0.1% formic acid, 95% ACN) using a flow rate of 300 nL/min. Instrument method consisted of one survey scan (AGC target value: 1e6; R = 70K; maximum ion time: 120 milliseconds; mass range: 400 to 1400 m/z, followed by data-dependent tandem mass spectra on the top 12 most abundant precursor ions ((isolation width: 1.6 m/z; HCD collision energy (NCE): 32; MS1 signal threshold: 2e4; AGC MS2 target value: 1e6; maximum MS/MS ion time: 200 milliseconds; dynamic exclusion: repeat count of 1, maximum exclusion list size, 20 seconds wide in time, +/−10 ppm wide in m/z; doubly-charged precursors only; Minimum signal threshold of 10,000. Analysis of Mass Spectrometry Data Raw data generated on the LTQ Orbitrap Velos or Q-Exactive Plus mass spectrometer (Thermo Fisher Scientific, Bremen, Germany) were processed with Proteome Discoverer (Version 1.4.1.14, Thermo Fisher Scientific) and subjected to database searching using an in-house Mascot server (Version 2.2.04, Matrix Science Ltd., London, UK). Database searches were performed with the following parameters: Database: annotated E. coli proteomes from ETEC H10407, AIEC LF82, NMEC IHE3034 and commensal MG1655; Trypsin as the enzyme allowing a maximum of one missed cleavages sites; Carbomidomethylation of Cys as fixed modification; Deamidation of Asn and Gln; Oxidation of Met allowed as variable modification; When required BEMAP mass tag 2-AEP ((C(2) H(6) N O(2) P) on Ser/Thr or Heptose on Ser/Thr or GlcNAc on Ser/Thr was set as variable modification. Precursor and fragment mass tolerance were set to 10 ppm and 0.05 Da, respectively. Precursor mass range set from 350 Da to 7,000 Da. False discovery rate was set to 1% at peptide level using the Percolator algorithm. Our in house developed tool Peptide Finder (accessible at www.microbiology.sdu.dk) was used to extract and compile a list of uniquely modified peptides using a ProteomeDiscover output text file as template. Similarly, our in house developed script MotifX aligner (accessible at www.microbiology.sdu.dk) was used to extract and centre Ser or Thr modified residues with nine flanking residues; Number of flanking amino acids is user defined. ProteinCenter (Thermo Scientific, Germany) was used to assign Gene Ontology terms to all identified proteins. Additional Information How to cite this article: Boysen, A. et al. A novel mass spectrometric strategy ‘‘BEMAP’’ reveals extensive O-linked protein glycosylation in enterotoxigenic Escherichia coli. Sci. Rep. 6, 32016; doi: 10.1038/srep32016 (2016). Supplementary Material Tables S1-S6 Figure S1-S7 This work was supported by a project grants from the Lundbeck Foundation, grant number R31-A2459 and The Villum Kann Rasmussen Foundation, grant number 436879. GP was supported by CNPq (GP 441878/2014) and FAPESP (GP: 2014/06863-3). Author Contributions A.B. and G.P. developed the BEMAP methodology; A.B., G.P., M.R.L. and J.M.J. designed the study; A.B., G.P. and T.J.K. performed the experiments; I.G.D., M.R.L. and J.M.J. supervised experiments; A.B., G.P., T.J.K., I.G.D., M.R.L. and J.M.J. analysed the data and wrote the manuscript. Figure 1 β-Elimination of glycan moiety and replacement with 2-AEP through Michael addition chemistry. (A) MALDI MS spectrum of TTVTSGGLQR (m/z = 1181.59 Da) synthetic O-linked glycopeptide. (B) The BEMAP reaction efficiently replaces the carbohydrate moiety with the 2-AEP molecule and produces a phosphopeptide with the mass of 1126.64 Da. Minor traces of beta-eliminated as well as intact peptide can be observed (m/z = 1001.62 Da and 1181.59, receptively). (C) The AEP modified peptide is selectively enriched with TiO2 as both the glycopeptide and the beta-eliminated peptide is absent in the MALDI MS spectrum. (D) MALDI MS peptide mass fingerprint of Tryptic digest of heptosylated protein Ag43. Ag43 can be digested into a mixture of heptosylated as well as unmodified peptides. Peptides marked with an asterisk indicate heptosylation. (E) BEMAP converts heptosylated peptides into phosphopeptides; modified peptides are indicated. (F) Specific TiO2 enrichment of phosphopeptides. Figure 2 Comparison of quantitative differences between ETEC and commensal E. coli O-linked glycosylated proteins associated with the outer membrane. (A) Proportional numerical representation of identified glycopeptides in pathogenic ETEC and commensal E. coli strain. Intersecting circles indicate glycopeptides expressed in both strains. Color code: Unique to ETEC; Glycopeptides identified in both strains; Unique to non-pathogenic E. coli. (B) Proportional numerical representation of identified glycoproteins in pathogenic and commensal strain. Intersecting circles indicate glycoproteins expressed in both strains. Color code: Unique to ETEC; Glycopeptides identified in both strains; Unique to commensal E. coli; Proteins encoded by ETEC only. The list of unique modified sites was extracted and compiled using our in house developed software tool, see methods and materials. Figure 3 O-linked S/T phosphorylation and O-linked protein glycosylation are two distinct groups of post translational modifications in E. coli. Proteins associated with the outer membrane were isolated and treated either with or without Alkaline Phosphatase (AP) enzyme prior to O-linked phosphospeptide enrichment using TiO2 and LC MS/MS analysis. Proportional numerical representation was used to show identified S/T phosphorylated residues. Intersecting circles indicate phosphopeptides identified under the two different experimental conditions. Proportional numerical representation and intersecting circles is used to compare the overlap between phosphorylated and glycosylated residues expressed in both strains. Color code: Number of phosphopeptides identified without AP treatment; Number of phosphopeptides identified with AP treatment; Number of phosphopeptide identified with or without AP treatment; Number of phosphopeptides identified without AP treatment which also can be glycosylated; Number of phosphopeptide identified with or without AP treatment which also can be glycosylated. Figure 4 Characterization of ETEC O-linked outer membrane glycoproteome. The predicted biological processes of the outer membrane glycoproteome changes with an increasing number of glycosylations/protein. Absolute number of proteins within each group is shown above each column. Percentile distribution of each GO term is displayed. Color code: Cell communication; Cell division; Cell organization and biogenesis; Cellular component movement; Cellular homeostasis; Metabolic process; Regulation of biological process; Response to stimulus; Transport; Unannotated. Figure 5 ETEC FliC protein glycosylation mapping. (A) Proportional numerical representation of identified Heptosylated peptides using either BEMAP or an orthogonal experimental approach. () Number of phosphopeptides identified BEMAP; () Number of phosphopeptides identified in both approaches; ()Number of phosphopeptide using an orthogonal approach (B) Clustal Omega W sequence alignment of Salmonella typhimurium, ETEC H10407 and E. coli K-12 MG16556 showing extent of sequence similarity. Boxed purple S and T amino acids indicate Heptosylated peptides identified using either BEMAP or an orthogonal experimental approach. Green colored S and T residues highlight glycosylation only identified using an orthogonal experimental workflow. Red colored S and T residues highlight ETEC glycosylation identified using BEMAP. The symbols *, and indicate sequence identity, sequence similarity among the three strains. (C) Visualization of ETEC FliC protein glycosylation using Salmonella typhimurium as template71. Glycosylated residues are shown as yellow spheres. Figure 6 Definition of ETEC protein glycosylation sequence motifs. Glycosylated ETEC S/T residues identified in proteins associated with the outer membrane and outer membrane vesicles were centered with nine flanking residues using an in house developed script, see methods and materials. The motif-x algorithm72 was used to extract overrepresented patterns among 389 Serine and 239 Threonine sequences (p value < 0.0001). Figure (A–G) shows probable Serine sequence logos whereas (H–K) depicts Theronine motifs. A schematic motif sequence is listed below each panel as well as the probability score within each dataset. Table 1 Strains used in this study. Name Genotype Source/Reference E. coli MG1655 Wild type K-12 strain ATCC number 700926 ETEC H10407 Wild type (serotype 078:H11) Gift from Statens Serum Institute ==== Refs Turner S. M. , Scott-Tucker A. , Cooper L. M. & Henderson I. R. Weapons of mass destruction: virulence factors of the global killer enterotoxigenic Escherichia coli . FEMS Microbiol. Lett. 263 , 10 –20 (2006 ).16958845 Liu L. . Global, regional, and national causes of child mortality: an updated systematic analysis for 2010 with time trends since 2000 . Lancet 379 , 2151 –2161 (2012 ).22579125 WHO Future directions for research on enterotoxigenic Escherichia coli vaccines for developing countries . Weekly epidemiological record 97 –104 (2006 ).16671213 Fleckenstein J. M. . Molecular mechanisms of enterotoxigenic Escherichia coli infection . Microbes. Infect. 12 , 89 –98 (2010 ).19883790 Crossman L. C. . A commensal gone bad: complete genome sequence of the prototypical enterotoxigenic Escherichia coli strain H10407 . J. Bacteriol. 192 , 5822 –5831 (2010 ).20802035 Szymanski C. M. & Wren B. W. Protein glycosylation in bacterial mucosal pathogens . Nat. Rev. Microbiol. 3 , 225 –237 (2005 ).15738950 Lithgow K. V. . A general protein O-glycosylation system within the Burkholderia cepacia complex is involved in motility and virulence . Mol. Microbiol. 92 , 116 –137 (2014 ).24673753 Karlyshev A. V. . The Campylobacter jejuni general glycosylation system is important for attachment to human epithelial cells and in the colonization of chicks . Microbiology 150 , 1957 –1964 (2004 ).15184581 Lindenthal C. & Elsinghorst E. A. Enterotoxigenic Escherichia coli TibA glycoprotein adheres to human intestine epithelial cells . Infect. Immun. 69 , 52 –57 (2001 ).11119488 Grass S. . The Haemophilus influenzae HMW1 adhesin is glycosylated in a process that requires HMW1C and phosphoglucomutase, an enzyme involved in lipooligosaccharide biosynthesis . Mol. Microbiol. 48 , 737 –751 (2003 ).12694618 Schirm M. . Structural, genetic and functional characterization of the flagellin glycosylation process in Helicobacter pylori . Mol. Microbiol. 48 , 1579 –1592 (2003 ).12791140 Iwashkiw J. A. . Identification of a general O-linked protein glycosylation system in Acinetobacter baumannii and its role in virulence and biofilm formation . PLoS. Pathog. 8 , e1002758 (2012 ).22685409 Moormann C. , Benz I. & Schmidt M. A. Functional substitution of the TibC protein of enterotoxigenic Escherichia coli strains for the autotransporter adhesin heptosyltransferase of the AIDA system . Infect. Immun. 70 , 2264 –2270 (2002 ).11953358 Sherlock O. , Dobrindt U. , Jensen J. B. , Munk V. R. & Klemm P. Glycosylation of the self-recognizing Escherichia coli Ag43 autotransporter protein . J. Bacteriol. 188 , 1798 –1807 (2006 ).16484190 Benz I. & Schmidt M. A. Glycosylation with heptose residues mediated by the aah gene product is essential for adherence of the AIDA-I adhesin . Mol. Microbiol. 40 , 1403 –1413 (2001 ).11442838 Fleckenstein J. M. , Roy K. , Fischer J. F. & Burkitt M. Identification of a two-partner secretion locus of enterotoxigenic Escherichia coli . Infect. Immun. 74 , 2245 –2258 (2006 ).16552055 Dube D. H. , Champasa K. & Wang B. Chemical tools to discover and target bacterial glycoproteins . Chem. Commun. (Camb.) 47 , 87 –101 (2011 ).20730171 Scott N. E. . Comparative proteomics and glycoproteomics reveal increased N-linked glycosylation and relaxed sequon specificity in Campylobacter jejuni NCTC11168 O . J. Proteome. Res. 13 , 5136 –5150 (2014 ).25093254 Geyer H. & Geyer R. Strategies for analysis of glycoprotein glycosylation . Biochim. Biophys. Acta 1764 , 1853 –1869 (2006 ).17134948 Pan S. . Mass spectrometry based targeted protein quantification: methods and applications . J. Proteome. Res. 8 , 787 –797 (2009 ).19105742 Hitchen P. G. & Dell A. Bacterial glycoproteomics . Microbiology 152 , 1575 –1580 (2006 ).16735721 Scott N. E. . Simultaneous glycan-peptide characterization using hydrophilic interaction chromatography and parallel fragmentation by CID, higher energy collisional dissociation, and electron transfer dissociation MS applied to the N-linked glycoproteome of Campylobacter jejuni . Mol. Cell Proteomics. 10 , M000031–MMCP201 (2011 ). Nothaft H. & Szymanski C. M. Protein glycosylation in bacteria: sweeter than ever . Nat. Rev. Microbiol. 8 , 765 –778 (2010 ).20948550 Champasa K. , Longwell S. A. , Eldridge A. M. , Stemmler E. A. & Dube D. H. Targeted identification of glycosylated proteins in the gastric pathogen Helicobacter pylori . Mol. Cell Proteomics. 12 (9), 2568 –2586 , doi:10.1074 (Sep. 2013 ).23754784 Scott N. E. . Diversity within the O-linked protein glycosylation systems of acinetobacter species . Mol. Cell Proteomics. 13 , 2354 –2370 (2014 ).24917611 Oda Y. , Nagasu T. & Chait B. T. Enrichment analysis of phosphorylated proteins as a tool for probing the phosphoproteome . Nat. Biotechnol. 19 , 379 –382 (2001 ).11283599 Wells L. . Mapping sites of O-GlcNAc modification using affinity tags for serine and threonine post-translational modifications . Mol. Cell Proteomics. 1 , 791 –804 (2002 ).12438562 Larsen M. R. , Thingholm T. E. , Jensen O. N. , Roepstorff P. & Jørgensen T. J. Highly selective enrichment of phosphorylated peptides from peptide mixtures using titanium dioxide microcolumns . Mol. Cell Proteomics. 4 , 873 –886 (2005 ).15858219 Huang H. . Simultaneous enrichment of cysteine-containing peptides and phosphopeptides using a Cysteine-specific Phosphonate Adaptable Tag (CysPAT) in combination with TiO2 chromatography . Mol. Cell Proteomics. (2016 ) Jun 8. pii: mcp.M115.054551. [Epub ahead of print]. Parker B. L. , Gupta P. , Cordwell S. J. , Larsen M. R. & Palmisano G. Purification and identification of O-GlcNAc-modified peptides using phosphate-based alkyne CLICK chemistry in combination with titanium dioxide chromatography and mass spectrometry . J. Proteome. Res. 10 , 1449 –1458 (2011 ).21158410 Jensen S. S. & Larsen M. R. Evaluation of the impact of some experimental procedures on different phosphopeptide enrichment techniques . Rapid Commun. Mass Spectrom. 21 , 3635 –3645 (2007 ).17939157 Knudsen S. K. , Stensballe A. , Franzmann M. , Westergaard U. B. & Otzen D. E. Effect of glycosylation on the extracellular domain of the Ag43 bacterial autotransporter: enhanced stability and reduced cellular aggregation . Biochem. J. 412 , 563 –577 (2008 ).18341480 Macek B. . Phosphoproteome analysis of E. coli reveals evolutionary conservation of bacterial Ser/Thr/Tyr phosphorylation . Mol. Cell Proteomics. 7 , 299 –307 (2008 ).17938405 Hansen A. M. . The Escherichia coli phosphotyrosine proteome relates to core pathways and virulence . PLoS. Pathog. 9 , e1003403 (2013 ).23785281 Qu Y. . Automated immobilized metal affinity chromatography system for enrichment of Escherichia coli phosphoproteome . Electrophoresis 34 , 1619 –1626 (2013 ).23494780 Soares N. C. , Spat P. , Krug K. & Macek B. Global dynamics of the Escherichia coli proteome and phosphoproteome during growth in minimal medium . J. Proteome. Res. 12 , 2611 –2621 (2013 ).23590516 Beis K. , Whitfield C. , Booth I. & Naismith J. H. Two-step purification of outer membrane proteins . Int. J. Biol. Macromol. 39 , 10 –14 (2006 ).16423387 Garcia E. C. , Brumbaugh A. R. & Mobley H. L. Redundancy and specificity of Escherichia coli iron acquisition systems during urinary tract infection . Infect. Immun. 79 , 1225 –1235 (2011 ).21220482 Lindenthal C. & Elsinghorst E. A. Identification of a glycoprotein produced by enterotoxigenic Escherichia coli . Infect. Immun. 67 , 4084 –4091 (1999 ).10417177 Ashburner M. . Gene Ontology: tool for the unification of biology . Nature Genetics 25 , 25 –29 (2000 ).10802651 Kulp A. & Kuehn M. J. Biological functions and biogenesis of secreted bacterial outer membrane vesicles . Annu. Rev. Microbiol. 64 , 163 –184 (2010 ).20825345 Kesty N. C. , Mason K. M. , Reedy M. , Miller S. E. & Kuehn M. J. Enterotoxigenic Escherichia coli vesicles target toxin delivery into mammalian cells . EMBO J. 23 , 4538 –4549 (2004 ).15549136 Boysen A. , Borch J. , Krogh T. J. , Hjerno K. & Moller-Jensen J. SILAC-based comparative analysis of pathogenic Escherichia coli secretomes . J. Microbiol. Methods 116 , 66 –79 (2015 ).26143086 Merino S. & Tomas J. M. Gram-negative flagella glycosylation . Int. J. Mol. Sci. 15 , 2840 –2857 (2014 ).24557579 Roy K. . Enterotoxigenic Escherichia coli EtpA mediates adhesion between flagella and host cells . Nature 457 , 594 –598 (2009 ).19060885 Schirm M. . Structural and genetic characterization of glycosylation of type a flagellin in Pseudomonas aeruginosa . J Bacteriol. 186 , 2523 –2531 (2004 ).15090491 Thibault P. . Identification of the carbohydrate moieties and glycosylation motifs in Campylobacter jejuni flagellin . J Biol. Chem. 276 , 34862 –34870 (2001 ).11461915 Schirm M. , Schoenhofen I. C. , Logan S. M. , Waldron K. C. & Thibault P. Identification of unusual bacterial glycosylation by tandem mass spectrometry analyses of intact proteins . Anal. Chem. 77 , 7774 –7782 (2005 ).16316188 Verma A. . Glycosylation of b-Type flagellin of Pseudomonas aeruginosa: structural and genetic basis . J. Bacteriol. 188 , 4395 –4403 (2006 ).16740946 Palmisano G. . Selective enrichment of sialic acid-containing glycopeptides using titanium dioxide chromatography with analysis by HILIC and mass spectrometry . Nat. Protoc. 5 , 1974 –1982 (2010 ).21127490 Beatson S. A. , Minamino T. & Pallen M. J. Variation in bacterial flagellins: from sequence to structure . Trends Microbiol. 14 , 151 –155 (2006 ).16540320 Yonekura K. . The bacterial flagellar cap as the rotary promoter of flagellin self-assembly . Science 290 , 2148 –2152 (2000 ).11118149 Smedley J. G. III . Influence of pilin glycosylation on Pseudomonas aeruginosa 1244 pilus function . Infect. Immun. 73 , 7922 –7931 (2005 ).16299283 Marceau M. , Forest K. , Beretti J. L. , Tainer J. & Nassif X. Consequences of the loss of O-linked glycosylation of meningococcal type IV pilin on piliation and pilus-mediated adhesion . Mol. Microbiol. 27 , 705 –715 (1998 ).9515697 Charbonneau M. E. . A structural motif is the recognition site for a new family of bacterial protein O-glycosyltransferases . Mol. Microbiol. 83 , 894 –907 (2012 ).22304382 Schwartz D. & Gygi S. P. An iterative statistical approach to the identification of protein phosphorylation motifs from large-scale data sets . Nat. Biotechnol. 23 , 1391 –1398 (2005 ).16273072 Stepan H. & Staudacher E. Optimization of monosaccharide determination using anthranilic acid and 1-phenyl-3-methyl-5-pyrazolone for gastropod analysis . Anal. Biochem. 418 , 24 –29 (2011 ).21802397 Jensen P. H. , Karlsson N. G. , Kolarich D. & Packer N. H. Structural analysis of N- and O-glycans released from glycoproteins . Nat. Protoc. 7 , 1299 –1310 (2012 ).22678433 Ong S. E. . Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics . Mol. Cell Proteomics. 1 , 376 –386 (2002 ).12118079 Ross P. L. . Multiplexed protein quantitation in Saccharomyces cerevisiae using amine-reactive isobaric tagging reagents . Mol. Cell Proteomics. 3 , 1154 –1169 (2004 ).15385600 Gault J. . Neisseria meningitidis Type IV Pili Composed of Sequence Invariable Pilins Are Masked by Multisite Glycosylation . PLoS. Pathog. 11 , e1005162 (2015 ).26367394 Weinert B. T. . Lysine succinylation is a frequently occurring modification in prokaryotes and eukaryotes and extensively overlaps with acetylation . Cell Rep. 4 , 842 –851 (2013 ).23954790 Weinert B. T. . Acetyl-phosphate is a critical determinant of lysine acetylation in E. coli . Mol. Cell 51 , 265 –272 (2013 ).23830618 Harding C. M. . Acinetobacter strains carry two functional oligosaccharyltransferases, one devoted exclusively to type IV pilin, and the other one dedicated to O-glycosylation of multiple proteins . Mol. Microbiol. 96 , 1023 –1041 (2015 ).25727908 Sambrook J. & Green M. R. Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory Press, New York, 2012 ). Wessel D. & Flügge U. I. A method for the quantitative recovery of protein in dilute solution in the presence of detergents and lipids . Anal. Biochem. 138 , 141 –143 (1984 ).6731838 Rappsilber J. , Mann M. & Ishihama Y. Protocol for micro-purification, enrichment, pre-fractionation and storage of peptides for proteomics using StageTips . Nat. Protoc. 2 , 1896 –1906 (2007 ).17703201 Munera D. , Hultgren S. & Fernandez L. A. Recognition of the N-terminal lectin domain of FimH adhesin by the usher FimD is required for type 1 pilus biogenesis . Mol. Microbiol. 64 , 333 –346 (2007 ).17378923 Candiano G. . Blue silver: a very sensitive colloidal Coomassie G-250 staining for proteome analysis . Electrophoresis 25 , 1327 –1333 (2004 ).15174055 Morsczeck C. . Streptococcus pneumoniae: proteomics of surface proteins for vaccine development . Clin. Microbiol. Infect. 14 , 74 –81 (2008 ).18034862 Yonekura K. , Maki-Yonekura S. & Namba K. Complete atomic model of the bacterial flagellar filament by electron cryomicroscopy . Nature 424 , 643 –650 (2003 ).12904785 Chou M. F. & Schwartz D. Using the scan-x Web site to predict protein post-translational modifications . Curr. Protoc. Bioinformatics. Chapter 13, Unit (2011 ).

PMC005xxxxxx/PMC5000013.txt

==== Front Sci RepSci RepScientific Reports2045-2322Nature Publishing Group srep3240510.1038/srep32405ArticleAIM 2 inflammasomes regulate neuronal morphology and influence anxiety and memory in mice Wu Pei-Jung 12Liu Hsin-Yu 2Huang Tzyy-Nan 2Hsueh Yi-Ping a121 Graduate Institute of Life Sciences, National Defense Medical Center, Taipei 114, Taiwan2 Institute of Molecular Biology, Academia Sinica, Taipei 115, Taiwana yph@gate.sinica.edu.tw26 08 2016 2016 6 3240506 04 2016 09 08 2016 Copyright © 2016, The Author(s)2016The Author(s)This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/Inflammasomes are the protein assemblies that consist of inflammasome sensors, adaptor apoptosis-associated speck-like proteins containing a CARD (ASC) and inflammasome caspase. Inflammasomes sense multiple danger signals via various inflammasome sensors and consequently use caspase to trigger proteolytic processing and secretion of IL-1β cytokines. Recent studies have suggested that neurons use their own innate immune system to detect danger signals and regulate neuronal morphology. Here, we investigate whether inflammasomes, the critical components of innate immunity, participate in regulation of neuronal morphology and function. Among various sensors, Absent in melanoma 2 (Aim2) expression in neurons is most prominent. Adding synthetic double-stranded DNA (dsDNA) to cultured neurons induces IL-1β secretion in an AIM2-dependent manner and consequently downregulates dendritic growth but enhances axon extension. The results of Aim2 knockout and knockdown show that AIM2 acts cell-autonomously to regulate neuronal morphology. Behavioral analyses further reveal that Aim2−/− mice exhibit lower locomotor activity, increased anxious behaviors and reduced auditory fear memory. In conclusion, our study suggests that AIM2 inflammasomes regulate neuronal morphology and influence mouse behaviors. ==== Body Innate immunity is a naive defense system that neutralizes and removes various foreign pathogens. It also acts as an alarm system to respond to intrinsic danger signals, such as stress and neurodegeneration12. In addition, immune responses are highly associated with neurodevelopmental disorders, such as autism spectrum disorders and schizophrenia3456. In brains, microglial cells are recognized as the specialized immune cells that respond to infection and neurodegeneration789. However, neurons have also been shown to express critical molecules that are involved in innate immune responses, such as Toll-like receptors (TLRs) and the downstream adaptors myeloid differentiation primary response protein 88 (MYD88), the Toll/IL-1 receptor (TIR) domain-containing adaptor inducing IFN-β (TRIF) and the Sterile α and TIR motif–containing protein 1 (SARM1), as well as inflammasomes1011121314. Neurons also produce inflammatory and antiviral cytokines, as well as corresponding receptors for these cytokines1315. Overall, neurons have their own innate immunity, which responds to endogenous ligands116 to regulate neurogenesis, neural differentiation and neurodegeneration1316171819. Among various TLRs, TLR7 has been extensively studied in the nervous system. TLR7 recognizes single-stranded RNAs, including bacterial and viral RNAs, as well as endogenous mRNAs and miRNAs1162021. TLR7 activation delivers a signal to MYD88 and c-FOS, thereby inducing IL-6 expression to restrict axon and dendrite outgrowth of cultured neurons through both autocrine and paracrine signaling mechanisms1316. Interestingly, TLR7 is activated in neuronal cultures in the absence of exogenous ligand. RNAs released from dead cells or miRNAs released through exosomes likely play roles in TLR7 activation of cultured neurons131619. Echoing the role of TLR7 in neural differentiation, exploratory activity of Tlr7 knockout mice is altered at the juvenile stage13. In addition to IL-6, neuronal TLR7 activation also increases the level of Il-1β mRNA expression, but the level of secreted IL-1β proteins in the supernatants of cultured neurons is not increased13. Cytoplasmic pro-IL-1β must be cleaved by an inflammasome—a composite assembly that consists of inflammatory caspases, adaptor apoptosis-associated speck-like proteins containing a CARD (ASC) and various sensors to detect inflammation-inducing stimuli22 —to produce and secrete the mature form of IL-1β2324252627. Different inflammatory sensors detect various secondary signals, rather than the signals that activate TLRs, to activate inflammasomes2728. It is likely that a proper secondary signal is necessary to activate neuronal inflammasomes in order to induce IL-1β processing in neurons. The inflammasome sensors belong to two families: the nucleotide-binding oligomerization domain-like receptor (NLR) family and the PYHIN (hematopoietic interferon-inducible nuclear antigens with 200-amino-acid repeat-domain-containing protein) family. Among these various sensors, NLRP3 is the most studied inflammasome sensor in brain2930313233. NLRP1, NLRC4 and AIM2 are also involved in brain injury and the innate immune response in brain14313435. Other sensors, such as NLRP6 and NLRP12, have been less extensively studied3637. AIM2 is a member of the PYHIN family, while NLRP1, NLRP3, NLRP6, NLRP12 and NLRC4 all belong to the NOD-like receptor family22. In this report, we explore whether neurons are able to sense stimulation and activate inflammasomes and investigate whether, similar to TLRs, inflammasome activation regulates neuronal morphogenesis and mouse behaviors. Results Aim2 is expressed in developing neurons To investigate the potential role of inflammasomes in neurons, we first examined expression of various inflammasome-related genes. Quantitative-PCR with the Universal ProbeLibrary system was used to detect the mRNA levels of Nlrp1a, Nlrp1b, Nlrp1c, Nlrp3, Nlrp6, Nlrp12, Nlrc4, Aim2 and Asc. The results showed that, in cultured mouse cortical and hippocampal neurons, Aim2 was the most highly expressed inflammasome sensor compared to the other inflammasome sensors (Fig. 1a). This result echoes a previous study showing Aim2 expression in neurons14. To further confirm this result, we performed quantitative-PCR using the SYBR Green system with another two different sets of primers for Aim2. Nlrp3, the most well-studied inflammasome sensor, was also included to compare with Aim2. Aim2 expression levels were much higher (~10-fold) than those of Nlrp3 (Fig. 1b). In addition to the cultured neurons, adult mouse brains were also analyzed. Although the expression levels of Nlrp3 and Nlrc4 were noticeably increased in adult brains compared with those of cultured neurons, Aim2 expression levels were still the most elevated among the different inflammasome sensors when detected with the Universal ProbeLibrary system, as well as with the SYBR Green system (Fig. 1c,d). To further confirm the higher expression levels of Aim2 in the brain, the actual copy number of Aim2 was measured by absolute quantitative RT-PCR. In adult mouse brains, the copy number of Aim2 mRNA was estimated to be ~190 per ng of total RNA, which was 6.2- and 18.2-fold higher than those of Nlrp3 and Nlrc4, respectively (Fig. 1e), further supporting the high expression of Aim2 in the brain. We then monitored the temporal expression of Aim2 and Nlrp3 by examining mouse brains at postnatal (P) day 0, 3 and 7, as well as at two months. Compared with P0 and P3, the brains of mice at two months of age had reduced (~50%) expression of Aim2 (Fig. 1f). In contrast, the mRNA expression level of Nlrp3 doubled in the brain during development from P0 to two months (Fig. 1g), though its expression level was still lower than that of Aim2 (Fig. 1e). These expression analyses suggest that AIM2 likely plays a role in brains. IL-1β is released by cultured neurons upon TLR7 and AIM2 activation To examine AIM2 function in neurons, we first monitored IL-1β secretion upon AIM2 activation. Since TLR7 activation increases IL-1β mRNA expression13, to maximize the effect of AIM2 activation, we applied CL075 and poly-deoxyadenylic-deoxythymidylic acid (poly dAdT) —the synthetic double-stranded DNA (dsDNA) —to activate TLR7 and AIM2, respectively, in cultured cortical and hippocampal neurons. The amounts of IL-1β in the supernatant of neurons treated with CL075 and poly dAdT increased approximately 5-6-fold when compared with the vehicle control (Fig. 1h). AIM2 activation by poly dAdT alone slightly induced IL-1β expression, but the difference was not significant (Fig. 1h). The combinatory effect of CL075 and poly dAdT on IL-1β production in the cultured neurons required AIM2 because Aim2−/− neurons did not produce IL-1β in the presence of CL075 and poly dAdT (Fig. 1h). These results suggest that AIM2 expression enables neurons to produce IL-1β under dsDNA stimulation and TLR7 activation. AIM2 regulates neuronal morphology A previous study suggested the role of IL-1β in controlling the cell morphology of cultured neurons38. It seems possible that AIM2 controls IL-1β secretion and thereby regulates neuronal morphology. We investigated the role of AIM2 in neuronal differentiation and studied whether IL-1β acts downstream of AIM2 in regulating neuronal development. We first manipulated the expression levels of Aim2 and monitored the dendritic and axonal growth of cultured neurons. Two different experiments were performed. The first experiment was to compare neuronal morphology of wild-type and Aim2−/− neurons. Total dendrite length, primary dendrite number and the number of dendritic branch tips were determined at 3, 5 and 7 DIV. The results showed that Aim2−/− neurons had longer total dendrite length at 5 and 7 DIV, more primary dendrites at 5 DIV and more branch tips at 7 DIV (Fig. 2a). In measuring axon formation, Aim2 deletion reduced both primary and total axon length and the number of axonal branch tips at 3 DIV (Fig. 2b). To confirm that the morphological changes in Aim2−/− neurons are due to Aim2 deletion, we reintroduced Aim2 to Aim2−/− neurons. The results showed that restoration of Aim2 expression reduced total dendrite length, primary dendrite number and the number of dendritic branch tips (Fig. 2c). Similarly, re-expression of Aim2 had the opposite effect on axonal growth—increasing both total and primary axon length (Fig. 2d). This suggests that the role of AIM2 in neurons is more complex than a simple control of cell growth or death, instead likely involving a specific function to influence dendritic and axonal differentiation. The second experiment was to examine the effect of AIM2 in vivo using Golgi staining to analyze Aim2−/− mouse brains. In four-month-old mouse brains, Aim2−/− CA1 neurons exhibited a noticeably different pattern of dendritic arbors compared with WT neurons (Fig. 3a,b). For basal dendrites, both total dendrite length and the number of total dendritic branch tips were noticeably increased in Aim2−/− CA1 neurons (Fig. 3c). In contrast, Aim2 deletion did not influence the length and branching level of apical dendrites (Fig. 3c). Because neonatal brains had higher Aim2 expression levels compared to adult brains (Fig. 1f), we also analyzed dendritic arbors in P7 brains. Similar to adult brains, total basal dendrite length was increased in Aim2−/− CA1 neurons at P7 (Fig. 3d,e), although the total number of branch tips did not differ between WT and Aim2−/− brains (Fig. 3d,e). Upon observing apical dendrites, we found that total dendrite length was not noticeably affected by Aim2 deletion. Nevertheless, the branching level of apical dendrites was higher in Aim2−/− CA1 neurons at P7 (Fig. 3d,e). These in vivo analyses support the role of AIM2 in controlling neuronal morphology at both neonatal and adult stages. Activation of AIM2 inflammasomes influences neuronal morphology In addition to manipulating Aim2 expression, we also investigated the role of AIM2 in neurons by activating AIM2 inflammasomes. We compared CL075 alone, poly dAdT alone, and a combination of CL075 and poly dAdT to the vehicle control. In our previous study, we showed that TLR7 activation by CL075 induces IL-6 expression, which ultimately restricts dendritic growth13. To exclude the contribution of IL-6, Il-6−/− cortical and hippocampal neurons were used. Consistent with our previous data13, CL075 alone did not alter the dendritic growth of Il-6−/− neurons (Fig. 4a,b). When CL075 was combined with poly dAdT, all dendritic parameters, including the total dendrite length, primary dendrite number and the number of dendritic branches, were noticeably reduced in Il-6−/− neurons (Fig. 4a,b), which suggests a negative role for AIM2 inflammasomes in dendritic growth. We also noticed that poly dAdT treatment alone was sufficient to downregulate the dendritic growth of Il-6−/− neurons (Fig. 4b), although it did not noticeably induce the amount of IL-1β expression observed for cultured neurons (Fig. 1h). To confirm the involvement of AIM2 in the effect of poly dAdT on neuronal morphology, poly dAdT was added into Aim2−/− neurons. Addition of poly dAdT did not influence dendritic growth of Aim2−/− neurons (Fig. 4c,d). These results suggest that AIM2 senses dsDNA in the environment to regulate morphology of cultured neurons. IL-1β signaling is required for AIM2 to regulate neuronal morphology The aforementioned results suggest that addition of poly dAdT is sufficient to activate AIM2 to restrict dendritic growth but is unable to noticeably increase IL-1β secretion into culture medium (Figs 1h and 4b). It is possible that the amount of IL-1β necessary to restrict dendritic growth was lower than the detection limit of our instruments. Alternatively, in addition to IL-1β, AIM2 might use another effector to downregulate dendritic growth. To investigate whether IL-1β is involved downstream of AIM2 in regulating neuronal morphology, we examined the effect of poly dAdT on two mutant neurons: 1) Il1r1−/− neurons, which lack the most well-studied receptor for IL-1; and 2) Myd88−/− neurons, which lack the key downstream adaptor molecule MYD88 for IL-1 receptor signaling39. Both these mutant neurons are unable to respond to IL-1β. We found that, unlike wild-type neurons, poly dAdT did not influence axonal and dendritic growth of Il1r1−/− neurons (Fig. 5a,b). Similarly, Myd88−/− neurons also did not respond to poly dAdT in terms of controlling axonal and dendritic growth (Fig. 5c,d). These data support the notion that IL-1β is required for poly dAdT to control neuronal morphology. AIM2 activation cell-autonomously regulates neuronal morphology Although we found that AIM2 responded to poly dAdT stimulation and used the IL-1β pathway to regulate neuronal morphology, our data also show that poly dAdT alone did not significantly increase IL-1β secretion into the culture supernatant. It seems possible that a very low amount of IL-1β is sufficient to regulate neuronal morphology, in which case an autocrine mechanism is likely the most effective way for a low amount of neuronal IL-1β to control neuronal morphology. To investigate whether AIM2 uses a cell-autonomous mechanism to influence neuronal morphology, we reduced Aim2 expression by transfecting an Aim2 knockdown construct. In our neuronal culture systems, transfection efficiency is usually less than 1%. Therefore, transfected neurons are surrounded by untransfected neurons, providing a good system to test the cell-autonomous effect of AIM2. If AIM2 acts cell-autonomously to regulate neuronal morphology, reducing Aim2 expression by an Aim2 knockdown construct should be sufficient to generate the abnormal neuronal morphology observed for transfected neurons. To investigate this possibility, we first constructed five different artificial miRNA knockdown constructs of Aim2. In HEK293T cells, miR-Aim2#2 exhibited the best effect in reducing the protein expression of Myc-tagged Aim2 (Fig. 6a). MiR-Aim2#2 (termed MiR-Aim2 hereafter) was then transfected into neurons. The results showed that miR-Aim2 expression in neurons resulted in longer total dendritic length and shorter total axonal length (Fig. 6b,c), which is similar to the phenotypes of Aim2−/− neurons (Fig. 2). Moreover, if AIM2 activation acts non-cell-autonomously to regulate neuronal morphology, Aim2 knockdown neurons should still be influenced by IL-1β secreted from neighboring neurons upon poly dAdT treatment. To investigate this possibility, poly dAdT treatment was included. We found that poly dAdT restricted dendritic arborization of neurons transfected with non-silencing control miR-Ctrl, but this effect did not occur in the Aim2 knockdown neurons (Fig. 6d). Taken together, these results suggest that the effect of poly dAdT on neuronal morphology is mediated by an AIM2-dependent cell-autonomous mechanism. Aim2 deletion induces anxious behaviors and impairs auditory fear memory The above results suggest the role of AIM2 in regulation of neuronal morphology. We then wondered whether AIM2 is relevant for brain functioning. A series of behavioral paradigms, including open field, light-dark box and fear conditioning, were applied to analyze Aim2−/− and WT mice. In open field, Aim2−/− mice had noticeably lower locomotor activity, as shorter moving distances were recorded on all three examination days (Fig. 7a). Because the moving speed of Aim2−/− mice was comparable to that of WT mice (Fig. 7a), the shorter moving distance was not caused by lower moving speed. The total number of rearing events also tended to be lower in Aim2−/− mice, but this was only statistically significant for Day 2 (Fig. 7a). We noticed that Aim2−/− mice preferred to stay in the corners of their boxes (Fig. 7a), and also defecated and urinated more in the open field (Fig. 7a), both of which are suggestive of anxious behaviors in Aim2−/− mice. To confirm this point, we performed light-dark box experiments. Indeed, compared with WT animals, Aim2−/− mice spent more time in the dark box, though the numbers of transitions between light and dark boxes were comparable between WT and Aim2−/− mice (Fig. 7b). These data support that Aim2 deletion induces anxiety. In fear conditioning, the freezing responses during habituation and the period right after foot shocks were comparable between Aim2−/− and WT mice (Fig. 7c). However, one day after auditory fear conditioning, Aim2−/− mice had a noticeably lower freezing percentage in response to tone stimulus (Fig. 7c), suggesting a role for AIM2 in regulating associative memory. In conclusion, our analyses suggest that Aim2 deletion influences brain functions. Discussions In this report, our data suggest that neuronal AIM2 inflammasomes are able to sense dsDNA to promote IL-1β secretion by neurons. IL-1β then acts as an autocrine signaling mechanism to downregulate dendritic growth, whilst promoting axonal growth. In classical innate immune responses, the combinatory signals that activate TLRs and inflammasomes provoke proteolytic processes and massive secretion of the IL-1 family of cytokines2740. In cultured neurons, we also found that, by adding exogenous agonists, activation of both TLR7 and the AIM2 inflammasomes promotes IL-1β secretion. However, in the absence of exogenous ligands, manipulation of Aim2 expression levels in cultured neurons was sufficient to regulate neuronal morphology, which is perhaps due to endogenous ligands released from dead cells or healthy neurons in cultures, echoing previous studies showing that endogenous ligands of TLR7 regulate neuronal death and morphology131619. In this scenario, it is likely that there is a basal level of IL-1β in cultures, which is also able to regulate cell growth of neurons. Manipulation of Aim2 expression levels may influence the processing of this basal level of IL-1β. Since the level of secreted IL-1β is so low, it can only act as an autocrine signal to cell-autonomously control neuronal morphology. This mechanism is likely involved in regulation of neural development in vivo. As for our analysis of Aim2−/− brains, we found that without any immune challenge, Aim2−/− CA1 neurons already had a more complex dendritic arbor at P7. Perhaps, apoptotic cells during development serve as a source of endogenous ligands to activate the AIM2 inflammasomes of neurons. These neuronal innate immune responses allow neurons to detect their growing environment and avoid extending their dendrites to regions with dead cells. Our study suggests that AIM2 activation restricts dendritic growth but promotes axonal extension. Thus, the effect of AIM2 on neuronal morphology is unlikely to be due to the induction of cell death. In fact, IL-1β has been shown to have a differential effect on dendrites and axons. It can downregulate the dendritic growth of cultured neurons38. For axonal growth, however, the reported function of IL-1β is controversial. During the development of sympathetic neurons, 100 ng/ml IL-1β inhibits nerve growth factor-induced axonal growth41. In contrast, IL-1β at a concentration of 50 ng/ml stimulates neurite growth of enteric neurons42. High doses (500 ng/ml) of IL-1β trigger axonal growth of entorhinal cortex slice cultures, but not transverse spinal cord slice cultures43. Thus, it seems likely that different types of neurons respond differently to IL-1β in terms of axonal growth. Nevertheless, taken with our findings, these studies suggest IL-1β may play a role in the regulation of neuronal development, not just induction of pyroptosis (a type of cell death induced by inflammation). In addition to IL-1β, there are several examples showing that cytokines or growth factors are able to act as autocrine signals regulating neuronal growth. Our previous study showed that TLR7 uses IL-6 to restrict axonal and dendritic growth16. Depending on the concentration, IL-6 may act as an autocrine or paracrine signal to regulate neuronal morphology1316. Neurons are able to secrete brain-derived neurotrophic factor (BDNF) at the growth cone. Locally-secreted BDNF then cell-autonomously regulates axonal differentiation44. Taken together, autocrine signaling may be a general mechanism for neurons to sense environmental cues and to control their growth and differentiation. Our study shows that Aim2 deletion induces anxious behaviors and impairs locomotion and associative memory. Previous human genetic and knockout mice have indicated the roles of IL-1 signaling in cognition45. The IL-1 receptor is heterodimeric, containing a ligand-binding subunit (IL-1R1) and essential accessory subunits. IL1RAPL1, one of the accessory subunits, regulates postsynaptic signaling46 and is associated with mental retardation4647. Deletion of Il1rapl1 in mice results in learning deficiency48, which is consistent with our observation that Aim2 deletion impairs auditory fear conditioning. However, in contrast to our finding, reduced anxiety behaviors were found in Il1rapl1−/− mice48. For Il1r1−/− mice, hyper locomotion and decreased anxiety in open field were observed49, which also contrast with the phenotype of Aim2−/− mice. Several possibilities may explain the conflicting observations in these mutant mice. For instance, there are at least nine different IL-1 receptors and accessory subunits5051. Different subunits have different expression patterns and vary in regulation or function. It is not clear if there is complex interplay between these different subunits, which could manifest in altered mouse behavior. It may also explain why many controversial results were observed for the IL-1 study45. Moreover, AIM2 is just one of various inflammasome sensors that processes pro-IL-1β and Aim2 deletion can only impair IL-1β production in response to dsDNA and perhaps only in a subset of cell types. Thus, the impact of Aim2 deletion and IL-1 receptor deficiency is likely to be different. To resolve this issue, it will be necessary to delete Aim2 or IL-1 receptors in a cell type-specific manner. Conditional knockout mice are therefore needed to address the contribution of different types of cells to the AIM2-dependent inflammatory response. Methods Animals Il-6−/− (stock number 002650)52, Aim2−/− (stock number 013144), Il1r1−/− (stock number 003245)53 and Myd88−/− (stock number 009088)54 mice in a C57BL/6 genetic background were purchased from Jackson Laboratory and housed in the animal facility of the Institute of Molecular Biology, Academia Sinica, under a 12 h light/12 h dark cycle. All animal experiments were performed with the approval of the Academia Sinica Institutional Animal Care and Utilization Committee and in strict accordance with its guidelines and those of the Council of Agriculture Guidebook for the Care and Use of Laboratory Animals. Chemicals and antibodies We used rabbit polyclonal GFP (Invitrogen), mouse monoclonal MAP2 (Sigma-Aldrich), mouse monoclonal SMI-312R (Covance), mouse monoclonal Myc (Cell Signaling, 9B11), rabbit monoclonal β-tublin (9F3, Cell Signaling), Alexa Fluor 488- and Alexa Fluor 594-conjugated secondary antibodies (Invitrogen), and HRP-conjugated secondary antibodies (GE Healthcare), as well as CL075 (InvivoGen) and poly dAdT (InvivoGen). Poly dAdT is known to activate AIM2 inflammasomes145556. The length of poly dAdT used in this report is 1184 bps. Plasmids The full-length of the Aim2 construct, pUNO1-Aim2, was purchased from InvivoGen and further subcloned into GW1-myc vector. Aim2 miRNA constructs were generated using the BLOCK-iTTM Pol II miR RNA1 Expression Vector Kit (Invitrogen). A plasmid pcDNA 6.2-GW/EmGFP-miR-neg (miR-ctrl, Invitrogen), which was predicted to not target any vertebrate gene, was used as a negative control. The detailed plasmid construction strategy and primer sequences are available in the Supplementary Materials. Cell culture, Transfection, Neuronal Morphometry, Q-PCR and ELISA Cortical and hippocampal neurons were collected from E17.5 mouse embryos and cultured in Neurobasal medium/DMEM (1:1) with B27 supplement, penicillin, streptomycin and glutamine (Invitrogen). For most transfections, the calcium phosphate precipitation method was used as described57. For poly dAdT transfection, we used the Lipofectamine 2000 system (Invitrogen). To investigate neuronal morphology, cells were seeded at a density of 2.5 × 105 cells/well in 12-well plates with 18 mm poly-L-lysine-coated coverslips. For ELISA and Q-PCR, cells were seeded at a density of 2.5 × 106 and 1 × 106 cells/well in poly-L-lysine-coated 6-well plates, respectively. The detailed methods for neuronal morphometry, Q-PCR and ELISA13 are available in the Supplementary Materials. Animals and behavioral analyses The animals used in the behavioral assays were offspring of Aim2 knockout or wild-type C57BL/6J. Male mice were used for behavioral assays to rule out a hormonal effect of female mice on offspring behaviors. All of the procedures for the behavioral analyses have been described previously155859 with minor modification. Behavioral analyses were performed in sequence as follows: week 12, open field; week 13–14, light-dark box; week 14–16, auditory fear conditioning. The animals were moved from the housing room to the behavioral room for accommodation for 1 week prior to the behavioral tests. Before tests, the mice were put into a new cage for 10 min habituation. After each test, the mice were put into another new cage until all mice had finished the test. Therefore, the tested mice did not affect the non-tested mice. All tests were performed from 14:00 to 19:00. The detailed protocols for each behavioral paradigm are available in the Supplementary Materials. Statistical analysis Statistical analyses were performed using GraphPad Prism. For data in Figs 2,3,5 and 7, an unpaired two-tailed t-test was used. For data in Figs 1 and 4, a one-way ANOVA with Bonferroni correction was used. For data in Fig. 6, a two-way ANOVA with variance was used. The data are presented as the mean ± s.d. (Fig. 1) or s.e.m. (remaining figures). For morphological experiments, n indicates the numbers of examined neurons. Additional Information How to cite this article: Wu, P.-J. et al. AIM 2 inflammasomes regulate neuronal morphology and influence anxiety and memory in mice. Sci. Rep. 6, 32405; doi: 10.1038/srep32405 (2016). Supplementary Material Supplementary Information This work was supported by grants from Academia Sinica (AS-103-TP-B05), the Ministry of Science and Technology (MOST 104-2321-B-001-050) and the Simons Foundation (SFARI# 388449) to Y.-P. H. and T.-N.H. was supported by a Postdoctoral Fellowship of Academia Sinica. We thank members of Dr. Yi-Ping Hsueh’s laboratory for relabeling samples for blind experiments and technical assistance, and Dr. John O’Brien for English editing. Author Contributions Conceptualization, P.-J.W., H.-Y.L., T.-N.H. and Y.-P.H.; investigation, P.-J.W.; writing, P.-J.W., H.-Y.L., T.-N.H. and Y.-P.H.; visualization, P.-J.W., Y.-P.H. and T.-N.H.; supervision, project administration and funding acquisition, Y.-P.H. Figure 1 Aim2 is expressed in neurons and is involved in IL-1β secretion. (a–g) Expression of a series of inflammasome-related genes in mouse cortical and hippocampal mixed cultures (CN + HN) at 4 days in vitro (DIV) and in mouse brains, examined by quantitative-PCR using both the Universal ProbeLibrary (a,c,e,f,g) and the SYBR Green (b,d) systems. The ages of examined mice are indicated. In cultured neurons, Cyp was used as internal control. In brains, Gapdh was used as internal control because its expression levels were more constant in brains of different ages. (e) Absolute copy number of Aim2, Nlrp3 and Nlrc4 transcripts in adult brains. (h) Activation of TLR7 and AIM2 induces IL-1β secretion. Cortical and hippocampal mixed cultures were pretreated with CL075 (6 μM) to activate TLR7 for 6 h and then transfected with poly dAdT (0, 2 or 5 μg) for a further 4 h. The levels of secreted IL-1β in cultured media were measured using ELISA. (h) left, WT; right, Aim2−/−. The data are presented as the mean ± s.d. *p < 0.05; **p < 0.01; ***p < 0.001. The experiments were repeated independently three times. Figure 2 Deletion of Aim2 inhibits axonal growth but promotes dendritic growth. (a,b) Cortical and hippocampal mixed cultures prepared from Aim2−/− and WT embryos were transfected with EGFP at 1 DIV and subjected to immunostaining using GFP and SMI-312 R (axonal marker) or MAP2 (dendritic marker) antibodies at different time points as indicted. After identifying neurite property (either axons or dendrites) based on SMI-312R or MAP2 signals, the lengths of axons and dendrites were determined by GFP signals. (a) Dendritic phenotypes. (b) Axonal phenotypes. (c,d) Aim2−/− cortical and hippocampal mixed cultures were transfected with Myc-tagged Aim2, EGFP and control vector, as indicated, at (c) 2 and (d) 1 DIV and harvested for immunostaining at (c) 5 and (d) 3 DIV. The phenotypes of (c) dendrites and (d) axons were quantified as indicated. In (a,b), only GFP images are shown; in (c,d), both GFP and Myc signals are shown. Scale bar, 50 μm. Neurons were collected from two independent experiments. The sample sizes (n) of examined neurons are indicated. The data represent the mean plus s.e.m. *p < 0.05; **p < 0.01; ***p < 0.001. Figure 3 Aim2 deletion alters dendritic arborization in vivo. (a) Representative images of hippocampal CA1 region in vivo revealed by Golgi staining of adult mice. (b,d) Images of individual CA1 neurons. (c,e) Quantitative results of basal and apical dendrites are shown. (b,c) Adult mice; (d,e) P7 mice. Neurons were collected from 5–7 neonatal and adult mice for each genotype. The sample sizes (n) of examined neurons are indicated. The data represent the mean plus s.e.m. *p < 0.05; **p < 0.01; ***p < 0.001. Scale bar, (a) 200 μm; (b,d) 50 μm. Figure 4 Activation of AIM2 inflammasomes inhibits dendritic growth. (a,b) Cortical and hippocampal mixed cultures that were prepared from IL-6−/− mice were first transfected with EGFP at 2 DIV, treated with CL075 (6 μM) for 6 hours, followed by transfection with poly dAdT (PdAdT, 2 μg) at 4 DIV, and finally fixed at 5 DIV. Fixed neurons were immunostained with EGFP and MAP2 antibody to outline cell morphology and to label dendrites. (a) Representative images. (b) Quantitative data of total dendrite length, primary dendrite number and the number of dendrite branch tips. (c,d) Aim2−/− neurons do not respond to poly dAdT to regulate neuronal morphology. Cortical and hippocampal mixed cultures prepared from Aim2−/− embryos were transfected with EGFP at 2 DIV and transfected with poly dAdT (PdAdT, 2 μg) one day later. Neurons were fixed at 4 DIV for immunostaining using GFP and MAP2 (dendritic marker) antibodies as indicated. (c) Representative images (d) Dendritic phenotype. Neurons were collected from three independent experiments. The sample sizes (n) of examined neurons are indicated. The data represent the mean plus s.e.m.. Scale bar, 50 μm. *p < 0.05; ***p < 0.001. Figure 5 Poly dAdT does not influence the neuronal morphology of Il1r1−/− and Myd88−/− neurons. Cortical and hippocampal mixed cultures prepared from Il1r1−/− and Myd88−/− embryos were transfected with EGFP at (a,c) 1 and (b,d) 2 DIV and transfected with poly dAdT (PdAdT, 2 μg) at (a,c) 2 and (b,d) 4 DIV. One day later, neurons were fixed for immunostaining using GFP and SMI-312R (axonal marker) or MAP2 (dendritic marker) antibodies as indicated. (a,c) Axonal phenotype and (b,d) dendritic phenotype were then determined based on GFP signals. Scale bar, 50 μm. The sample sizes (n) of examined neurons are indicated. The data represent the mean plus s.e.m. ns, non-significant. Figure 6 Aim2 knockdown inhibits axonal growth while promoting dendritic growth in WT cultured neurons. (a) Knockdown effect of miR-Aim2 constructs on Myc-tagged Aim2 expression in HEK293T cells. Five miR-Aim2 constructs were cotransfected with Myc-tagged Aim2 into HEK293T cells. Immunoblotting was performed using Myc and β-tubulin antibodies. (b,c) Cultured cortical and hippocampal neurons were transfected with Aim2 knockdown construct miR-Aim2 and non-silencing control miR-Ctrl at (b) 2 DIV and (c) 1 DIV and subjected to immunostaining using GFP and (b) MAP2 or (c) SMI-312R antibodies. The axonal and dendritic lengths were analyzed based on the EGFP signals. Only GFP images are shown. (d) Transfection of miRNA constructs in WT cultured neurons at 2 DIV, followed by transfection with 2 μg poly dAdT (PdAdT) at 4 DIV. Dendrite morphology was analyzed at 5 DIV. Scale bar, 50 μm. The sample sizes (n) of analyzed neurons are indicated. The data collected from three independent experiments are presented as the means plus s.e.m. *p < 0.05; ***p < 0.001. Figure 7 Aim2 deletion reduces locomotor activity, induces anxious behaviors and impairs associative memory. (a) Open field. The total travel distance (locomotion), moving speed, number of total rearing events (exploration), the percentage of time spent at a corner (anxiety), numbers of urine stains and fecal pellets in the arena are shown. (b) Light-dark box. The number of transitions between light and dark boxes and the percentage of time spent in the light or dark was analyzed. (c) Auditory fear conditioning was analyzed. Basal, the freezing response during habituation. AS, freezing response right after stimulation. Memory, the freezing response one day after training. Sample numbers (n) are indicated in each panel. Aim2−/− and WT mice are compared. Data represent mean plus s.e.m. The results of unpaired t tests are shown. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant. ==== Refs Dellacasagrande J. Ligands, cell-based models, and readouts required for Toll-like receptor action . Methods in molecular biology (Clifton, N. J.) 517 , 15 –32 (2009 ).19378025 Czirr E. & Wyss-Coray T. The immunology of neurodegeneration . J Clin Invest 122 , 1156 –1163 (2012 ).22466657 Gottfried C. , Bambini-Junior V. , Francis F. , Riesgo R. & Savino W. The Impact of Neuroimmune Alterations in Autism Spectrum Disorder . Frontiers in psychiatry 6 , 121 (2015 ).26441683 Estes M. L. & McAllister A. K. Immune mediators in the brain and peripheral tissues in autism spectrum disorder . Nat Rev Neurosci 16 , 469 –486 (2015 ).26189694 Stuart M. J. , Singhal G. & Baune B. T. Systematic Review of the Neurobiological Relevance of Chemokines to Psychiatric Disorders . Front Cell Neurosci 9 , 357 (2015 ).26441528 Khandaker G. M. . Inflammation and immunity in schizophrenia: implications for pathophysiology and treatment . The lancet. Psychiatry 2 , 258 –270 (2015 ).26359903 Olson J. K. & Miller S. D. Microglia initiate central nervous system innate and adaptive immune responses through multiple TLRs . J Immunol 173 , 3916 –3924 (2004 ).15356140 van Noort J. M. & Bsibsi M. Toll-like receptors in the CNS: implications for neurodegeneration and repair . Prog Brain Res 175 , 139 –148 (2009 ).19660653 Cunningham C. Microglia and neurodegeneration: The role of systemic inflammation . Glia 61 , 71 –90 (2013 ).22674585 Bsibsi M. , Ravid R. , Gveric D. & van Noort J. M. Broad expression of Toll-like receptors in the human central nervous system . J Neuropathol Exp Neurol 61 , 1013 –1021 (2002 ).12430718 Chen C. Y. , Lin C. W. , Chang C. Y. , Jiang S. T. & Hsueh Y. P. Sarm1, a negative regulator of innate immunity, interacts with syndecan-2 and regulates neuronal morphology . J Cell Biol 193 , 769 –784 (2011 ).21555464 Kaul D. . Expression of Toll-like receptors in the developing brain . Plos One 7 , e37767 (2012 ).22666391 Liu H. Y. . TLR7 negatively regulates dendrite outgrowth through the Myd88-c-Fos-IL-6 pathway . J Neurosci 33 , 11479 –11493 (2013 ).23843519 Adamczak S. E. . Pyroptotic neuronal cell death mediated by the AIM2 inflammasome . J Cereb Blood Flow Metab 34 , 621 –629 (2014 ).24398937 Lin C. W. & Hsueh Y. P. Sarm1, a neuronal inflammatory regulator, controls social interaction, associative memory and cognitive flexibility in mice . Brain Behav Immun 37 , 142 –151 (2014 ).24321214 Liu H. Y. , Huang C. M. , Hung Y. F. & Hsueh Y. P. The microRNAs Let7c and miR21 are recognized by neuronal Toll-like receptor 7 to restrict dendritic growth of neurons . Experimental neurology 269 , 202 –212 (2015 ).25917529 Okun E. , Griffioen K. J. & Mattson M. P. Toll-like receptor signaling in neural plasticity and disease . Trends Neurosci 34 , 269 –281 (2011 ).21419501 Cameron J. S. . Toll-like receptor 3 is a potent negative regulator of axonal growth in mammals . J Neurosci 27 , 13033 –13041 (2007 ).18032677 Lehmann S. M. . An unconventional role for miRNA: let-7 activates Toll-like receptor 7 and causes neurodegeneration . Nat Neurosci 15 , 827 –835 (2012 ).22610069 Lehmann S. M. . Extracellularly Delivered Single-Stranded Viral RNA Causes Neurodegeneration Dependent on TLR7 . J Immunol 189 , 1448 –1458 (2012 ).22745379 Park C. K. . Extracellular microRNAs activate nociceptor neurons to elicit pain via TLR7 and TRPA1 . Neuron 82 , 47 –54 (2014 ).24698267 Hauenstein A. V. , Zhang L. & Wu H. The hierarchical structural architecture of inflammasomes, supramolecular inflammatory machines . Current opinion in structural biology 31 , 75 –83 (2015 ).25881155 Martinon F. , Burns K. & Tschopp J. The inflammasome: a molecular platform triggering activation of inflammatory caspases and processing of proIL-beta . Molecular cell 10 , 417 –426 (2002 ).12191486 Martinon F. & Tschopp J. Inflammatory caspases: linking an intracellular innate immune system to autoinflammatory diseases . Cell 117 , 561 –574 (2004 ).15163405 Ogura Y. , Sutterwala F. S. & Flavell R. A. The inflammasome: first line of the immune response to cell stress . Cell 126 , 659 –662 (2006 ).16923387 Vanaja S. K. , Rathinam V. A. & Fitzgerald K. A. Mechanisms of inflammasome activation: recent advances and novel insights . Trends Cell Biol 25 , 308 –315 (2015 ).25639489 Strowig T. , Henao-Mejia J. , Elinav E. & Flavell R. Inflammasomes in health and disease . Nature 481 , 278 –286 (2012 ).22258606 Hanamsagar R. , Hanke M. L. & Kielian T. Toll-like receptor (TLR) and inflammasome actions in the central nervous system . Trends Immunol 33 , 333 –342 (2012 ).22521509 Halle A. . The NALP3 inflammasome is involved in the innate immune response to amyloid-beta . Nat Immunol 9 , 857 –865 (2008 ).18604209 Heneka M. T. . NLRP3 is activated in Alzheimer’s disease and contributes to pathology in APP/PS1 mice . Nature 493 , 674 –678 (2013 ).23254930 Fann D. Y. . Intravenous immunoglobulin suppresses NLRP1 and NLRP3 inflammasome-mediated neuronal death in ischemic stroke . Cell death & disease 4 , e790 (2013 ).24008734 Cho M. H. . Autophagy in microglia degrades extracellular beta-amyloid fibrils and regulates the NLRP3 inflammasome . Autophagy 10 , 1761 –1775 (2014 ).25126727 Malhotra S. . NLRP3 inflammasome is associated with the response to IFN-beta in patients with multiple sclerosis . Brain 138 , 644 –652 (2015 ).25586466 de Rivero Vaccari J. P. . Therapeutic neutralization of the NLRP1 inflammasome reduces the innate immune response and improves histopathology after traumatic brain injury . J Cereb Blood Flow Metab 29 , 1251 –1261 (2009 ).19401709 Denes A. . AIM2 and NLRC4 inflammasomes contribute with ASC to acute brain injury independently of NLRP3 . Proc Natl Acad Sci USA 112 , 4050 –4055 (2015 ).25775556 Nagyoszi P. . Regulation of NOD-like receptors and inflammasome activation in cerebral endothelial cells . J Neurochem 135 , 551 –564 (2015 ).26083549 Lech M. , Avila-Ferrufino A. , Skuginna V. , Susanti H. E. & Anders H. J. Quantitative expression of RIG-like helicase, NOD-like receptor and inflammasome-related mRNAs in humans and mice . International immunology 22 , 717 –728 (2010 ).20584763 Gilmore J. H. , Fredrik Jarskog L. , Vadlamudi S. & Lauder J. M. Prenatal infection and risk for schizophrenia: IL-1beta, IL-6, and TNFalpha inhibit cortical neuron dendrite development . Neuropsychopharmacology 29 , 1221 –1229 (2004 ).15085088 Cohen P. The TLR and IL-1 signalling network at a glance . J Cell Sci. 127 , 2383 –2390 (2014 ).24829146 Yin Q. , Fu T. M. , Li J. & Wu H. Structural biology of innate immunity . Annual review of immunology 33 , 393 –416 (2015 ). Nolan A. M. , Nolan Y. M. & O’Keeffe G. W. IL-1beta inhibits axonal growth of developing sympathetic neurons . Mol Cell Neurosci 48 , 142 –150 (2011 ).21791245 Gougeon P. Y. . The pro-inflammatory cytokines IL-1beta and TNFalpha are neurotrophic for enteric neurons . J Neurosci 33 , 3339 –3351 (2013 ).23426662 Boato F. . Interleukin-1 beta and neurotrophin-3 synergistically promote neurite growth in vitro . Journal of neuroinflammation 8 , 183 (2011 ).22200088 Cheng P. L. . Self-amplifying autocrine actions of BDNF in axon development . Proc Natl Acad Sci USA 108 , 18430 –18435 (2011 ).22025720 Huang Z. B. & Sheng G. Q. Interleukin-1beta with learning and memory . Neuroscience bulletin 26 , 455 –468 (2010 ).21113196 Pavlowsky A. . A postsynaptic signaling pathway that may account for the cognitive defect due to IL1RAPL1 mutation . Curr Biol 20 , 103 –115 (2010 ).20096586 Jin H. , Gardner R. J. , Viswesvaraiah R. , Muntoni F. & Roberts R. G. Two novel members of the interleukin-1 receptor gene family, one deleted in Xp22.1-Xp21.3 mental retardation . Eur J Hum Genet 8 , 87 –94 (2000 ).10757639 Yasumura M. . IL1RAPL1 knockout mice show spine density decrease, learning deficiency, hyperactivity and reduced anxiety-like behaviours . Scientific reports 4 , 6613 (2014 ).25312502 Murray C. L. , Obiang P. , Bannerman D. & Cunningham C. Endogenous IL-1 in cognitive function and anxiety: a study in IL-1RI−/− mice . Plos One 8 , e78385 (2013 ).24205219 Andre R. , Lerouet D. , Kimber I. , Pinteaux E. & Rothwell N. J. Regulation of expression of the novel IL-1 receptor family members in the mouse brain . J Neurochem 95 , 324 –330 (2005 ).16086690 Smith D. E. . A central nervous system-restricted isoform of the interleukin-1 receptor accessory protein modulates neuronal responses to interleukin-1 . Immunity 30 , 817 –831 (2009 ).19481478 Kopf M. . Impaired immune and acute-phase responses in interleukin-6-deficient mice . Nature 368 , 339 –342 (1994 ).8127368 Glaccum M. B. . Phenotypic and functional characterization of mice that lack the type I receptor for IL-1 . J Immunol 159 , 3364 –3371 (1997 ).9317135 Hou B. , Reizis B. & DeFranco A. L. Toll-like receptors activate innate and adaptive immunity by using dendritic cell-intrinsic and -extrinsic mechanisms . Immunity 29 , 272 –282 (2008 ).18656388 Pierini R. . AIM2/ASC triggers caspase-8-dependent apoptosis in Francisella-infected caspase-1-deficient macrophages . Cell Death Differ 19 , 1709 –1721 (2012 ).22555457 Zhao C. , Gillette D. D. , Li X. , Zhang Z. & Wen H. Nuclear factor E2-related factor-2 (Nrf2) is required for NLRP3 and AIM2 inflammasome activation . J Biol Chem 289 , 17020 –17029 (2014 ).24798340 Lin Y. L. , Lei Y. T. , Hong C. J. & Hsueh Y. P. Syndecan-2 induces filopodia and dendritic spine formation via the neurofibromin-PKA-Ena/VASP pathway . J Cell Biol 177 , 829 –841 (2007 ).17548511 Chung W. C. , Huang T. N. & Hsueh Y. P. Targeted Deletion of CASK-Interacting Nucleosome Assembly Protein Causes Higher Locomotor and Exploratory Activities . NeuroSignals 19 , 128 –141 (2011 ).21576927 Huang T. N. . Tbr1 haploinsufficiency impairs amygdalar axonal projections and results in cognitive abnormality . Nat Neurosci 17 , 240 –247 (2014 ).24441682

PMC005xxxxxx/PMC5000014.txt

==== Front Sci RepSci RepScientific Reports2045-2322Nature Publishing Group srep3231910.1038/srep32319ArticleSonochemical Assisted Solvothermal Synthesis of Gallium Oxynitride Nanosheets and their Solar-Driven Photoelectrochemical Water-Splitting Applications Iqbal Naseer 12Khan Ibrahim 13Yamani Zain H. 14http://orcid.org/0000-0002-6031-9385Qurashi Ahsanullhaq a131 Center of Research Excellence in Nanotechnology, King Fahd University of Petroleum and Minerals, Dhahran, 31261, Saudi Arabia2 Department of Biosciences, COMSATS Institute of Information Technology, Park Road, Chak Shahzad, Islamabad, 45550, Pakistan3 Department of Chemistry, King Fahd University of Petroleum and Minerals, Dhahran, 31261, Saudi Arabia4 Department of Physics, King Fahd University of Petroleum and Minerals, Dhahran, 31261, Saudi Arabiaa ahsanulhaq@kfupm.edu.sa26 08 2016 2016 6 3231926 04 2016 02 08 2016 Copyright © 2016, The Author(s)2016The Author(s)This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/Gallium oxynitride (GaON) nanosheets for photoelectrochemical (PEC) analysis are synthesized via direct solvothermal approach. Their FE-SEM revealed nanosheets morphology of GaON prepared at a reaction time of 24 hours at 180 °C. The elemental composition and mapping of Ga, O and N are carried out through electron dispersive X-ray spectroscopy (EDX). The cubic structure of GaON nanosheets is elucidated by X-ray diffraction (XRD)analysis. The X-ray Photoelectron Spectroscopy (XPS) further confirms Ga, O and N in their respective ratios and states. The optical properties of GaON nanosheets are evaluated via UV-Visible, Photoluminescence (PL) and Raman spectroscopy’s. The band gap energy of ~1.9 eV is calculated from both absorption and diffused reflectance spectroscopy’s which showed stronger p-d repulsions in the Ga (3d) and N (2p) orbitals. This effect and chemical nitridation caused upward shift of valence band and band gap reduction. The GaON nanosheets are investigated for PEC studies in a standard three electrode system under 1 Sun irradiation in 0.5 M Na2SO4. The photocurrent generation, oxidation and reduction reactions during the measurements are observed by Chronoampereometry, linear sweep Voltametry (LSV) and Cyclic Voltametry (CV) respectively. Henceforward, these GaON nanosheets can be used as potential photocatalyts for solar water splitting. ==== Body Gallium compounds12345 including gallium nitride (GaN), gallium oxide (Ga2O3) and more recently gallium oxynitrides (GaON/Ga3O3N/GaxOyNz) have been known for their semiconducting and optoelectronic properties6789. These gallium materials in different morphologies have enlightened ample of applications e.g. High electron mobility transistors (HEMTs)10, UV-Blue light emitting diode and laser diodes11, logic gates1213, field effect transistors (FET)14 and gas sensing devices15. Despite of their diverse applications, very limited gallium oxynitride (GaON) compounds have been reported for solar water splitting studies1617. Synthesis of GaON can be of significant interest to understand its intriguing properties. Formation of a direct band gap spinal structured gallium oxynitride (Ga3O3N) material18 has been reported within a gallium oxide and gallium nitride system with potential electronic properties1819. Similarly, cubic gallium oxynitride crystal phase material of composition Ga2.8O3.5N0.5 is also carried out in a metastable state during the formation of GaN thin films by a chemical approach2021. Literature revealed that GaON can be prepared from both GaN and Ga2O3 by oxidation and nitridation/ammonolysis respectively. However, ammonolysis of Ga2O3 is favored as compared to the oxidation of GaN, because the subsequent method revokes the original structure via homogeneous formation of GaN22. Though, there are many factors that remained unclear in formation of GaON from ammonolysis route where a complete Ga2O3 layer growth over GaN crystal was observed22.Hu et al.17 presented a strategy in which they considered Ga(OH)3 as more appropriate precursor for GaON synthesisth an Ga2O3 and developed a visible-light active wurtzite-like gallium oxynitride (GaON) photocatalyst by nitridation of crystalline Ga(OH)3 with NH3 at elevated temperatures (550 and 900 °C). Their experimentation revealed the utilization of unoccupied 12-coordinate sites in crystal lattice of Ga(OH)3 that expedited ionic transportation during the nitridation. As incorporation of nitrogen is particularly helpful in reducing band gap of photocatalyst16 hence the GaON prepared showed band gap energies in the range of 2.2 eV to 2.8 eV. In addition, GaON showed superior PEC properties for production of H2 and O2 gases from methanol and silver nitrate (AgNO3) solutions respectively under visible light irradiation. Since photocatalyts based solar water splitting for H2 and O2 production requires a suitable material with considerable band gap energy that could facilitate the electron mobility in an effective way16. Generally nitrides or oxynitride photocatalytic materials involving d0 or d10 electronic configuration with band gap energy below 3 eV are considered efficient for solar water splitting2324. In case of d10 photocatalytic materials, the occurrence of conduction band of semiconductor is considered by hybridization of sp orbitals with large band dispersion. Thus observed high electron mobility and enhanced photoelectrochemical characteristics1725. Group III nitrides such as GaN with band gap of 3.4 eV and wurtzite morphology showed limited photocatalytic water splitting. Its wide band gap can be tuned by doping metallic (M) nuclei (Zn, Ru) or preparing its oxynitride1726. In such photocatalyts, the existence of M3d and N2p orbitals in the upper valence band provide p-d repulsions as a result upward switch of valence band ensued band gap reduction. Besides this the large dispersion of hybridized M3d, N2p, and O2p orbitals in doped nitrides or oxynitrides possibly augmented photogenerated hole mobility in valence band and thus further promoted photocatalytic activity17262728. Herein, our research focuses on Gallium oxynitride (GaON),fabricated from gallium metal and organic diamine in a single-step at 180 °C and at different reaction times following sonochemical and hydrothermal approaches729. The surface morphology under FE-SEM showed formation of uniform nanosheets containing nanopores and defects in the structure obtained at a reaction time of 24 hours. The as prepared GaON nanosheets at 180 °C for 24 hours is characterized by EDX, XRD and XPS for elemental and structural insights. Furthermore, its optical properties are examined by FTIR, Raman UV-Visible spectroscopy’s and Photoluminescence studies. In addition to aforementioned characterization of as prepared GaON nanosheets at 180 °C for 24 hours, we carried out FE-SEM, XRD studies of GaON prepared at 3, 6 and 12 hours reaction time. However, based on our experimental observations we conducted photoelectrochemical studies of GaON prepared at 180 °C for 12 hours, 24 hours as well as GaON prepared at 180 °C for 24 hours and annealed at 500 °C respectively. Thin films of GaON samples are deposited over FTO substrates and tested by employing standard three electrode system for photoelectrochemical studies under chopped solar (1 Sun) irradiation source. Results and Discussions The FE-SEM analysis at first explored nanosheets like morphology of our proposed GaON material through a solvothermal reaction at slightly better temperature of 180 °C for 24 hours. This strategy is straightforward and facile in comparison to high temperature synthesis that are reported elsewhere19222630. To our best of knowledge, this kind of morphology with one-step synthetic strategy and a single GaON phase photocatalyts for solar water splitting is not reported. We carried out FE-SEM studies as shown in Fig. 1, the FE-SEM micrographs presented a uniform growth of variable size and width of GaON nanosheets. Moreover, we performed FE-SEM analysis for as prepared GaON at different reaction times (3, 6 & 12 hours @ 180 °C) as well as for GaON prepared at 24 hours and annealed at 500 °C for 4 hours, the data is presented in Fig. S1 of Supplementary information. The Fig. 1(a,b), shows lower magnification micrographs of GaON nanosheets, a clear arrangement of nanosheets of variable size is explicable. Whereas Fig. 1(c,d) further elucidate high resolution micrographs of GaON nanosheets. It is evident from Fig. 1(d) that the thickness of the nanosheets lies between 15–30 nm whilst the length is variable we could assume between 400–1000 nm. Furthermore, the petals like flower arrangement of nanosheets has made its morphology inimitable with interfacial distance of 100–500 nm. Moreover, the nanosheets surface also showed nanoporesas well as ordered defects that are supposed to show some shifts in XRD patterns in addition to better photoelectrochemical properties. This morphology is attributed to the in-situ incorporation of N and O in the chemical reaction of Ga-ethylenediamine-water complex through a hydrothermal reaction at 180 °C. In order to confirm the elemental existence of each component and our assumed product, energy dispersive spectroscopy (EDX) is performed. The Fig. 2 shows elemental mapping/distribution encompassing composition of each Gallium, Oxygen and Nitrogen entities in the nanosheets and EDX spectrum respectively. Their weigh and atomic ratios are well justifiable for the single phase as prepared GaON nanosheets, however the appearance of gold in the Fig. 2(e) spectrum is because of gold coating used for FE-SEM and EDX analysis. The powder XRD analysis enlightened diffraction patterns and crystalline phase of as prepared GaON nanosheets at 180 °C for 24 hours. Figure 3 shows, characteristic peaks that could be indexed partially and assigned a cubic crystal structure with space group Fd-3m(227) and composition of Ga2.79O3.05N0.76 via JSPDF card number 01-078-8634 from XRD data base and literature31. Furthermore, the XRD patterns of GaON prepared at 180 °C for 3, 6 and 12 hours as well as GaON nanosheets prepared at 180 °C for 24 hours and annealed at 500 °C for 4 hours are compared in Fig. S2 of Supplementary information. In Fig. 3, the insights to XRD spectrum exposed highest intensity peaks corresponding to cubic GaON nanosheets in their most probable forms which are recorded at 2 theta (θ) range of 27°, 34°, 37°, 42° and 54° with indices 220, 311, 222, 400 and 422 respectively. However the median to lower range XRD patterns observed at 2 theta (θ) values i.e., 17°,48°, 58°, 63°,65° and 67° with indices 111, 331, 511, 440, 531 and 442 also substantiate with the cubic crystal structure XRD patterns of Ga2.79O3.05N0.76 via JSPDF card number 01-078-8634. The peak shifting and appearance of new patterns(red star pointed in Fig. 3) observed are assumed due to formation of single phase GaON. In addition, some peaks in XRD spectra (tagged with green circle and black square) are also compared with GaN and Ga2O3 XRD patterns in the data base. However their peaks showed low intensities and slight shifting in the XRD region therefore no crystallization of the two stable phases GaN and Ga2O3 could be observed. These differences in the lattice parameters are obvious depending upon the experimental conditions commenced since cubic gallium oxynitride has been known to exist in a metastable state during oxidation or ammonolysis of GaN and Ga2O3 respectively6719223032. Consequently, all such broad reflections in XRD are indicative of crystallinity as well as somewhat ordered defects in composition of as prepared GaON nanosheetswhich are also depicted in FE-SEM and EDX studies. To further elucidate the structure of GaON nanosheets, XPS measurements are carried out, all the peaks observed are fitted according to the standard methods17. Figure 4(a–d) show XPS binding energy profiles of various chemical states of Ga, O and N. A deeper insight to Fig. 4(a) gave clue about two peaks for Ga 3 d at binding energies 19.77 eV and 21.87 eV respectively that also validated with the literature values for Ga compounds17333435. The shifting of these peaks to higher values as well as appearance of d-p repulsions is obviously due to incorporation of nitrogen/oxygen in the gallium metallic center from the reaction precursors. Hu et al. reported this in their studies17, furthermore, they explained strong hybridization between the valence orbitals of Ga, N and O atoms that can be observed in binding energy range of 0–14 eV and assigned to hybrid Ga4p-N2p, Ga4s-N2p, and Ga4s-O2p chemical states363738. These hybrid states are also observed in our as prepared GaON nanosheets at 24 hours and are presented in Fig. S3 along with XPS survey spectra in Supplementary information. The occurrence of this phenomenon as evident from XPS spectra also strengthen the efficacy of our straightforward synthetic approach that incorporated nitrogen and oxygen simultaneously during reaction between Gallium-ethylenediamine-water at reaction temperature of 180 °C. Figure 4(b) provides extended information of fitted peaks recorded at binding energy range 397.86 eV (main peak) and 394.52 eV (shoulder peak) which are assigned to Nitrogen (1s). This assignment is straight forward for existence of Ga-N bond as reported for various nitrogen containing Ga materials39404142. Similarly, the Fig. 4(c) shows high resolution spectra of fitted peaks near binding energies 531.86–534.44 eV assigned for core level O1s4344454647. The intense peak for O1s is observed at 531.86 eV whereas the shoulder peak appears at 534.44 eV.Literature4849 revealed that the main O1s peak in Gallium oxide occurs at a binding energy of ~531 eV and supports the formation of Ga-O bonding with the highest oxidation state of Ga is (Ga3+). Figure 4(d) shows the doublet peaks observed at binding energies of 1120.11–1122.84 eV and 1146.51–1149.79 eV that are characteristics for Gallium (Ga) and thus assigned as Ga 2P3/2 and Ga 2P1/2 respectively50. These binding energies are different from Ga 2p levels in metallic Gallium i.e., 1117.0 and 1144.0 eV for Ga 2p3/2 and Ga 2p1/2, respectively51. Nonetheless, this reflects a positive shift in our as prepared GaON material which is attributed to redistribution of electronic charge around the reacting atoms5253. Therefore, the difference in Ga chemical bonding causes a binding energy shift that can be used to extract chemical states information. This further enlighten oxidation state of Ga metal in the growing GaON nanosheets in its highest valence state (Ga3+)5054. The XPS studies thus comply the formation of GaON nanosheets via our straightforward and facile approach at 180 °C for 24 hours that can be used as photoactive material for enhanced water splitting performance. The absorbance and diffused reflectance studies are performed by UV-VIS spectroscopy in order to estimate band gap energy of as prepared GaON nanosheets synthesized from a solvothermal route. The powder sample is used for absorbance and diffused reflectance analysis. Figure 5(a) shows the absorbance curve, the spectrum shows a strong cut-off wave length observed at 650 nm where comparatively low absorbance is observed. This wavelength is used to calculate the band gap energy of as prepared GaON nanosheets according to the methods reported elsewhere5556. The band energy is calculated around ~1.9 eV, further to confirm this band gap energy we evaluated the diffused reflectance spectroscopy as shown in Fig. 5(b) for band gap calculation. Thus the band gap estimated from this also comparable with that observed from absorbance curve. The reduction of band gap energy is attributed to incorporation of nitrogen and the existence of Ga3d and N2p orbitals in the upper valence band of GaON that encouraged more p-d repulsions hence caused an upward shift in its valence band161726. In order to elucidate the structural composition of the cubic GaON nanosheets we investigated it for Raman, FTIR and Photo luminescence studies (FT-IR and PL are discussed in Supplementary information Fig. S4 and Fig. S5 respectively). The Raman spectra of GaON is shown in Fig. 6. The gallium oxynitride cubic phase showed broad Raman shifts that symbolizes significant disorder among O and N atoms in the cubic structure and/or the presence of cation (Ga3+) or anion vacancies7 during the solvothermal synthesis of GaON. The maxima of these good intensity broad bands are observed around 250 cm−1, 425 cm−1, 600 cm−1 730 cm−1and 860 cm−1. However, some narrow bands of medium to lower intensity are also observed at 265 cm−1, 520 cm−1and 540 cm−1 respectively. These fingerprints phonon frequencies are comparatively in good agreement with those reported in literature for Ga3O3N67. Soignard et al.6 carried out theoretical calculations by first principles local density approximation (LDA) methods for a Ga3O3N pseudocubic R3hm model structure. They proposed nine Raman-active modes (4A1g + 5Eg), with zone-center frequencies calculated at 213, 219, 367, 379, 499, 512, 634, 647, and 782 cm−1 respectively. However they concluded that the zone-center modes acted as poles for broadened spectra with complete vibrational density of states (vDOS). Hence, their expected mode frequencies for pseudocubic Ga3O3N were grouped into five phonon frequencies i.e., 216, 373, 506, 640, and 782 cm−1. Thus for an ideal spinal structure gallium oxynitride one should expect these phonon frequencies with slight acceptable variations depending upon the nature of chemical reaction undertaken and morphology achieved. This perception is further supported by Oberlaender et al.7 for their cubic spinal-type Ga3O3N material. They described five active Raman modes i.e., ГRaman = 3T2g + 1Eg + 1 A1g and assigned their phonon frequencies for the respective Raman bands. The broadness of Raman bands is attributed to the high disorder of nitrogen and oxygen atoms in the anionic cites during growth of GaON as described earlier i.e., above 8% of nitrogen in our as synthesized GaON. In light of these experimental and theoretical studies the high intensity Raman bands in our as synthesized GaON nanosheets are assigned in the following order: Phonon frequency/modes, 250 cm−1(T2g), 425 cm−1(Eg), 600 cm−1 (T2g), 730 cm−1 (T2g) and, 860 cm−1 (A1g). The FTO coated GaON nanosheets are used for probing photoelectrochemical performance in a standard three electrode setup i.e., a reference standard calomel electrode (SCE), platinum counter electrode and GaON coated FTO photoanode as working electrode respectively. The PEC cell containing FTO/GaON photoanode and other respective electrodes are dipped in 0.5 M Na2SO4 electrolyte solution at neutral pH (7) and are exposed to illumination source (1 Sun) in an on/off fashion with regular intervals of time to investigate the photoresponse of the GaON nanosheets. Figure 7 (a) shows the Jp – t profile where photocurrent generated is plotted as function of time. The current produced from this device is in the range of nanoamperes (140 nA–80 nA at OCP ~ 0 V = applied potential). Whereas this photocurrent density substantially increased to ~240 μA at applied potential of −1.2 V as shown in Fig. 7(c). Besides this, we observed enhanced photocurrent density in mA range i.e., ~3.75 mA at higher potential values recorded from GaON nanosheets in linear sweep voltametry (LSV) experiments as presented in Fig. 8. In chronoampereometry, the first Jp (photocurrent density) spike is recorded to a higher value upon GaON exposure to 1 Sun irradiation but later on relaxed to a stable plateau that explicit the significant stability of the GaON film with the passage of time as can be seen in Fig. 7(a), in onset Fig. 7(b,c) respectively. The initial decline in the photocurrent response is observed in course of instituting equilibrium between FTO coated electrode layer and the electrolyte solution upon instant exposure to solar irradiation source. Furthermore, these chronoampereometric measurements carried out at different applied potentials (0 V and −1.2 V) for longer time span showed significant stability for several minutes, as also presented in Supplementary information Fig. S6. Inprogression of PEC measurements, shifting of the Jp − t profile to its normal baseline under dark (no illumination) exhibited a reversible responsehence acclaimed that photocurrent generation is solely due to solar driven water splitting reaction. Nonetheless, the structural morphology, reduced band gap and a straight forward synthetic approach are advantageous features of as prepared GaON nanosheets that corroborate to better PEC characteristics. The Fig. 8 shows linear sweep voltammetry (LSV) curves of the as prepared GaON at 180 °C for 12 hours, 24 hours and GaON annealed sample in a standard three electrode configuration using 0.5 M Na2SO4 as electrolyte. Figure 8(a) shows comparative LSV profiles of bare FTO and FTO coated with different GaON samples tested at applied potential range versus SCE under 1 Sun simulated solar light. The as prepared GaON nanosheets at 180 °C for 24 hours sample showed highest photocurrent of magnitude 3.75 mA whereas its annealed GaON nanosheets produced more than 2 fold lower photocurrent i.e., 1.25 mA. Likewise, GaON sample obtained at 180 °C for 12 hours generated photocurrent in ~500 μA range under the applied potential, this current density is more than three times lower as compared to current density observed from as prepared GaON nanosheets at 180 °C for 24 hours. The most probable reasons of this decreased photocurrent recorded from GaON obtained at 12 hours and GaON annealed samples are attributed to in complete nanosheets growth and deformation of nanosheets morphology which resulted in thickening and distortion of the nanoflakes hence caused reduced electron mobility and charge separation respectively. An additional LSV measurement is presented in Fig. 8(b) onset. The as prepared GaON nanosheets at 180 °C for 24 hours showed highest photocurrent in comparison to competing GaON samples when the applied potential is in the range of −0.4 to −1.3 V. While further insights show that a photocurrent of ~250 μA is observed at applied potential −1.2 V which elucidate solar water splitting and H2 production by GaON under the same experimental conditions as discussed in chronoampereometric studies earlier and presented in Fig. 7(c). In contrast the LSV curves obtained for GaON 12 hours and GaON annealed samples showed significantly decreased photoresponse. To further support our preceding findings for as prepared GaON nanosheets at 180 °C for 24 hours sample we also carried out its cyclic voltammetry (CV) (details provided in Supplementary information Fig. S7). The reduction potential observed in CV curves is between −1.1 V to −1.2 V and photocurrent of ~280 μA is also recorded. This indicated that reduction reaction i.e., hydrogen generation is favorable by Gallium oxynitride under solar driven photocatalytic water splitting. Thus in all PEC experiments, the enhanced photocurrent generation from as prepared GaON nanosheets at 180 °C for 24 hours sample is assumed due to nanoporous texture and appearance of defects on the surfaces of GaON nanosheets. Furthermore, the in-situ embedding of nitrogen and oxygen to Gallium metal during its solvothermal synthesis not only causes defects on Gallium sites but also leave substantial O/N disorder. This in turn helps in better electron mobility and minimal electron hole recombination.The schematic for electron hole pair generation, charge separation mechanisms and hydrogen production under solar light with perspective future green energy fuel are shown in Fig. 9. In summary, the as prepared GaON nanosheets by sonochemically assisted solvothermal process at comparatively low temperature showed significantly enhanced photoelectrochemical properties. The XRD, XPS and Raman spectroscopy’s revealed the cubic crystal structure in which gallium exists in it highest chemical state of Ga (3+) and showed further interaction with N and O in a very disordered way thus cause nanopores and defects on the surface. This ultimately helped in better optical and photoelectrochemical properties by providing better electron mobility and minimal electron hole recombination. The facile synthetic route and the nanosheets like morphology are the advantageous features of GaON which also contributed in enhanced physicochemical properties. Furthermore, the appreciable photoelectrochemical characteristics such as, generation of photocurrent, reversible response, stability over certain period of time etc. are the key parameters of this single phase GaON nanosheets material that make it significantly better in the series of gallium compounds used for solar water splitting. Nevertheless, the photoelectrochemical properties of such GaON materials can be further enhanced by tuning its material’s chemistry such as by doping different other metals with GaON via physical methods or incorporation of certain co-catalyst in GaON nanosheets by chemical process or by adopting an electrochemical method to develop its well-aligned hetero nanostructures. Methods Materials Synthesis All the chemicals and reagents i.e., Gallium metal, ethylene diamine, sodium sulphate (Na2SO4), ethanol and acetone were purchased from Sigma Aldrich and used as received unless otherwise stated. For synthesis, few grams of Gallium metal are mixed with ethylenediamine and placed in an ultrasonic bath for an hour at 75 °C. During the ultrasonification, deionized water is added in 5 ml portions after every 5 min. time span. The appearance of black suspension showed the formation of Ga- ethylenediamine complex. After complete dissolution of gallium metal in ethylenediamine water solution, the reaction mixture is shifted into a stainless steel autoclave containing a Teflon vessel. The solvothermal reaction is carried out for 3, 6, 12 and 24 hours at 180 °C respectively. Afterwards the reaction mixture containing precipitates of Galliumoxynitride (GaON) in each case is centrifuged at 4000 RPM for 5 mins washed with ethanol and acetone respectively before drying in a vacuum oven at 100 °C for two hours. Furthermore, the as prepared GaON nanosheets at 180 °C for 24 hours is further annealed at 500 °C for 4 hours owing to its best morphology achieved for comparative PEC studies. Fabrication of FTO/GaON nanosheets based device and photoelectrochemical setup Photoelectrochemical characteristics of the GaON nanosheets are investigated using FTO conducting glass substrates. Initially, FTO glass substrates are washed by acetone (10 min) and water (10 min) respectively with continuous ultrasonification. In the next step slurry of GaON nanosheets samples (taken from GaON obtained at 12 hours, 24 hours reaction times and 24 hours but annealed at 500 °C respectively) with ethanol are prepared and drop coated over the pre-treated FTO glass substrates to produce a smooth film. The FTO/GaON substrates are heated at 100 °C for 2 hours in order to evaporate the solvent and harden the GaON nanosheets layers over to FTO substrate that could withstand during the PEC measurements. The photoelectrochemical measurements are carried out by a three electrode system in 0.5 M Na2 SO4 (pH = 7) solution, where a Pt wire served as auxiliary electrode, FTO coated with different GaON substrate served as photoanode or working electrode and the standard calomel Ag/AgCl reference electrode (SCE). All the photoelectrochemical experiments are performed through Metrohm Autolab Potentiostat (PGSTAT302N) instrument. However, for Solar light in lab we used Oriel sol 3A class AAA solar simulator-Newport with following specifications i.e., power 100 mW.cm−2(1 Sun), IEC/JIS/ASTM certified containing 450 Watt Xenon lamp, Air Mass 1.5G Filter, UV cut off filter and 2 × 2 inch aperture for output beam. The data obtained from photoelectrochemical measurements is discussed according to the standard calomel electrode (SCE). Characterization of Gallium Oxynitride Nanostructures The purity and crystalline phases of as synthesized GaON nanomaterials are characterized by X-ray diffraction (XRD) technique using a Benchtop MiniFlex -X-ray Diffraction instrument (Mini-XRD) from Rigaku (40 kV and 15 mA) in the range of 10–70° (2θ) at a scanning rate of 3° min-1 with CuK alpha radiation and wave length of 1.54060. The structural composition and elemental states involved in GaON formation are also verified by x-ray photoelectron spectroscopy (XPS) using PHI 5000 VersaProbe II spectrometer (UlVAC-PHI), employing Al Kα as the incident radiation source. The C1s (E = 284.5 eV) level was served as the internal standard. The morphology of the GaON nanomaterial is observed under TESCAN Lyra 3 Field Emission Dual Beam (Electron/Focused Ion Beam) system combined high‐end field‐emission scanning electron microscope (FE-SEM) also facilitated with EDX for elemental determination/mapping. Optical properties like absorption and diffused reflectance spectroscopy of GaON nanosheets are carried out using Agilent Cary 5000 high performance UV-Vis-NIR spectrophotometer containing praying mantis accessory with alignment tools and powder cell sample cups. In situ Raman spectroscopy of GaON is performed with laser (300 mW Green 532 nm) by iHR320 Horiba imaging spectrometer. The vibrational modes in the GaON are recorded by FT-IR 6700 Nicolet™ FT-IR spectrometer from Thermo Electron Corporation. The material is also tested for photoluminescence studies carried out with fluorolog-3 Imaging Spectrophotometer at an excitation wavelength of 350 nm and slit width of 2 nm. Additional Information How to cite this article: Iqbal, N. et al. Sonochemical Assisted Solvothermal Synthesis of Gallium Oxynitride Nanosheets and their Solar-Driven Photoelectrochemical Water-Splitting Applications. Sci. Rep. 6, 32319; doi: 10.1038/srep32319 (2016). Supplementary Material Supplementary Information This project was funded by the National Plan for Science, Technology and Innovation (MAARIFAH) – King Abdulaziz City for Science and Technology –through the Science and Technology unit at King Fahd University of Petroleum and Minerals (KFUPM) – the Kingdom of Saudi Arabia, award number (13-NAN1600-04). Author Contributions A.Q., N.I. and Z.H.Y. designed the experimental strategy. N.I. conducted experiments and A.Q. also supervised the experiments. N.I. and I.K. characterized the materials, collected and analyzed the data, N.I. wrote the paper. A.Q. gave suggestions to revise the manuscript. Finally, all authors analyzed data, discussed the results, and reviewed the manuscript. All the authors have equal contribution in this work. Figure 1 SEM images of GaON nanosheets (a,b) low resolution micrographs at 2μm and 1μm showing growth patterns of nanosheets (c,d) higher magnification micrographs show internal texture and defects attributing towards availability of high surface area. Figure 2 EDX elemental composition and mapping of the GaON nanosheets. Where (a) showing Fe-SEM micrograph undertaken for elemental composition analysis (b–d) presenting elemental maps of Ga, O and N respectively. (e) Expressing the spectrum investigated for atomic composition, the onset in (e) also shows their corresponding weight and atomic percentages. Figure 3 XRD patterns of as prepared GaON nanosheets prepared via a solvothermal route @ 180 °C for 24 hours. Where (*), (■), () represents new XRD peaks, patterns of XRD compared with GaN and Ga2O3 respectively. Figure 4 XPS photoelectron spectroscopy of as synthesized GaON nanosheets at 180 °C @ 24 hours presenting binding energy profiles for (a) Ga 3d (b) N1s (c) O1s and (d) Ga2p3/2 & Ga 2p1/2. Figure 5 UV-Vis spectroscopic studies of as prepared GaON nanosheets at 180 °C @ 24 hours in powder form (a) UV/Vis absorption curve (b) diffused reflectance spectrum for band gap energy calculation. Figure 6 Raman spectroscopy of as prepared single phase GaON nanosheets at 180 °C @ 24 hours through a solvothermal approach. Figure 7 Chronoampereometry showing photocurrent densities for as prepared GaON nanosheets @ 180 °C for 24 hours. (a) Photocurrent recorded at 0 V applied potential (b) onset shows photocurrent stability for longer period of time and (c) photoresponse recorded at −1.2 V applied potential. All measurements are observed under chopped on/off light from a solar simulator of AM 1.5G filter and 100 mW cm−2(1Sun) light intensity of Xe lamp. Figure 8 Linear sweep voltammograms of GaON samples coated on FTO glass under simulated solar light (1 Sun) with scan rate of 100 mV/sec in a standard three electrode system, SCE, Pt-cathode and GaON coated FTOs as anode, (a) shows comparative photocurrent densities of GaON prepared at different reaction conditions, photocurrent density in mA range is recorded against applied potential (−0.4 to −2.0) whereas the onset (b) is presenting μA range photocurrent density, when the applied potential is in the range of −0.4 to −1.3 V respectively. Figure 9 A schematic illustration of photo-electrochemical solar-driven hydrogen production device via water splitting. ==== Refs Masubuchi Y. . Nanowire of hexagonal gallium oxynitride: Direct observation of its stacking disorder and its long nanowire growth . Journal of the European Ceramic Society 32 , 1989 –1993 , 10.1016/j.jeurceramsoc.2011.10.038 (2012 ). Figueiredo J. F. D. . Zinc-gallium oxynitride powders: effect of the oxide precursor synthesis route . Cerâmica 59 , 269 –276 (2013 ). Kamler G. , Weisbrod G. & Podsiadło S. Formation and Thermal Decomposition of Gallium Oxynitride Compounds . Journal of Thermal Analysis and Calorimetry 61 , 873 –877 , 10.1023/a:1010142424089 . Li C. . Positive onset potential and stability of Cu2O-based photocathodes in water splitting by atomic layer deposition of a Ga2O3 buffer layer . Energy & Environmental Science 8 , 1493 –1500 , 10.1039/C5EE00250H (2015 ). Ganesh V. , Suresh S. , Balaji M. & Baskar K. Synthesis and characterization of nanocrystalline gallium nitride by nitridation of Ga-EDTA complex . Journal of Alloys and Compounds 498 , 52 –56 , 10.1016/j.jallcom.2010.03.068 (2010 ). Soignard E. . Spinel-Structured Gallium Oxynitride (Ga3O3N) Synthesis and Characterization:  An Experimental and Theoretical Study . Chemistry of Materials 17 , 5465 –5472 , 10.1021/cm051224p (2005 ). Oberländer A. , Kinski I. , Zhu W. , Pezzotti G. & Michaelis A. Structure and optical properties of cubic gallium oxynitride synthesized by solvothermal route . Journal of Solid State Chemistry 200 , 221 –226 , 10.1016/j.jssc.2013.01.040 (2013 ). Oshima C. . Photocatalytic activity of ZnO/GaP1-xNx for water splitting . Journal of Materials Chemistry A 3 , 18083 –18089 , 10.1039/C5TA04732C (2015 ). Qin L. , Xue C. , Duan Y. & Shi L. Synthesis and Characterization of Glomerate GaN Nanowires . Nanoscale Research Letters 4 , 584 –587 , 10.1007/s11671-009-9285-y (2009 ).20596317 Karmalkar S. , Shur M. S. & Gaska R. 173–213 (World Scientific, 2003). Nakamura S. , Mukai T. & Senoh M. Candela‐class high‐brightness InGaN/AlGaN double‐heterostructure blue‐light‐emitting diodes . Applied Physics Letters 64 , 1687 –1689 , 10.1063/1.111832 (1994 ). Johnson J. C. . Single gallium nitride nanowire lasers . Nat Mater 1 , 106 –110 (2002 ).12618824 Huang Y. , Duan X. , Cui Y. & Lieber C. M. Gallium Nitride Nanowire Nanodevices . Nano Letters 2 , 101 –104 , 10.1021/nl015667d (2002 ). Chen Q. , Yang J. W. , Kahn M. A. , Ping A. T. & Adesida I. High transconductance AlGaN/GaN heterostructure field effect transistors on SiC substrates . Electronics Letters 33 , 1413 –1415 (1997 ). Kerlau M. . Detection of propane by “GaON” thick-film gas sensors . Sensors and Actuators B: Chemical 122 , 14 –19 , 10.1016/j.snb.2006.05.001 (2007 ). Dharmagunawardhane H. A. N. . Photocatalytic hydrogen evolution using nanocrystalline gallium oxynitride spinel . Journal of Materials Chemistry A 2 , 19247 –19252 , 10.1039/C4TA03676J (2014 ). Hu C.-C. & Teng H. Gallium Oxynitride Photocatalysts Synthesized from Ga(OH)3 for Water Splitting under Visible Light Irradiation . The Journal of Physical Chemistry C 114 , 20100 –20106 , 10.1021/jp1070083 (2010 ). Lowther J. , Wagner T. , Kinski I. & Riedel R. Potential gallium oxynitrides with a derived spinel structure . Journal of alloys and compounds 376 , 1 –4 (2004 ). Kroll P. Spinel-type gallium oxynitrides attainable at high pressure and high temperature . Physical Review B 72 , 144407 (2005 ). Wolter S. , DeLucca J. , Mohney S. , Kern R. & Kuo C. An investigation into the early stages of oxide growth on gallium nitride . Thin Solid Films 371 , 153 –160 (2000 ). Puchinger M. , Kisailus D. J. , Lange F. F. & Wagner T. Microstructural evolution of precursor-derived gallium nitride thin films . Journal of Crystal Growth 245 , 219 –227 , 10.1016/S0022-0248(02)01712-8 (2002 ). Martin M. . Thermodynamics, structure and kinetics in the system Ga–O–N . Progress in Solid State Chemistry 37 , 132 –152 , 10.1016/j.progsolidstchem.2009.11.005 (2009 ). Maeda K. & Domen K. New Non-Oxide Photocatalysts Designed for Overall Water Splitting under Visible Light . The Journal of Physical Chemistry C 111 , 7851 –7861 , 10.1021/jp070911w (2007 ). Kudo A. & Miseki Y. Heterogeneous photocatalyst materials for water splitting . Chemical Society Reviews 38 , 253 –278 , 10.1039/B800489G (2009 ).19088977 Arai N. . Overall Water Splitting by RuO2-dispersed Divalent-ion-doped GaN Photocatalysts with d10 Electronic Configuration . Chemistry Letters 35 , 796 –797 , 10.1246/cl.2006.796 (2006 ). Chen H. . In Situ XRD Studies of ZnO/GaN Mixtures at High Pressure and High Temperature: Synthesis of Zn-Rich (Ga1−x Zn x)(N1−x O x) Photocatalysts . The Journal of Physical Chemistry C 114 , 1809 –1814 (2010 ). Maeda K. . Overall Water Splitting on (Ga1-x Zn x)(N1-x O x) Solid Solution Photocatalyst: Relationship between Physical Properties and Photocatalytic Activity . The Journal of Physical Chemistry B 109 , 20504 –20510 (2005 ).16853653 Hirai T. . Origin of Visible Light Absorption in GaN-Rich (Ga1-x Zn x)(N1-x O x) Photocatalysts . The Journal of Physical Chemistry C 111 , 18853 –18855 (2007 ). Chengchun T. . Synthesis of nanoporous spheres of cubic gallium oxynitride and their lithium ion intercalation properties . Nanotechnology 21 , 115705 (2010 ).20173228 Kinski I. . High-pressure synthesis of a gallium oxonitride with a spinel-type structure . Zeitschrift fur Naturforschung - Section B Journal of Chemical Sciences 60 , 831 –836 (2005 ). Huppertz H. . High-Pressure Synthesis, Electron Energy-Loss Spectroscopy Investigations, and Single Crystal Structure Determination of a Spinel-Type Gallium Oxonitride Ga2.79□0.21(O3.05N0.76□0.19) . Chemistry of Materials 21 , 2101 –2107 , 10.1021/cm803371k (2009 ). Okeke O. U. & Lowther J. Theoretical electronic structures and relative stabilities of the spinel oxynitrides M 3 N O 3 (M = B, Al, Ga, In) . Physical Review B 77 , 094129 (2008 ). Carin R. , Deville J. P. & Werckmann J. An XPS study of GaN thin films on GaAs . Surface and Interface Analysis 16 , 65 –69 , 10.1002/sia.740160116 (1990 ). Bhaviripudi S. , Qi J. , Hu E. L. & Belcher A. M. Synthesis, Characterization, and Optical Properties of Ordered Arrays of III-Nitride Nanocrystals . Nano Letters 7 , 3512 –3517 , 10.1021/nl072129d (2007 ).17949047 Procop M. XPS data for sputter-cleaned In0. 53Ga0. 47As, GaAs, and InAs surfaces . Journal of electron spectroscopy and related phenomena 59 , R1 –R10 (1992 ). Martin G. . Valence-band discontinuity between GaN and AIN measured by x-ray photoemission spectroscopy . Journal of electronic materials 24 , 225 –227 (1995 ). Maruyama T. . Electronic structure of wurtzite-and zinc blende-GaN studied by angle-resolved photoemission . Journal of electronic materials 27 , 200 –205 (1998 ). Fiorentini V. , Methfessel M. & Scheffler M. Electronic and structural properties of GaN by the full-potential linear muffin-tin orbitals method: The role of the d electrons . Physical Review B 47 , 13353 (1993 ). Bois L. . Characterization of a boro-silicon oxynitride prepared by thermal nitridation of a polyborosiloxane . Journal of Alloys and Compounds 232 , 244 –253 , 10.1016/0925-8388(95)01982-0 (1996 ). Grünert W. . A new facility for inert transfer of reactive samples to XPS equipment . Journal of Electron Spectroscopy and Related Phenomena 40 , 187 –192 , 10.1016/0368-2048(86)80017-2 (1986 ). Kawamura M. , Abe Y. , Yanagisawa H. & Sasaki K. Characterization of TiN films prepared by a conventional magnetron sputtering system: influence of nitrogen flow percentage and electrical properties . Thin Solid Films 287 , 115 –119 (1996 ). Ingo G. & Zacchetti N. XPS investigation on the growth model of a-SiNx and silicon and nitrogen chemical bondings . High temperature science 28 , 137 –151 (1988 ). Apai G. & Frederick B. G. X-ray photoelectron spectroscopic characterization of carbon monoxide bonding sites on supported small rhodium clusters . Langmuir 3 , 395 –400 , 10.1021/la00075a021 (1987 ). Hollinger G. , Skheyta-Kabbani R. & Gendry M. Oxides on GaAs and InAs surfaces: An x-ray-photoelectron-spectroscopy study of reference compounds and thin oxide layers . Physical Review B 49 , 11159 –11167 (1994 ). Mizokawa Y. , Iwasaki H. , Nishitani R. & Nakamura S. ESCA studies of Ga, As, GaAs, Ga 2 O 3, As 2 O 3 and As 2 O 5 . Journal of Electron Spectroscopy and Related Phenomena 14 , 129 –141 (1978 ). Zhdan P. , Shepelin A. , Osipova Z. & Sokolovskii V. The extent of charge localization on oxygen ions and catalytic activity on solid state oxides in allylic oxidation of propylene . Journal of Catalysis 58 , 8 –14 (1979 ). Iwakuro H. , Tatsuyama C. & Ichimura S. XPS and AES studies on the oxidation of layered semiconductor GaSe . Japanese Journal of Applied Physics 21 , 94 (1982 ). Perron H. . Combined investigation of water sorption on TiO2 rutile (1 1 0) single crystal face: XPS vs. periodic DFT . Surface Science 601 , 518 –527 , 10.1016/j.susc.2006.10.015 (2007 ). Dupin J.-C. , Gonbeau D. , Vinatier P. & Levasseur A. Systematic XPS studies of metal oxides, hydroxides and peroxides . Physical Chemistry Chemical Physics 2 , 1319 –1324 , 10.1039/A908800H (2000 ). Ramana C. V. . Chemical bonding, optical constants, and electrical resistivity of sputter-deposited gallium oxide thin films . Journal of Applied Physics 115 , 043508 , 10.1063/1.4862186 (2014 ). Priyantha W. , Radhakrishnan G. , Droopad R. & Passlack M. In-situ XPS and RHEED study of gallium oxide on GaAs deposition by molecular beam epitaxy . Journal of Crystal Growth 323 , 103 –106 , 10.1016/j.jcrysgro.2010.11.122 (2011 ). Zheng B. , Hua W. , Yue Y. & Gao Z. Dehydrogenation of propane to propene over different polymorphs of gallium oxide . Journal of Catalysis 232 , 143 –151 , 10.1016/j.jcat.2005.03.001 (2005 ). Atuchin V. V. . Spectroscopic ellipsometry and x-ray photoelectron spectroscopy of La2O3 thin films deposited by reactive magnetron sputtering . Journal of Vacuum Science & Technology A 29 , 021004 , 10.1116/1.3539069 (2011 ). Kong L. , Ma J. , Luan C. , Mi W. & Lv Y. Structural and optical properties of heteroepitaxial beta Ga2O3 films grown on MgO (100) substrates . Thin Solid Films 520 , 4270 –4274 , 10.1016/j.tsf.2012.02.027 (2012 ). Dharma J. & Pisal A. Simple method of measuring the band gap energy value of TiO2 in the powder form using a UV/Vis/NIR spectrometer . Perkin Elmer (2012 ). Hoffmann M. R. , Martin S. T. , Choi W. & Bahnemann D. W. Environmental Applications of Semiconductor Photocatalysis . Chemical Reviews 95 , 69 –96 , 10.1021/cr00033a004 (1995 ).

PMC005xxxxxx/PMC5000015.txt

==== Front Sci RepSci RepScientific Reports2045-2322Nature Publishing Group srep3208010.1038/srep32080ArticleNon-invasive MRI Assessments of Tissue Microstructures and Macromolecules in the Eye upon Biomechanical or Biochemical Modulation Ho Leon C. 123Sigal Ian A. 2356Jan Ning-Jiun 24Yang Xiaoling 12van der Merwe Yolandi 12456Yu Yu 7Chau Ying 78Leung Christopher K. 910Conner Ian P. 256Jin Tao 1Wu Ed X. 3Kim Seong-Gi 145111213Wollstein Gadi 2456Schuman Joel S. 2451114Chan Kevin C. a12456111 NeuroImaging Laboratory , University of Pittsburgh, Pittsburgh, PA, USA2 UPMC Eye Center, Ophthalmology and Visual Science Research Center, Department of Ophthalmology, School of Medicine, University of Pittsburgh, Pittsburgh, PA, USA3 Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, China4 Department of Bioengineering, Swanson School of Engineering, University of Pittsburgh, Pittsburgh, PA, USA5 McGowan Institute for Regenerative Medicine, University of Pittsburgh, Pittsburgh, PA, USA6 Louis J. Fox Center for Vision Restoration, University of Pittsburgh, Pittsburgh, PA, USA7 Department of Chemical and Biomolecular Engineering, Hong Kong University of Science and Technology, Hong Kong, China8 Division of Biomedical Engineering, Hong Kong University of Science and Technology, Hong Kong, China9 University Eye Center, Hong Kong Eye Hospital, Hong Kong, China10 Department of Ophthalmology and Visual Sciences, The Chinese University of Hong Kong, Hong Kong, China11 Center for the Neural Basis of Cognition, University of Pittsburgh and Carnegie Mellon University, Pittsburgh, PA, USA12 Center for Neuroscience Imaging Research, Institute for Basic Science, Suwon, Korea13 Department of Biomedical Engineering, Sungkyunkwan University, Suwon, Korea14 Clinical and Translational Science Institute, University of Pittsburgh, Pittsburgh, PA, USAa chuenwing.chan@fulbrightmail.org26 08 2016 2016 6 3208027 05 2016 02 08 2016 Copyright © 2016, The Author(s)2016The Author(s)This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/The microstructural organization and composition of the corneoscleral shell (CSS) determine the biomechanical behavior of the eye, and are important in diseases such as glaucoma and myopia. However, limited techniques can assess these properties globally, non-invasively and quantitatively. In this study, we hypothesized that multi-modal magnetic resonance imaging (MRI) can reveal the effects of biomechanical or biochemical modulation on CSS. Upon intraocular pressure (IOP) elevation, CSS appeared hyperintense in both freshly prepared ovine eyes and living rat eyes using T2-weighted MRI. Quantitatively, transverse relaxation time (T2) of CSS increased non-linearly with IOP at 0–40 mmHg and remained longer than unloaded tissues after being unpressurized. IOP loading also increased fractional anisotropy of CSS in diffusion tensor MRI without apparent change in magnetization transfer MRI, suggestive of straightening of microstructural fibers without modification of macromolecular contents. Lastly, treatments with increasing glyceraldehyde (mimicking crosslinking conditions) and chondroitinase-ABC concentrations (mimicking glycosaminoglycan depletion) decreased diffusivities and increased magnetization transfer in cornea, whereas glyceraldehyde also increased magnetization transfer in sclera. In summary, we demonstrated the changing profiles of MRI contrast mechanisms resulting from biomechanical or biochemical modulation of the eye non-invasively. Multi-modal MRI may help evaluate the pathophysiological mechanisms in CSS and the efficacy of corneoscleral treatments. ==== Body The sclera and cornea are dense and fibrous connective tissues that form the outer coat of the eye, which act to support and protect the eye from the surrounding environments. These load-bearing connective tissues can dynamically interact with changing physiological conditions12. For example, varying intraocular pressures (IOP) alter the fiber orientations34 and collagen crimp structures5 in the corneoscleral shell and modify the biomechanical properties including strain, stiffness and hysteresis of the eye6. The corneoscleral shell in turn transfers tensile, compression and shear stresses on the lamina cribrosa in response to the IOP changes7, which contributes to the deformation of the optic nerve head and may result in neurodegeneration in the visual system in glaucoma8. The extracellular matrix in the corneoscleral shell may also be remodeled in other conditions such as aging9, myopia10 and keratoconus11, and plays an important role in transcorneal and transscleral drug delivery1213. Having a non-invasive and non-destructive tool to better understand the microstructural organization and compositions of the corneoscleral shell in different environments can help determine the dynamics and the biomechanical and biochemical properties of the eye. Such knowledge is important for resolving the mechanisms of vision-related diseases involving the sclera and cornea, and may guide strategies for more effective drug delivery and corneoscleral treatment for vision preservation and restoration. The corneoscleral shell structure has been studied extensively using atomic force14, transmission15 and scanning electron microscopies16, polarized light5, bright field10 and nonlinear microscopies17, as well as small-angle light scattering18 and small and wide-angle x-ray scattering19202122. Although these ex vivo imaging techniques have provided fundamental information on the microarchitecture and biochemistry of the tissue microstructures and collagen across ocular tissues with excellent resolution and sensitivity7, most of these techniques require chemical tissue processing or tissue sectioning, which is highly invasive and can be destructive or affect the true morphology of the collagen (e.g. substantial tissue shrinkage during preparation for histological labeling). These methods tend to have limited depth penetration, precluding analysis of the whole eye. Methods with wide fields of view, such as small-angle light scattering, have low resolution in the order of hundreds of microns, whereas methods with high resolution at sub-micron scale, such as transmission electron microscopy, have small fields of view, even when coupled with modern block-imaging techniques2324. Altogether, these limitations prevent multi-modality assessments on the same sample, and prevent in vivo applications. Limited methods have been available for non-invasive and quantitative 3D assessments of the sclera and cornea in the whole globe25, and the exact contributions of the microstructural organization and composition of the corneoscleral shell to the pathogenesis of vision-related diseases remains largely undetermined. Magnetic resonance imaging (MRI) offers non-invasive and multi-parametric methods for assessing the impaired visual system without depth limitation262728293031323334. However, MRI studies of the corneoscleral shell have been limited, partly due to the intrinsically fast transverse magnetic resonance relaxation of the fibrous tissues and the resulting low MRI signal intensities available for biological examinations. Recently, our group demonstrated the use of the magic-angle enhancement effect to improve MRI sensitivity for detecting the structural details of the collagen-rich sclera and cornea tissues, and their changes in T2 and T2* transverse relaxation times in tissues fixed at various levels of IOP25. To date, it remains unclear how the MRI contrast mechanisms of fresh sclera and cornea tissues may respond dynamically with changing IOP, and whether the magic-angle effect can be applied to the living tissues to differentiate the biological conditions between normotensive and hypertensive eyes in vivo. Recent studies also suggested the possible role of scleral treatment as potential therapeutics to glaucoma35. However, its effectiveness on neuroprotection remains controversial, partly because of the lack of tools to characterize the biomechanical properties of the sclera in vivo, and to monitor their global and regional changes under stress or treatments. Herein we report the effects of stepwise IOP changes on the fibrous tissues of freshly prepared, unfixed ovine eyes by examining the corresponding dynamic changes in T2 relaxation time and T2-weighted MRI signal intensities in the corneoscleral shell. In addition, we demonstrated in vivo the effects of chronic IOP elevation on the MRI contrasts in the sclera using a hydrogel-induced rat model of experimental glaucoma and T2-weighted MRI. Apart from basic MR relaxometry of the collagen-rich fibers36, advanced imaging techniques such as diffusion tensor MRI (DTI) and magnetization transfer MRI (MTI) can detect more specifically how collagen fibers are organized microstructurally, and how the extracellular matrix can be remodeled3738. DTI measures water diffusion patterns in free, restricted or hindered compartments, and can reveal microstructural organization such as primary fiber orientation, directionality and directional diffusivities. MTI measures the magnetization transfer between free water and water bounded to macromolecules through chemical exchange and/or dipole-dipole interactions39, and can reflect changes in macromolecular structures and contents. The potentials of DTI and MTI for examinations of collagen-rich fibrous tissues have been demonstrated recently in studies of normal and diseased tendons4041, menisci42, ligaments4344 and cartilages3845 but have not been investigated in the sclera or cornea for ophthalmic research. Such examinations are particularly important in diseases such as glaucoma, where vision loss cannot be recovered. In this study, we tested the central hypothesis that multi-modal MRI can detect and differentiate the effects of biomechanical or biochemical modulation on the sclera and cornea via IOP loading (mimicking ocular hypertension46), collagen cross-linking (mimicking aging conditions47 and corneoscleral stiffening248) and glycosaminoglycan cleavage (mimicking pathological changes in extracellular matrix compositions495051). We used T2-weighted MRI, DTI and MTI to characterize the sclera and cornea with and without different levels of acute and chronic IOP loading. Furthermore, we tested the feasibility and sensitivity of T2-weighted MRI, DTI and MTI profiling to assess the treatment effects of cross-linking and glycosaminoglycan depletion on the sclera and cornea with different concentrations of glyceraldehyde and chondroitinase-ABC solutions, respectively. Results Effects of dynamic IOP changes on corneoscleral tissues by ex vivo T2 magnetic resonance relaxometry Figure 1 shows the effects of stepwise changes in IOP on the transverse relaxation time (T2) in the sclera and cornea of freshly prepared ovine eyes (Group 1). When IOP initially rose from 0 to 10 mmHg, no apparent T2 change was observed in either the sclera or the cornea. T2 began to increase when IOP was elevated from 10 to 20 mmHg. At 20 to 40 mmHg, T2 in sclera and cornea further increased and was significantly higher than the unloaded control. After being unpressurized from 40 mmHg back to 0 mmHg, T2 values of the ocular tissues remained higher than the unloaded control for about 6 hours. Effects of chronic IOP elevation on rat scleral tissues by in vivo T2-weighted MRI The effects of IOP loading on the ocular morphology and T2-weighted signal intensity in the living animals were illustrated in Fig. 2. In the rat model of experimental glaucoma, intracameral hydrogel injection to the right eye induced sustained IOP elevation from 2 days post-hydrogel injection up to MRI experiment at 1.5 weeks post-hydrogel injection. The IOP levels immediately before MRI experiment were 28.6 ± 4.7 mmHg in the right, hydrogel-injected eye and 13.7 ± 3.2 mmHg in the left, untreated control eye (Two-tailed paired t-test: p < 0.001). The right anterior chamber significantly enlarged at 1.5 weeks after hydrogel injection with no significant difference in the vitreous size between contralateral eyes (Fig. 2a,c). In addition, the sclera of the hypertensive right eye showed significantly higher normalized T2-weighted signal intensity than the normotensive left eye near the magic angle at 55° to the main magnetic field (Bo) but not at 0° to Bo (Fig. 2b,d). Tissue magnetic resonance relaxation, diffusion and magnetization transfer properties in unloaded eyes by T2-weighted MRI, diffusion tensor MRI and magnetization transfer MRI In addition to T2 relaxometry, MRI allows multi-modal profiling among different structures in the whole globe of the eye. Within the unloaded ovine eye (Group 2), the lens and sclera appeared the darkest in T2-weighted MRI followed by the cornea, iris and optic nerve, whereas the anterior chamber and vitreous appeared the brightest (Fig. 3a). In the color-encoded fractional anisotropy map of the same eye in DTI (Fig. 3b), the principal diffusion orientations of the fibrous tissues including the sclera, cornea, lens cortex, retina, optic nerve and dura generally corroborated the fiber arrangements in the polarized light microscopy images of histological sections (Fig. 3c,d). In addition, a sharp boundary was observed in the color-encoded fractional anisotropy map between the outer cortex and inner nucleus of the lens, as well as near the lamina cribrosa in the optic nerve head. Quantitative measurements of the T2 relaxation times, diffusion properties and magnetization transfer properties of the fresh ovine eyes at 10 mmHg IOP were tabulated in Table 1 showing distinct MRI contrasts among cornea, sclera, retina, optic nerve, lens, anterior chamber and vitreous body. Specifically, the sclera had the lowest T2 value and the highest magnetization transfer ratio among all ocular components measured. The anterior chamber and vitreous body had the highest T2 and mean diffusivity, but the lowest magnetization transfer ratio and fractional anisotropy. The optic nerve and the lens cortex had the highest fractional anisotropy, while the optic nerve had the lowest mean diffusivity among all measured ocular components (Tukey’s multiple comparisons tests: p < 0.05). Microstructural organization and macromolecular contents of loaded and unloaded sclera, cornea and tendon tissues by T2-weighted MRI, diffusion tensor MRI and magnetization transfer MRI In the tissue strips of the loaded and unloaded ovine eyes and tendons (Group 2) (Fig. 4), different profiles of T2-weighted signal intensity, fractional anisotropy and magnetization transfer ratio were observed in T2-weighted MRI, DTI and MTI among loaded and unloaded sclera, cornea and tendon when the tissues were oriented near the magic angle at about 55° to Bo. In Fig. 4b, normalized T2-weighted signal intensities were higher in the cornea than the sclera and tendon under both loaded and unloaded conditions. In addition, all cornea, sclera and tendon tissues became hyperintense in T2-weighted MRI after loading similar to our previous study25. In Fig. 4c, the unloaded tendon exhibited the highest magnetization transfer ratio followed by the unloaded sclera and cornea. However, no significant difference in magnetization transfer ratio was found between loaded and unloaded tissues. DTI quantitation in Fig. 4d–g showed the highest fractional anisotropy and the lowest directional diffusivities (λ// and λ⊥) in unloaded tendon followed by unloaded sclera and cornea. Loaded sclera, cornea and tendon tissues also showed significantly higher fractional anisotropy than unloaded tissues. A trend of lower λ⊥ was observed in loaded sclera than unloaded sclera. Effects of collagen crosslinking or glycosaminoglycan depletion on sclera and cornea tissues by T2-weighted MRI, diffusion tensor MRI and magnetization transfer MRI When the fresh ovine cornea and sclera tissues were treated with increasing concentrations of glyceraldehyde cross-linking solutions (Group 3a), the occurrence of non-enzymatic glycation was clearly evident by a gradual color change from whitish to brownish yellow similar to a recent study52. In MTI, magnetization transfer ratio in both cornea and sclera tissues significantly increased with increasing glyceraldehyde concentrations (Fig. 5c), whereas DTI quantitation in Fig. 5d–g showed that λ//, λ⊥ and mean diffusivity in cornea significantly decreased with increasing glyceraldehyde concentrations. Similar changes were detected in chondroitinase-ABC treated ovine cornea but not sclera (Group 3b), whereby significantly increasing magnetization transfer ratio and decreasing λ//, λ⊥ and mean diffusivity were observed with increasing chondroitinase-ABC concentrations (Fig. 6). No significant difference in fractional anisotropy or normalized T2-weighted signal intensity was detected in the cornea or sclera among different glyceraldehyde or chondroitinase-ABC concentrations. Discussion In this study, we used freshly prepared ovine eyes to evaluate the tissue magnetic resonance relaxation in response to dynamic IOP changes, since tissue fixation may alter the MRI contrast mechanisms in the fibrous tissues37 as well as the biomechanical responses of the eye to changing IOP12. When the fresh ocular tissues were loaded beyond physiological IOP levels, the consistent T2 increase observed in Fig. 1 might be explained by the diminishing magnetization spin dephasing between increasingly distant neighboring protons during collagen fiber straightening36. More importantly, the non-linearity of T2 changes with increasing IOP at 0–40 mmHg echoed with recent histological studies showing non-linear uncrimping of collagen fibrils in the corneoscleral shell with IOP loading5354. After being unpressurized, T2 of the fresh IOP-loaded ovine ocular tissues remained significantly higher than the unloaded control for few hours, suggestive of hysteresis or hydraulic effect after IOP loading55. MRI detection of hysteresis is potentially important to determine globally the biomechanical state of the corneoscleral shell and its interplay with the surrounding ocular structures and the behavioral outcomes. Apart from dynamic MRI assessments of fresh ocular tissues ex vivo, the present study demonstrated the use of in vivo magic angle-enhanced MRI to differentiate sclera tissues between normotensive and hypertensive eyes in the rat glaucoma model of chronic IOP elevation as shown in Fig. 2. These initial results opened up the possibility for future non-invasive MRI assessments of the corneoscleral biomechanics in the living eyes, which may help better predict the dynamic profiles of IOP and identify new modifiable risk factors for glaucomatous optic neuropathy. In addition to the sclera and cornea, multi-modal MRI allows non-invasive profiling and comparisons among different structures in the whole globe of the eye without depth limitation. In the normal ovine eye in Fig. 3 and Table 1, the short T2 and the resulting low T2-weighted signals in the lens and sclera could be explained by the abundance of tightly bound protons associated to macromolecules such as collagen. Furthermore, the decreasing T2-weighted signal intensities within the lens indicated increasing bound water and less free water from increasing protein concentration toward the center of the ocular lens56. The long T2 and bright T2-weighted signals in the anterior chamber and vitreous body reflected the abundance of mobile water protons in the aqueous humor and vitreous humor. Apart from conventional T2-weighted MRI, advanced diffusion MRI techniques such as DTI allow microstructural assessments by measuring water diffusion patterns in biological tissues. The diffusivity patterns observed in the ovine lens generally agreed with other species56, which showed anisotropic water movements and stronger diffusivity in the outer cortex containing elongated lens fiber cells, and isotropic patterns and weaker diffusivity in the protein core of the lens in the absence of a blood supply. The high fractional anisotropy in the intraorbital optic nerve indicated the unilaterality of myelinated axonal fibers, whereas the sharp boundary at the lamina cribrosa and between the outer cortex and inner nucleus of the lens in the fractional anisotropy map indicated the presence of a diffusion barrier limiting the transport of water and other metabolites across the boundary. While DTI may help examine ocular tissue microstructures, MTI may probe macromolecular contents in the ocular tissues. Although the short T2 of the less mobile protons associated with macromolecules may not be easily detected directly in conventional MRI, the coupling between the macromolecular protons and the mobile protons allows the spin state of the macromolecular protons to influence the spin state of the liquid protons through exchange processes39, leading to the magnetization transfer contrast. The magnetization transfer effect is dependent on the concentration, mobility, and surface chemistry of the macromolecules57. In the present study, the sclera, cornea, optic nerve and lens cortex had 6–10 times higher magnetization transfer ratio than the anterior chamber and vitreous body, indicative of the higher macromolecular contents in these ocular tissues compared to the water-rich ocular chambers. In vivo measurement of magnetization transfer ratio has been demonstrated to reflect the collagen status and may be applied for the routine evaluation of normal and abnormal fibrous tissues such as articular cartilage58. Previous MTI studies have shown that magnetization transfer can give better image contrasts between the lens cortex and nucleus than conventional T2-weighted MRI sequences in normal and cataratous lenses57. Taken together, combined T2-weighted MRI, DTI and MTI may offer promises to improve the specificity in evaluating the conditions of normal and abnormal eye globes. To validate the feasibility and sensitivity of multi-modal MRI use in IOP-loaded and unloaded eyes, we compared the T2-weighted MRI, DTI and MTI profiles between IOP-loaded and unloaded cornea and sclera tissues in Fig. 4. Stretch-loaded and unloaded tendons were also scanned as a positive control given the similar structural compositions of the tendon and the corneoscleral shell as well as existing reports detailing MRI evaluations of the tendon253641. In DTI, among the 3 unloaded tissues, the highest fractional anisotropy observed in tendon conformed with the nearly parallel alignment of tendon collagen fibrils to the long axis59. Although fibers in the corneoscleral shell are also highly aligned in-plane, they may traverse each other in slightly different directions to provide maximal mechanical strength within the curved globe2260. Such crossing fibers may lead to lower overall fractional anisotropy in ocular tissues compared to the tendon. The highest directional diffusivities (λ// and λ⊥) in the cornea followed by the sclera and tendon likely reflect the rich water content and more regular lattice arrangement in the corneal stroma. Upon pressure loading, the collagen-rich tissues experienced stretch and compression leading to fiber straightening3 and a more anisotropic microstructural environment61. Our results of the significantly higher fractional anisotropy but not directional diffusivities in loaded than unloaded tisues suggested that fractional anisotropy is a more sensitive DTI marker than λ// and λ⊥ to detect pressure loading in ocular tissues and tendons. Previous studies suggested that the dynamic responses of the sclera to experimental glaucoma may be as important as the baseline anatomic features in explaining susceptibility to neuronal damage61. Here, the increase in fractional anisotropy of loaded sclera and cornea tissues may indicate short-term plasticity in the interwoven collagen lamellae in response to elevated IOP during glaucoma exposure62. In MTI, the magnetization transfer effect in the unloaded tendon was slightly stronger than the unloaded sclera and significantly stronger than the unloaded cornea. This may be explained by the most abundant collagen contents in tendon (~100% dry mass) compared to sclera and cornea (~70–90% dry weight)63. The insignificant difference in magnetization transfer ratio between loaded and unloaded tissues suggests no apparent alteration in macromolecular contents despite microstructural changes upon acute pressure loading4. Although ocular and tendon tissues have similar extracellular matrix compositions, there are some important differences in their structural hierarchies. Tendon has a highly ordered fascicular structural hierarchy whereas corneal and scleral collagen does not. The molecular tilt in the collagen fibrils are also different between the tissues, with a more oblique molecular tilt in the cornea (~15°) vs. the tendon (~4°) with respect to the fibril axis19. It has been suggested that this molecular tilt is a result of the lower axial periodicity of microfibrils in the cornea compared to the tendon19. The consequences of this difference in molecular tilt on our MRI results are unclear. In a previous MRI study examining the extents of T2 and T2* signal enhancement at multiple angles to the main magnetic field, we did not detect a deviation of angle at maximum signal enhancement from the magic angle between tissues25. Whether this tilt may have contributed to the lower magic-angle effect strength and the sensitivity to DTI and MTI changes in the ocular tissues compared with the tendon will be the subject of further investigation. It is critical to determine whether the corneoscleral reorganization in human eyes is a beneficial adaptation or a detrimental contributor to optic nerve head injury in glaucoma. Multi-modal MRI may be useful to measure the biomechanical behavior of the eyes non-invasively in future, both to monitor the baseline state of the eye as a risk factor for future development of glaucoma, and to assess disease progression. MRI of ocular tissue structures and compositions may provide a non-destructive and non-invasive means for screening patients’ susceptibility to glaucomatous damage and for determining whether the greater stiffness of human glaucoma eyes is present at baseline, whether it develops as a response to the disease, or both2, in order to design optimal biomechanical modification therapies35. In MTI, it has been suggested that magnetization transfer ratio may differentiate between various pathomimetic degradative procedures64. While the baseline magnetization transfer ratio in fibrous tissues may be primarily due to the tissue collagen concentration, changes in magnetization transfer ratio may be due to physiological or pathophysiological changes in tissue structure and glycosaminoglycan concentration58. In the present study, the increased magnetization transfer ratio in the cross-linked cornea and sclera in Fig. 5 could be explained by the larger pool of immobile water molecules available for transfer of magnetization to free water when cross-linking packing density increased at a constant macromolecular concentration37. Glyceraldehyde cross-linking may also reduce permeability or induce tissue shrinkage leading to the reduced diffusivity observed in the cornea65. These findings introduced the possibility of multi-modal MRI to characterize the state of collagen cross-linking in different regions of the eye in future in vivo studies. Pathological changes in glycosaminoglycan content have been observed in eyes with glaucoma and myopia495051, whereas glycosaminoglycan removal may contribute to an increase in hydration and altered creep and stiffness of the sclera66. In the present study, glycosaminoglycan depletion by chondroitinase ABC did not alter T2-weighted signals in cornea and sclera in Fig. 6 similar to a recent study on chondroitinase ABC-treated cartilage cultures with altered biomechanical properties but not collagen contents or T2 values67. On the other hand, chondroitinase ABC increased magnetization transfer ratio and decreased diffusivities but not fractional anisotropy in the cornea, whereas no significant change was observed in the sclera under the same chondroitinase-ABC concentrations. Whether the observed differences in treatment responses between cornea and sclera tissues were due to different extents of enzyme penetration, different tissue compositions or other factors will require additional studies in future employing a larger range of chondroitinase-ABC concentrations and considering other MRI modalities such as T1ρ and chemical exchange saturation transfer to further improve specificity. Complementary to other imaging modalities such as spectral-domain optical coherence tomography and ultrasound7, multi-modal MRI may further improve the understanding of the remodeling of the collagen and proteoglycan structures in the corneoscleral shell, and the biomechanical characteristics that predispose clinically to the development of ocular pathologies. This study has several limitations. While this study demonstrated the feasibility and sensitivity to detect changes in MRI contrast mechanisms in response to biomechanical or biochemical manipulation in the eye, the direct linkages between MRI contrast mechanisms and biomechanical properties, despite demonstrated in other fibrous tissues under various conditions6869, remain to be elucidated in the eye in the future. On the other hand, although our MRI methods had achieved sub-millimeter in-plane resolution to examine the collagen organization relevant to lamellar and crimp properties, the current MRI study is limited in its out-of-plane resolution due to the relatively large 1 mm slice thickness. It should be noted that this slice thickness was chosen as an early step to guarantee sufficient signal-to-noise ratio for detecting tissue changes in response to biomechanical or biochemical modulation. Since MRI has no depth limitation and the slice thickness can be further optimized, we expect that multi-modal MRI has the potential to resolve collagen organization in thinner slices for more detailed 3D analyses of the collagen structure. This may provide a more complete evaluation of the eye’s anisotropic structure and biomechanical performance in future studies. Lastly, while ovine eyes were chosen in this study for ex vivo exmainations to mimic human eyes of similar sizes, when translating this study to imaging humans or larger animals in vivo, a larger scanner bore size and lower magnetic field strengths are expected for more practical use, and limited image signal-to-noise ratio may be resulted at high resolutions. Eye motion is another potential challenge for in vivo ocular imaging. Using specialized orbital coils, recent studies had achieved sub-millimeter resolutions for human ocular MR imaging in the clinical 3-Tesla scanners and research-based 7-Tesla human scanners7071. Eye tracking and gaze fixation can also help minimize motion artifacts and improve image quality within clinically acceptable scan time. By instructing our subject to maintain stable eye fixation on several targets at various angles relative to the main magnetic field, our pilot study demonstrated the magic-angle effect in the human sclera in vivo using 3-Tesla MRI72. These studies provide initial evidence to support the translation of the current high-field sclera and cornea MRI toward in vivo imaging in humans. Conclusion T2-weighted MRI, DTI and MTI non-destructively distinguished between different ocular tissues with and without biomechanical or biochemical modulation via acute and chronic IOP loading, collagen crosslinking and glycosaminoglycan depletion. Multi-modal MRI may provide cross-sectional and longitudinal monitoring of the ocular fiber organization and remodeling in aging and diseases involving the corneoscleral shell. This may help evaluate the pathophysiological mechanisms in the corneoscleral shell and the efficacy of corneoscleral treatments in a variety of ophthalmic diseases. Materials and Methods Ovine eye and tendon tissue preparation for ex vivo imaging Twenty three pairs of ovine eyes were obtained from the local abattoir and processed within 12 hours of death to minimize mechanical deterioration73. The ovine eyes were divided into 3 groups: Group 1: Effects of dynamic biomechanical loading on fresh ovine eyes by T2 magnetic resonance relaxometry Twelve fresh, unfixed ovine eyes underwent anterior chamber perfusion in the 9.4 Tesla MRI scanner using a plastic cannula connected to a saline bag (0.9% sodium chloride; Baxter International Inc., Deerfield, IL, USA) and a pressure transducer (BIOPAC Systems, Goleta, CA, USA) so as to image the dynamic effects of stepwise changes in IOP on the tissue transverse relaxation times (T2) of the corneoscleral shell in the same eyes (Fig. 1a,b). Six of these 12 ovine eyes were loaded at 0, 10, 20 and 40 mmHg by raising the saline bag at different heights consecutively inside the MRI scanner, followed by unpressurization back to 0 mmHg until the end of the experiment. Although 0 mmHg is not a normal physiological condition, it was selected as the lowest pressure to provide fundamental insights into the state of the tissues without the load from IOP, which is essential in the formulation of a comprehensive mechanistic model of the tissues5535474. The pressure transducer was used to ensure a constant pressure was applied at the desired level during each experimental session. Six other fresh ovine eyes were cannulated but kept at 0 mmHg throughout the entire MRI experiment as an unloaded control. Hydrogel (McNeil-PPC, Inc., Skillman, NJ, USA) was applied to keep the surface of the ovine eyes moist throughout the experiment. Group 2: Effects of biomechanical loading on ovine cornea, sclera and tendon tissues by T2-weighted MRI, diffusion tensor MRI and magnetization transfer MRI Eight ovine eyes were loaded at an IOP of 50 mmHg using a gravity perfusion fixation system with a cannula inserted into the anterior chamber25 outside the MRI scanner, so as to image the effects of IOP loading on the corneoscleral shell with T2-weighted MRI, DTI and MTI. Twelve additional ovine eyes were cannulated but unpressurized. Ten minutes later, all eyes were immersion fixed with 10% formalin for 24 hours while remained loaded or unloaded, and then washed in phosphate buffered saline (PBS) (Fisher Scientific Inc., Pittsburgh, PA, USA). Three of the 12 unloaded ovine eyes underwent whole-globe MRI while another unloaded eye was processed in sucrose, cryosectioned axially into 30 μm thick histological sections and imaged with polarized light microcopy75. The remaining 8 loaded and 8 unloaded eyes were dissected using razors and surgical scissors to isolate the sclera and cornea of 8–12 mm long, 2–5 mm wide and 2–3 mm thick. Given the similar structural compositions between the tendon and corneoscleral shell, 13 ovine Achilles tendons from the same set of animals were washed in PBS and dissected to isolate thin strips of similar dimensions to compare with the ocular tissues. Four tendon strips were randomly selected and loaded using alligator clips attached to a rod to keep stretched for 10 minutes25, followed by immersion fixation with 10% formalin for 24 hours while remaining stretched. The remaining 9 strips were fixed but unloaded. Group 3: Effects of cross-linking and glycosaminoglycan depletion on ovine cornea and sclera tissues by T2-weighted MRI, diffusion tensor MRI and magnetization transfer MRI Sixteen freshly prepared ovine eyes were divided into 2 sub-groups of 8 eyes each in order to image the effects of glyceraldehyde cross-linking (Group 3a) and glycosaminoglycan depletion by chondroitinase-ABC treatment (Group 3b) on the sclera and cornea with T2-weighted MRI, DTI and MTI. Each eye was dissected at the central globes using razors and surgical scissors to give 4 sclera and 4 cornea fresh tissue strips of similar dimensions as in Group 2. In Group 3a, the 4 sclera and 4 cornea strips from each eye were treated with D-(+)-glyceraldehyde (Sigma-Aldrich, St. Louis, MO, USA) solutions at 0, 0.05, 0.1, or 0.2 M at 37 °C in the incubator for 12 hr. In Group 3b, the 4 sclera and 4 cornea strips from each eye were treated with chondroitinase-ABC (Sigma-Aldrich, St. Louis, MO, USA) solutions at 0, 0.06, 0.5 or 2 unit/ml at 37 °C in the incubator for 12 hr. The different glyceraldehyde concentrations were prepared by diluting the solutions with 1x PBS at pH 7.4 (Fisher Scientific Inc., Pittsburgh, PA, USA), whereas the different chondroitinase-ABC concentrations were prepared by diluting the solutions with 1 M Tris-HCl at pH 8.0 (Mediatech, Inc, Manassas, VA, USA ), 3 M sodium acetate buffer solution (Sigma-Aldrich, St. Louis, MO, USA), and 0.1% bovine serum albumin solution (Fisher Scientific Inc., Pittsburgh, PA, USA). Samples treated with sham solution without glyceraldehyde or chondroitinase-ABC were indicated as 0 M or 0 unit/ml. Rat model preparation for in vivo imaging All in vivo animal experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and protocols reviewed and approved by the University of Pittsburgh’s Institutional Animal Care and Use Committee. Five adult female Long Evans rats were first anesthetized by inhaling a mixture of air and isoflurane (3% for induction and 1.25% for maintenance). Proparacaine (Bausch & Lomb, Inc., Rochester, NY, USA) was topically administered to anesthetize the surface of the eye. The right eye was then intracamerally injected with 20 μL of a solution containing 6% vinysulfonated hyaluronic acid and 6% thiolated hyaluronic acid by a microinjection system through a sharp glass micropipette (World Precision Instruments, Sarasota, FL, USA) under a surgical microscope. The hyaluronic acid derivatives were synthesized as reported previously76. The polymer mixture solidified shortly after injection to an optically clear cross-linked hydrogel causing aqueous outflow obstruction and sustained IOP elevation77. The left eye did not receive any injection and served as an internal control. The IOPs of both eyes were measured using the TonoLab rebound tonometer (Colonial Medical Supply, Franconia, NH, USA) under light isoflurane gas anesthesia. At least 18 valid, non-failed IOP values from each eye were obtained and averaged. IOP measurements were performed every 3–4 days after hydrogel injection. MRI protocol All MRI experiments were performed using a 9.4-Tesla/31-cm Varian/Agilent horizontal MRI scanner (Santa Clara, CA, USA) with a 38 mm-diameter transmit-receive volume coil for ex vivo ovine eye and tendon imaging, and a volume transmit and surface receive coil for in vivo rat eye imaging. For T2 relaxometry in Group 1, the whole globe of the cannulated fresh eye underwent T2 mapping at each pressure level using a spin-echo MRI pulse sequence with the following imaging parameters: Repetition time (TR)/echo time (TE) = 1000/9.48 ms, echo space time = 9.48 ms, number of echoes = 5, in-plane resolution = 130 × 130 μm2, slice thickness = 1 mm and number of repetitions = 2. Diffusion tensor MRI (DTI) and magnetization transfer imaging (MTI) were also performed at 10 mmHg using fast spin-echo sequences with the following imaging parameters: (i) DTI: 2 non-diffusion-weighted (b0) images and 12 diffusion weighted images with 12 gradient directions at diffusion weighting factor (b) = 1000 s/mm2, diffusion gradient duration time (δ)/diffusion gradient separation time (Δ) = 5/17 ms, TR/TE = 2300/27.8 ms, echo train length = 8 and number of repetitions = 1; (ii) MTI: 9.5 μT saturation pulses at +6000 Hz off-resonance and 150 ms pulse length, TR/TE = 1500/8.43 ms, echo train length = 8 and number of repetitions = 2. The MTI parameters were selected so as to minimize direct saturation of the bulk water64. Both DTI and MTI shared the same slice geometry, with in-plane resolution = 140 × 140 μm2 and slice thickness = 1 mm. The total scanning duration for each pressure level was limited to 1 hr, and MRI scans began at about 20 min after one stepwise change in pressure to allow the ocular tissues to reach viscoelastic equilibrium78. In Group 2, DTI was performed to the whole globe of the fixed, unloaded eye under similar DTI parameters as in Group 1 except b = 500 s/mm2, δ/Δ = 5/12 ms, TR/TE = 2300/21.6 ms and number of repetitions = 4. The sectioned eye and tendon tissue samples in Groups 2 and 3 were suspended in agarose gel and were orientated near the magic angle at ~55° to the main magnetic field (Bo) during imaging. DTI and MTI were performed to the sectioned tissues with the same imaging parameters as in Group 1 except number of repetitions = 4 for DTI and in-plane resolution = 125 × 125 μm2 for both DTI and MTI. For the rat model of experimental glaucoma, T2-weighted MRI was acquired at 1.5 weeks after induction of chronic IOP elevation under ketamine (75 mg/kg) and xylazine (10 mg/kg) anaesthesia using a spin-echo imaging sequence with TR/TE = 1000/13.6 ms, in-plane resolution = 55 × 55 μm2, slice thickness = 1.0 mm, number of repetitions = 2 and total scanning duration = 17 min. Data analysis In Group 1, the T2 transverse relaxation times of the eyes at each pressure level were derived from curve fitting of the T2-weighted signal decay with echo time using the equation S(TE) = S(0)e(−TE/T2) where S is the signal intensity (S) from the spin-echo MR image at a particular echo time (TE)25. T2-weighted MRI in Group 1 was obtained from T2 mapping, whereas T2-weighted images in Groups 2 and 3 were obtained from the non-diffusion-weighted b0 images in DTI. For DTI, co-registration between non-diffusion-weighted b0 images and diffusion-weighted images was performed using SPM8 (Wellcome Department of Imaging Neuroscience, University College, London, UK). Using DTIStudio v3.02 (Johns Hopkins University, Baltimore, MD), 3 × 3 diffusion tensors were fitted on a pixel-by-pixel basis from the non-diffusion-weighted b0 images and the diffusion-weighted images. The eigenvectors and eigenvalues of the diffusion tensors were derived to compute the DTI parametric maps including the fractional anisotropy directionality color map, fractional anisotropy value map, and axial diffusivity (λ//), radial diffusivity (λ⊥) and mean diffusivity maps. For MTI, magnetization transfer ratio was calculated by the equation (M0 − Msat)/M0 where Msat and M0 are the magnetization signals with and without saturation pulse respectively. Regions of interest were manually drawn and measured on the cornea, sclera or tendon near the magic angle at ~55° to the main magnetic field (Bo), as well as the retina, optic nerve, lens cortex, anterior chamber, vitreous body using ImageJ v1.47 (Wayne Rasband, NIH, USA) for quantitative analyses. T2-weighted signal intensities in the eye or tendon tissues were normalized to the nearby gel for the ovine in Groups 2 and 3, and to the nearby vitreous for rats to account for coil sensitivity and magnetic field inhomogeneity. Results were presented as mean ± standard deviation. Comparisons among conditions were performed using analyses of variance (ANOVA) followed by post-hoc multiple comparisons tests using GraphPad Prism v6.00 (GraphPad Software Inc., La Jolla, CA, USA) unless otherwise specified. Results were considered significant when p < 0.05. Additional Information How to cite this article: Ho, L. C. et al. Non-invasive MRI Assessments of Tissue Microstructures and Macromolecules in the Eye upon Biomechanical or Biochemical Modulation. Sci. Rep. 6, 32080; doi: 10.1038/srep32080 (2016). This work was supported by the National Institutes of Health P30-EY008098, R01-EY023966, R01-EY025011 and UL1-TR000005 (Bethesda, Maryland); BrightFocus Foundation G2013077 (Clarksburg, Maryland); Stimulating Pittsburgh Research in Geroscience Pilot Project Program Award (Pittsburgh, PA); Eye and Ear Foundation (Pittsburgh, Pennsylvania); and Research to Prevent Blindness (New York, New York). The author(s) have made the following disclosure(s): J.S.S.: Royalties e Zeiss, Dublin, CA (for intellectual property licensed by the Massachusetts Institute of Technology and Massachusetts Eye and Ear Infirmary). Author Contributions Study conception and design: L.C.H., I.A.S., N.J.J., Y.C., C.K.L., E.X.W., S.G.K., G.W., J.S.S. and K.C.C.; Data collection: L.C.H., I.A.S., N.J.J., X.L.Y., Y.V.D.M., Y.Y., I.P.C., T.J. and K.C.C.; Data analysis: L.C.H., I.A.S., N.J.J., T.J. and K.C.C.; Manuscript writing: L.C.H., I.A.S., N.J.J., X.L.Y., Y.Y., I.P.C., T.J., G.W. and K.C.C. All authors read and approved the final manuscript. Figure 1 Dynamic imaging of the effects of stepwise intraocular pressure (IOP) changes on magnetic resonance relaxometry in the corneoscleral tissues of freshly prepared ovine eyes. (a) Schematic diagram of anterior chamber perfusion to a fresh, unfixed ovine eye in the 9.4 Tesla MRI scanner. A plastic cannula was inserted into the anterior chamber and connected to a saline bag that was risen at different heights to induce different levels of IOP elevation. (b) Experimental paradigm of the stepwise IOP loading MRI experiment and the IOP-unloaded control experiment. (c) Representative ex vivo T2-weighted MRI (T2WI) of the same ovine eye loaded at 0, 10, 20 and 40 mmHg. (d,e) Quantitative comparisons (mean ± standard deviation) of transverse relaxation time (T2) in cornea (d) and sclera (e) upon stepwise IOP loading (dashed lines) and in unloaded control (solid lines). Both cornea and sclera T2 gradually increased as IOP increased from 0 to 40 mmHg (Tukey’s multiple comparisons tests between first MRI session and subsequent sessions, ##p < 0.01, ###p < 0.001; comparisons between other sessions are not shown here for clarity) but remained unchanged after being unpressurized (Tukey’s multiple comparisons tests between MRI sessions, p > 0.05). No significant change was observed in the unloaded control tissues over time (Tukey’s tests between MRI sessions, p > 0.05) (Sidak’s multiple comparison tests between IOP loaded and unloaded control tissues: *p < 0.05, **p < 0.01) (Bo: main magnetic field). Figure 2 In vivo imaging of the effects of chronic IOP elevation on the rat scleral tissues. (a) Representative in vivo T2-weighted MRI (T2WI) of the hypertensive right eye and normotensive left eye of adult Long Evans rats at 1.5 weeks after intracameral hydrogel injection to the right eye. (b) Enlarged view of the rat eyes from the broken-line boxes in (a). Blue and red arrows in both eyes indicated the sclera tissues oriented at 0° to the main magnetic field (Bo) and near the magic angle at 55° to Bo respectively. (c) Quantitative comparisons (mean ± standard deviation) of the cross-sectional areas of the anterior chamber and vitreous between the hydrogel-injected right eye and the uninjected left eye. (d) T2-weighted signal intensities (SI) in the sclera tissues oriented at about 0° or 55° (the magic angle) to Bo in both eyes. (Sidak’s multiple comparisons tests between left and right eyes: §p < 0.05; **p < 0.01; Tukey’s multiple comparisons tests between 0° to Bo and 55° to Bo on the same eye: #p < 0.05). Figure 3 Whole eye imaging and histological confirmation. Representative (a) T2-weighted MRI, (b) color-encoded fractional anisotropy map in diffusion tensor MRI (DTI), (c) histological section imaged with polarized light microscopy, and (d) energy, or intensity-map of the collagen density parallel to the plane of the section, calculated from a series of polarized light microscopy images of the unloaded ovine eye. (Color representations for the principal diffusion directions in the color-encoded fractional anisotropy map: Blue: caudal-rostral; red: left-right; green: dorsal-ventral) (Bo: main magnetic field). Figure 4 Microstructural organization and macromolecular contents of IOP-loaded sclera and cornea tissues, and stretch-loaded tendon tissues by T2-weighted MRI (T2WI), diffusion tensor MRI (DTI) and magnetization transfer MRI (MTI). (a) Representative T2WI (top), fractional anisotropy (FA) maps (middle) and magnetization transfer ratio (MTR) maps (bottom) of loaded (red arrows) and unloaded (blue arrows) cornea, sclera and tendon tissue strips. The tissues were oriented and measured near the magic angle at ~55° to the main magnetic field (Bo) to enhance MRI signals for more sensitive examinations; (b–g) Quantitative comparisons (mean ± standard deviation) of (b) T2-weighted signal intensity (SI), (c) MTR, (d) FA, (e) axial diffusivity (λ//), (f) radial diffusivity (λ⊥) and (g) mean diffusivity (MD) between loaded and unloaded cornea, sclera and tendon strips. ANOVA tests showed significant differences between cornea, sclera and tendon tissues under the same loading conditions for each MRI parameter in (b–g) (p < 0.05) except loaded tissues in MTR (c) (p > 0.05) (Paired t-tests between loaded and unloaded cornea or sclera: *p < 0.05, **p < 0.01; ***p < 0.001; Unpaired t-tests between loaded and unloaded tendon: *p < 0.05). Figure 5 Microstructural and macromolecular changes in the fresh sclera and cornea tissues after glyceraldehyde cross-linking treatment with T2-weighted MRI (T2WI), diffusion tensor MRI (DTI) and magnetization transfer MRI (MTI). (a) Representative T2WI (top), mean diffusivity (MD) maps (middle) and magnetization transfer ratio (MTR) maps (bottom) of cross-linked cornea (left panel) and sclera (right panel) after treating with 0.2 M glyceraldehyde solution or sham solution (0 M). The tissue strips were oriented and measured near the magic angle at ~55° to main magnetic field (Bo); (b–g) Quantitative comparisons (mean ± standard deviation) of (b) T2-weighted signal intensity (SI), (c) MTR, (d) fractional anisotropy (FA), (e) axial diffusivity (λ//), (f) radial diffusivity (λ⊥) and (g) MD in cornea and sclera strips after treatments with glyceraldehyde cross-linking solutions at 0, 0.05, 0.10 and 0.20 M concentrations. (Tukey’s multiple comparisons tests: *p < 0.05, **p < 0.01, ***p < 0.001). Figure 6 Microstructural and macromolecular changes in the fresh sclera and cornea tissues after glycosaminoglycan depletion by chondroitinase-ABC with T2-weighted MRI (T2WI), diffusion tensor MRI (DTI) and magnetization transfer MRI (MTI). (a) Representative T2WI (top), mean diffusivity (MD) maps (middle) and magnetization transfer ratio (MTR) maps (bottom) of glycosaminoglycan removal in cornea (left panel) and sclera (right panel) after treating with 2 units/ml (U/ml) of chondroitinase-ABC solution or sham solution (0 U/ml). Tissue strips were oriented and measured about the magic angle at ~55° to main magnetic field (Bo); (b–g) Quantitative comparisons (mean ± standard deviation) of (b) T2-weighted signal intensity (SI), (c) MTR, (d) fractional anisotropy (FA), (e) axial diffusivity (λ//), (f) radial diffusivity (λ⊥) and (g) mean diffusivity (MD) in cornea and sclera strips after treatment with chondroitinase-ABC solutions at concentrations of 0, 0.06, 0.5 and 2 U/ml. (Tukey’s multiple comparisons tests: *p < 0.05, **p < 0.01, ***p < 0.001). Table 1 Summary of T2-weighted MRI, magnetization transfer MRI (MTI) and diffusion tensor MRI (DTI) measurements in different components of the fresh ovine eyes at the physiological intraocular pressure of 10 mmHg (mean ± standard deviation).   T2 (ms) MTR FA λ// (μm2/ms) λ⊥ (μm2/ms) MD (μm2/ms) Cornea 43.3 ± 14.1 26.6% ± 4.8% 0.126 ± 0.038 1.638 ± 0.068 1.375 ± 0.121 1.463 ± 0.102 Sclera 26.5 ± 4.2 43.1% ± 8.4% 0.386 ± 0.061 1.270 ± 0.067 0.738 ± 0.078 0.915 ± 0.064 Retina 122.5 ± 18.3 15.1% ± 4.8% 0.283 ± 0.152 0.565 ± 0.053 0.380 ± 0.096 0.441 ± 0.068 Optic Nerve 51.3 ± 7.9 40.5% ± 5.3% 0.513 ± 0.105 0.452 ± 0.116 0.195 ± 0.024 0.281 ± 0.047 Lens Cortex 33.9 ± 7.7 40.0% ± 7.4% 0.597 ± 0.050 1.289 ± 0.097 0.464 ± 0.069 0.739 ± 0.075 Anterior Chamber 188.6 ± 36.9 4.0% ± 0.7% 0.063 ± 0.003 2.144 ± 0.081 1.956 ± 0.068 2.019 ± 0.072 Vitreous Body 189.4 ± 12.8 3.8% ± 0.5% 0.065 ± 0.003 2.140 ± 0.025 1.945 ± 0.017 2.010 ± 0.019 (T2: transverse relaxation time; MTR: magnetization transfer ratio; FA: fractional anisotropy; λ//: axial diffusivity; λ⊥: radial diffusivity; MD: mean diffusivity). ==== Refs Liu J. & He X. Corneal stiffness affects IOP elevation during rapid volume change in the eye . Invest Ophthalmol Vis Sci 50 , 2224 –2229 (2009 ).19151396 Kimball E. C. . Experimental scleral cross-linking increases glaucoma damage in a mouse model . Exp Eye Res 128 , 129 –140 (2014 ).25285424 Sigal I. A. . Eye-specific IOP-induced displacements and deformations of human lamina cribrosa . Investigative ophthalmology & visual science 55 , 1 –15 (2014 ).24334450 Pijanka J. K. . Changes in scleral collagen organization in murine chronic experimental glaucoma . Invest Ophthalmol Vis Sci 55 , 6554 –6564 (2014 ).25228540 Jan N. J. . Polarization microscopy for characterizing fiber orientation of ocular tissues . Biomed Opt Express 6 , 4705 –4718 (2015 ).26713188 Coudrillier B. . Biomechanics of the human posterior sclera: age- and glaucoma-related changes measured using inflation testing . Investigative ophthalmology & visual science 53 , 1714 –1728 (2012 ).22395883 Nguyen T. D. & Ethier C. R. Biomechanical assessment in models of glaucomatous optic neuropathy . Exp Eye Res (2015 ). Burgoyne C. F. , Downs J. C. , Bellezza A. J. , Suh J. K. & Hart R. T. The optic nerve head as a biomechanical structure: a new paradigm for understanding the role of IOP-related stress and strain in the pathophysiology of glaucomatous optic nerve head damage . Progress in retinal and eye research 24 , 39 –73 (2005 ).15555526 Coudrillier B. . Collagen structure and mechanical properties of the human sclera: analysis for the effects of age . Journal of biomechanical engineering 137 , 041006 (2015 ).25531905 Rada J. A. , Shelton S. & Norton T. T. The sclera and myopia . Experimental eye research 82 , 185 –200 (2006 ).16202407 Goldich Y. . Clinical and corneal biomechanical changes after collagen cross-linking with riboflavin and UV irradiation in patients with progressive keratoconus: results after 2 years of follow-up . Cornea 31 , 609 –614 (2012 ).22378112 Cassagne M. . Iontophoresis transcorneal delivery technique for transepithelial corneal collagen crosslinking with riboflavin in a rabbit model . Invest Ophthalmol Vis Sci (2014 ). Suen W. L. . Ultrasound-mediated transscleral delivery of macromolecules to the posterior segment of rabbit eye in vivo . Invest Ophthalmol Vis Sci 54 , 4358 –4365 (2013 ).23722390 Choi S. . AFM study for morphological and mechanical properties of human scleral surface . Journal of nanoscience and nanotechnology 11 , 6382 –6388 (2011 ).22121720 Doughty M. J. Observations on the ultrastructure of equatorial scleral collagen fibrils in sheep eyes . Veterinary ophthalmology 15 , 71 –80 (2012 ).22050829 Meek K. M. & Fullwood N. J. Corneal and scleral collagens–a microscopist’s perspective . Micron 32 , 261 –272 (2001 ).11006506 Brown D. J. , Morishige N. , Neekhra A. , Minckler D. S. & Jester J. V. Application of second harmonic imaging microscopy to assess structural changes in optic nerve head structure ex vivo . J Biomed Opt 12 , 024029 (2007 ).17477744 Girard M. J. . Quantitative mapping of scleral fiber orientation in normal rat eyes . Invest Ophthalmol Vis Sci 52 , 9684 –9693 (2011 ).22076988 Meek K. M. & Boote C. The organization of collagen in the corneal stroma . Exp Eye Res 78 , 503 –512 (2004 ).15106929 Meek K. M. & Boote C. The use of X-ray scattering techniques to quantify the orientation and distribution of collagen in the corneal stroma . Progress in retinal and eye research 28 , 369 –392 (2009 ).19577657 Boote C. , Dennis S. , Newton R. H. , Puri H. & Meek K. M. Collagen fibrils appear more closely packed in the prepupillary cornea: optical and biomechanical implications . Invest Ophthalmol Vis Sci 44 , 2941 –2948 (2003 ).12824235 Pijanka J. K. . Quantitative mapping of collagen fiber orientation in non-glaucoma and glaucoma posterior human sclerae . Invest Ophthalmol Vis Sci 53 , 5258 –5270 (2012 ).22786908 Lewis P. N. . Three-dimensional arrangement of elastic fibers in the human corneal stroma . Exp Eye Res 146 , 43 –53 (2016 ).26704458 Hanlon S. D. , Behzad A. R. , Sakai L. Y. & Burns A. R. Corneal stroma microfibrils . Exp Eye Res 132 , 198 –207 (2015 ).25613072 Ho L. C. . Magic angle-enhanced MRI of fibrous microstructures in sclera and cornea with and without intraocular pressure loading . Investigative ophthalmology & visual science 55 , 5662 –5672 (2014 ).25103267 Ho L. C. . In Vivo Evaluation of White Matter Integrity and Anterograde Transport in Visual System After Excitotoxic Retinal Injury With Multimodal MRI and OCT . Invest Ophthalmol Vis Sci 56 , 3788 –3800 (2015 ).26066747 Ho L. C. . In vivo assessment of aqueous humor dynamics upon chronic ocular hypertension and hypotensive drug treatment using gadolinium-enhanced MRI . Invest Ophthalmol Vis Sci 55 , 3747 –3757 (2014 ).24764067 Chan K. C. , Kancherla S. , Fan S. J. & Wu E. X. Long-term effects of neonatal hypoxia-ischemia on structural and physiological integrity of the eye and visual pathway by multimodal MRI . Invest Ophthalmol Vis Sci 56 , 1 –9 (2015 ).25491295 Chan K. C. , Fu Q. L. , Hui E. S. , So K. F. & Wu E. X. Evaluation of the retina and optic nerve in a rat model of chronic glaucoma using in vivo manganese-enhanced magnetic resonance imaging . Neuroimage 40 , 1166 –1174 (2008 ).18272401 Chan K. C. . In vivo evaluation of retinal and callosal projections in early postnatal development and plasticity using manganese-enhanced MRI and diffusion tensor imaging . Neuroimage 59 , 2274 –2283 (2012 ).21985904 Chan K. C. . In vivo visuotopic brain mapping with manganese-enhanced MRI and resting-state functional connectivity MRI . Neuroimage 90 , 235 –245 (2014 ).24394694 Chan K. C. . In vivo retinotopic mapping of superior colliculus using manganese-enhanced magnetic resonance imaging . Neuroimage 54 , 389 –395 (2011 ).20633657 Chan K. C. , So K. F. & Wu E. X. Proton magnetic resonance spectroscopy revealed choline reduction in the visual cortex in an experimental model of chronic glaucoma . Exp Eye Res 88 , 65 –70 (2009 ).18992243 Murphy M. C. . Top-down influence on the visual cortex of the blind during sensory substitution . Neuroimage 125 , 932 –940 (2016 ).26584776 Strouthidis N. G. & Girard M. J. Altering the way the optic nerve head responds to intraocular pressure-a potential approach to glaucoma therapy . Current opinion in pharmacology 13 , 83 –89 (2013 ).22999652 Mountain K. M. , Bjarnason T. A. , Dunn J. F. & Matyas J. R. The functional microstructure of tendon collagen revealed by high-field MRI . Magnetic resonance in medicine 66 , 520 –527 (2011 ).21674618 Fishbein K. W. . Effects of formalin fixation and collagen cross-linking on T2 and magnetization transfer in bovine nasal cartilage . Magn Reson Med 57 , 1000 –1011 (2007 ).17534923 Raya J. G. Techniques and applications of in vivo diffusion imaging of articular cartilage . Journal of magnetic resonance imaging: JMRI 41 , 1487 –1504 (2015 ).25865215 Henkelman R. M. , Stanisz G. J. & Graham S. J. Magnetization transfer in MRI: a review . NMR Biomed 14 , 57 –64 (2001 ).11320533 Sarman H. . Assessment of Postoperative Tendon Quality in Patients With Achilles Tendon Rupture Using Diffusion Tensor Imaging and Tendon Fiber Tracking . The Journal of foot and ankle surgery: official publication of the American College of Foot and Ankle Surgeons (2015 ). Chang E. Y. . Off-resonance saturation ratio obtained with ultrashort echo time-magnetization transfer techniques is sensitive to changes in static tensile loading of tendons and degeneration . Journal of magnetic resonance imaging: JMRI (2015 ). Szeverenyi N. M. & Bydder G. M. Dipolar anisotropy fiber imaging in a goat knee meniscus . Magnetic resonance in medicine: official journal of the Society of Magnetic Resonance in Medicine/Society of Magnetic Resonance in Medicine 65 , 463 –470 (2011 ). Yang X. . Diffusion tensor imaging for anatomical and quantitative evaluation of the anterior cruciate ligament and ACL grafts: a preliminary study . Journal of computer assisted tomography 38 , 489 –494 (2014 ).24651754 Rautiainen J. . Adiabatic rotating frame relaxation of MRI reveals early cartilage degeneration in a rabbit model of anterior cruciate ligament transection . Osteoarthritis and cartilage/OARS, Osteoarthritis Research Society 22 , 1444 –1452 (2014 ). Yao W. , Qu N. , Lu Z. & Yang S. The application of T1 and T2 relaxation time and magnetization transfer ratios to the early diagnosis of patellar cartilage osteoarthritis . Skeletal radiology 38 , 1055 –1062 (2009 ).19688346 Nguyen T. D. & Ethier C. R. Biomechanical assessment in models of glaucomatous optic neuropathy . Exp Eye Res 141 , 125 –138 (2015 ).26115620 Stewart J. M. , Schultz D. S. , Lee O. T. & Trinidad M. L. Exogenous collagen cross-linking reduces scleral permeability: modeling the effects of age-related cross-link accumulation . Invest Ophthalmol Vis Sci 50 , 352 –357 (2009 ).18708616 Coudrillier B. . Effects of Peripapillary Scleral Stiffening on the Deformation of the Lamina Cribrosa . Invest Ophthalmol Vis Sci 57 , 2666 –2677 (2016 ).27183053 Norton T. T. & Rada J. A. Reduced extracellular matrix in mammalian sclera with induced myopia . Vision Res 35 , 1271 –1281 (1995 ).7610587 Avetisov E. S. , Savitskaya N. F. , Vinetskaya M. I. & Iomdina E. N. A study of biochemical and biomechanical qualities of normal and myopic eye sclera in humans of different age groups . Metab Pediatr Syst Ophthalmol 7 , 183 –188 (1983 ).6678372 Tezel G. , Edward D. P. & Wax M. B. Serum autoantibodies to optic nerve head glycosaminoglycans in patients with glaucoma . Arch Ophthalmol 117 , 917 –924 (1999 ).10408457 Danilov N. A. . Stabilization of scleral collagen by glycerol aldehyde cross-linking . Biochim Biophys Acta 1780 , 764 –772 (2008 ).18328828 Grytz R. . Material properties of the posterior human sclera . J Mech Behav Biomed Mater 29 , 602 –617 (2014 ).23684352 Sigal Ian A. , Jan Ning-Jiun & Saundria Moed. A microstructural basis for nonlinear effects of IOP on the lamina cribrosa and sclera . In ARVO (Denver, Colorado, 2015 ). Glass D. H. , Roberts C. J. , Litsky A. S. & Weber P. A. A viscoelastic biomechanical model of the cornea describing the effect of viscosity and elasticity on hysteresis . Invest Ophthalmol Vis Sci 49 , 3919 –3926 (2008 ).18539936 Vaghefi E. & Donaldson P. J. An exploration into diffusion tensor imaging in the bovine ocular lens . Frontiers in physiology 4 , 33 (2013 ).23459990 Lizak M. J. , Datiles M. B. , Aletras A. H. , Kador P. F. & Balaban R. S. MRI of the human eye using magnetization transfer contrast enhancement . Invest Ophthalmol Vis Sci 41 , 3878 –3881 (2000 ).11053289 Laurent D. . Quantitative and qualitative assessment of articular cartilage in the goat knee with magnetization transfer imaging . Magn Reson Imaging 19 , 1279 –1286 (2001 ).11804755 Fullerton G. D. & Rahal A. Collagen structure: the molecular source of the tendon magic angle effect . J Magn Reson Imaging 25 , 345 –361 (2007 ).17260393 Winkler M. . Three-dimensional distribution of transverse collagen fibers in the anterior human corneal stroma . Investigative ophthalmology & visual science 54 , 7293 –7301 (2013 ).24114547 Cone-Kimball E. . Scleral structural alterations associated with chronic experimental intraocular pressure elevation in mice . Molecular vision 19 , 2023 –2039 (2013 ).24146537 Nguyen C. . Studies of scleral biomechanical behavior related to susceptibility for retinal ganglion cell loss in experimental mouse glaucoma . Invest Ophthalmol Vis Sci 54 , 1767 –1780 (2013 ).23404116 DM M. The Cornea and Sclera (New York, Academic Press, 1984 ). Gray M. L. , Burstein D. , Lesperance L. M. & Gehrke L. Magnetization transfer in cartilage and its constituent macromolecules . Magn Reson Med 34 , 319 –325 (1995 ).7500869 Boubriak O. A. , Urban J. P. , Akhtar S. , Meek K. M. & Bron A. J. The effect of hydration and matrix composition on solute diffusion in rabbit sclera . Exp Eye Res 71 , 503 –514 (2000 ).11040086 Murienne B. J. , Jefferys J. L. , Quigley H. A. & Nguyen T. D. The effects of glycosaminoglycan degradation on the mechanical behavior of the posterior porcine sclera . Acta biomaterialia 12 , 195 –206 (2015 ).25448352 Chen C. T. . Matrix fixed-charge density as determined by magnetic resonance microscopy of bioreactor-derived hyaline cartilage correlates with biochemical and biomechanical properties . Arthritis and rheumatism 48 , 1047 –1056 (2003 ).12687548 Lammentausta E. . T2 relaxation time and delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) of human patellar cartilage at 1.5T and 9.4T: Relationships with tissue mechanical properties . Journal of orthopaedic research: official publication of the Orthopaedic Research Society 24 , 366 –374 (2006 ).16479569 Nieminen M. T. . Prediction of biomechanical properties of articular cartilage with quantitative magnetic resonance imaging . Journal of biomechanics 37 , 321 –328 (2004 ).14757451 Zhang Y. . Lamina-specific anatomic magnetic resonance imaging of the human retina . Investigative ophthalmology & visual science 52 , 7232 –7237 (2011 ).21828153 Stachs O. . High resolution MR imaging of the human eye and orbit at 7T using a dedicated 6-channel eye coil - impact of an optoacoustic triggered MR protocol on imaging quality . Investigative ophthalmology & visual science 55 , 5845 –5845 (2014 ). Ho L. C. . Non-invasive assessments of biomechanical and biochemical properties in animal and human eyes using multi-modal MRI. In Proc Int Soc Magn Reson Med Vol. 24 (in press) (2016). Girard M. , Suh J. K. , Hart R. T. , Burgoyne C. F. & Downs J. C. Effects of storage time on the mechanical properties of rabbit peripapillary sclera after enucleation . Current eye research 32 , 465 –470 (2007 ).17514532 Kling S. , Remon L. , Perez-Escudero A. , Merayo-Lloves J. & Marcos S. Corneal biomechanical changes after collagen cross-linking from porcine eye inflation experiments . Investigative ophthalmology & visual science 51 , 3961 –3968 (2010 ).20335615 Jan N. J. . Measuring Micron-Scale Collagen Fiber Orientation in the Lamina Cribrosa and Peripapillary Sclera . In Investigative Ophthalmology and Visual Science 54 (15), 65 (2013 ). Yu Y. & Chau Y. Formulation of in situ chemically cross-linked hydrogel depots for protein release: from the blob model perspective . Biomacromolecules 16 , 56 –65 (2015 ).25314589 Leung C. K. , Yu Y. , Xu G. , Biswas S. & Y. C. A Novel Glaucoma Model Induced by Intracameral Injection of Cross-Linking Hydrogel . In Invest Ophthalmol Vis Sci 55 , 2087 (2014 ). Sigal I. A. , Flanagan J. G. , Tertinegg I. & Ethier C. R. Reconstruction of human optic nerve heads for finite element modeling . Technology and health care: official journal of the European Society for Engineering and Medicine 13 , 313 –329 (2005 ).16055979

PMC005xxxxxx/PMC5000080.txt

==== Front J Biol ChemJ. Biol. ChemjbcjbcJBCThe Journal of Biological Chemistry0021-92581083-351XAmerican Society for Biochemistry and Molecular Biology 11200 Rockville Pike, Suite 302, Rockville, MD 20852-3110, U.S.A. M115.69065110.1074/jbc.M115.690651Protein Structure and FoldingMechanism of Histone H3K4me3 Recognition by the Plant Homeodomain of Inhibitor of Growth 3* Histone Recognition by the ING3 PHD FingerKim Sophia ‡1Natesan Senthil ‡Cornilescu Gabriel §Carlson Samuel ‡Tonelli Marco §McClurg Urszula L. ¶234Binda Olivier ¶45Robson Craig N. ¶34Markley John L. §Balaz Stefan ‡http://orcid.org/0000-0002-2761-733XGlass Karen C. ‡6From the ‡ Department of Pharmaceutical Sciences, Albany College of Pharmacy and Health Sciences, Colchester, Vermont 05446, the § National Magnetic Resonance Facility at Madison and Department of Biochemistry, University of Wisconsin-Madison, Madison, Wisconsin 53706, and the ¶ Newcastle Cancer Centre at the Northern Institute for Cancer Research, Newcastle University, Newcastle Upon Tyne NE2 4HH, United Kingdom6 Supported in part by National Institutes of Health Grant 1R15GM104865 from NIGMS. To whom correspondence should be addressed: Dept. of Pharmaceutical Sciences, Albany College of Pharmacy and Health Sciences, 261 Mountain View Dr., Colchester, VT 05446. Tel.: 802-735-2636; Fax: 802-654-0716; E-mail: karen.glass@acphs.edu.1 Recipient of an Albany College of Pharmacy and Health Sciences student summer research award. 2 Supported in part by Newcastle Healthcare Charity. 3 Supported in part by Prostate Cancer United Kingdom Grant PG09-23. 4 Supported in part by Joint Research Executive Scientific Committee JG/ML/0414 and Cancer Research United Kingdom C27826/A15994. 5 Supported by the Newcastle's Biomedical Fellowship Programme, which is in part funded through the Wellcome Trust's Institutional Strategic Support Fund, and by Breast Cancer Campaign Charity Grant 2013MaySP005. 26 8 2016 8 6 2016 8 6 2016 291 35 18326 18341 4 9 2015 7 6 2016 © 2016 by The American Society for Biochemistry and Molecular Biology, Inc.2016The American Society for Biochemistry and Molecular Biology, Inc.Author's Choice—Final version free via Creative Commons CC-BY license.Aberrant access to genetic information disrupts cellular homeostasis and can lead to cancer development. One molecular mechanism that regulates access to genetic information includes recognition of histone modifications, which is carried out by protein modules that interact with chromatin and serve as landing pads for enzymatic activities that regulate gene expression. The ING3 tumor suppressor protein contains a plant homeodomain (PHD) that reads the epigenetic code via recognition of histone H3 tri-methylated at lysine 4 (H3K4me3), and this domain is lost or mutated in various human cancers. However, the molecular mechanisms targeting ING3 to histones and the role of this interaction in the cell remain elusive. Thus, we employed biochemical and structural biology approaches to investigate the interaction of the ING3 PHD finger (ING3PHD) with the active transcription mark H3K4me3. Our results demonstrate that association of the ING3PHD with H3K4me3 is in the sub-micromolar range (KD ranging between 0.63 and 0.93 μm) and is about 200-fold stronger than with the unmodified histone H3. NMR and computational studies revealed an aromatic cage composed of Tyr-362, Ser-369, and Trp-385 that accommodate the tri-methylated side chain of H3K4. Mutational analysis confirmed the critical importance of Tyr-362 and Trp-385 in mediating the ING3PHD-H3K4me3 interaction. Finally, the biological relevance of ING3PHD-H3K4me3 binding was demonstrated by the failure of ING3PHD mutant proteins to enhance ING3-mediated DNA damage-dependent cell death. Together, our results reveal the molecular mechanism of H3K4me3 selection by the ING3PHD and suggest that this interaction is important for mediating ING3 tumor suppressive activities. epigeneticshistone methylationisothermal titration calorimetry (ITC)nuclear magnetic resonance (NMR)PHD fingerInhibitor of Growth 3histone acetyltransferaseplant homeodomainNational Institutes of Health http://dx.doi.org/10.13039/1000000021R15GM104865P41GM103399S10RR02781S10RR08438S10RR023438S10RR025062S10RR029220Breast Cancer Campaign http://dx.doi.org/10.13039/5011000003012013MaySP005Prostate Cancer UK http://dx.doi.org/10.13039/501100000771PG09–23Joint Research Executive Scientific Committee JG/ML/0414Cancer Research UK http://dx.doi.org/10.13039/501100000321Cys-27826/A15994National Science Foundation http://dx.doi.org/10.13039/100000001DMB-8415048OIA-9977486BIR-9214394 ==== Body Introduction Although the human genome was sequenced over a decade ago and its structure has been investigated for nearly half a century, the molecular mechanisms that regulate access to genetic information remain largely unknown. One recently identified mechanism is based on the physical association of chromatin reader domains with the histone scaffolding proteins that condense the genome within the nucleus of the cell. Essentially, the DNA strands of the genome are spooled around histone octamers (two copies of each histone H2A, H2B, H3, and H4) to form nucleosomes, the basic unit of chromatin. These histone proteins harbor an N-terminal tail that protrudes outside of the core nucleosome (1). The histone tails are readily available for post-translational modifications, such as lysine acetylation and methylation. Methyltransferases modify lysine in a stepwise manner to generate mono- (Kme1), di- (Kme2), or tri-methylated (Kme3) lysine. Each methylation state can be recognized by specific histone mark reader modules and thus lead to divergent biological outcomes. Specifically, reader domains serve as bridges between chromatin and enzymatic activities that open or close the structure of the chromatin, thereby regulating access to genetic information and control gene expression. Aberrant access to genetic information leads to pathologies ranging from cancer to neurological disorders. Several histone mark readers have been described over the last 15 years, with the chromodomain of the heterochromatin protein HP1α that binds to histone H3 tri-methylated on lysine 9 (H3K9me3) (2) as a prime example. Since then, other reader domains have been discovered, including H3K9me3 readers (ADD, WD40, plant homeodomain, and chromodomains); H3K36me3 readers (tudor, chromobarrel, and PWWP domains); H4K20me2 readers (tandem tudor domain and BAH domain); and H3K4me3 readers (plant homeodomains, tandem tutor domains, double chromodomains, and the zf-CW domain) (3, 4). The plant homeodomain (PHD7 finger) is found in several nuclear proteins, but their function was first described in the INhibitor of Growth (ING) family of tumor suppressors. The founding member, p33ING1b, was cloned in a genetic suppressor element screen (5). Then, ING2 (6), ING3 (7, 8), ING4 (9–11), and ING5 (9) were identified, essentially based on sequence similarities with p33ING1b. The PHD finger of ING proteins was shown to mediate interactions with H3K4me3 at the transcriptional start site of genes to stabilize enzymatic activities that regulate access to genetic information (12). Specifically, ING4 binds H3K4me3 at the transcriptional start site to recruit histone acetyltransferase activity and facilitate access to genetic information (13), whereas ING2 bridges histone deacetylase activity to H3K4me3 at transcriptional start site to silence gene expression (14). ING proteins are broadly lost or mutated in various types of human cancers (15, 16). Overexpression studies suggest their involvement in preventing cellular proliferation, while enhancing cell contact inhibition and DNA damage-induced cell death (12). A comprehensive biochemical characterization of ING proteins demonstrated that p33ING1b and ING2 associate with the mSIN3A-HDAC1 histone deacetylase complex (17, 18), whereas ING3, ING4, and ING5 are found in histone acetyltransferase (HAT) complexes (17). Specifically, ING3 associates with the TIP60 (Tat-interacting protein of 60 kDa) complex (17, 19), ING4 with the HBO1-JADE complex (17), and ING5 with both HBO1-JADE and MOZ-BRPF complexes (17). Thus, p33ING1b and ING2 are believed to primarily function as transcriptional repressors via histone deacetylase activity, whereas ING3, ING4, and ING5 are mainly involved in transcriptional activation via associated HAT activities. ING3 functions in the multiprotein TIP60 HAT complex, which also includes the ATPases EP400, RUVBL1 (Pontin), and RUVBL2 (Reptin), TRRAP, the bromodomain protein BRD8, the polycomb proteins EPC1 and EPCys-2, the DNA methyltransferase-associated protein DMAP1, actin, and the actin-like protein BAF53A (ACTL6A), the mortality factor MRG15 (MORF4L1), MEAF6, and GAS41 (YEATS4) (17). Interestingly, like ING4 (13), TIP60, RUVBL1, and RUVBL2 are involved in the DNA damage response (20, 21). Moreover, the RUVBL1 and RUVBL2 proteins in the ING3 complex regulate hypoxia signaling much the same as observed with ING4 (22–25), suggesting possible redundant or complementary functions between ING3 and ING4 in the regulation of cellular stress responses. Finally, the components of the ING3 complex described above were recently found to associate with the histone chaperon ANP32E and the histone variant H2AZ (26). Although the role of ING3 in histone exchange remains elusive, ING3 is required along with EPC1 for acetylation of nucleosomes by TIP60 (19). Importantly, ING2 was found to interact with histone H3. Specifically, the PHD of ING2 binds to H3K4me3 via a conserved aromatic cage (14, 27). The PHD finger-H3K4me3 interaction is critical for both ING2 (14) and ING4 (13) to associate with transcriptional start sites and repress or activate gene expression, respectively. Furthermore, the ING4PHD-H3K4me3 interaction was demonstrated to be essential for ING4 tumor suppressive activities, including enhancing DNA damage-induced cell death, inhibition of cellular proliferation, and colony formation (13). Thus, the ability of the PHD finger domain to bridge ING proteins to H3K4me3 is thought to be critical for the tumor suppressor activity of all ING family proteins. The ING3 locus is lost or mutated in several human cancers. Notably, frequent loss of heterozygosity is detected at the ING3 locus, 7q31, in invasive epithelial ovarian carcinomas (28, 29), prostate (30), colorectal (31), as well as human head and neck cancers (8). The 7q31 region contains four candidate tumor suppressor genes, CAV1, CAVal-2, ST7, and ING3. Because mutations in the PHD domain of ING3 are also reported in the genomes of various cancers (Cys-376 (frameshift), D380H, and H387P (33)), we decided to investigate the molecular mechanisms that regulate the association of ING3 with histone post-translational modifications. ITC and nuclear magnetic resonance (NMR) experiments demonstrate that the ING3PHD selects for H3K4me3 > H3K4me2 > H3K4me1 > unmodified histone H3 and identified residues critical for ligand coordination. These results are further supported by molecular dynamic (MD) simulation studies of the ING3PHD with modified and unmodified histone H3 and H4 peptides. The trajectory analysis of our MD simulations revealed a conserved aromatic cage composed of residues Tyr-362, Ser-369, and Trp-385, which imparts the affinity and specificity for histone H3 methylated at lysine 4. A structural comparison of the time-averaged structure from the MD simulation of the ING3PHD-H3K4me3 complex with x-ray crystal structures of the other ING family member proteins bound to H3K4me3 revealed that the mechanism of histone ligand binding is universally conserved within this protein family. Furthermore, we found that full-length ING3 proteins defective in H3K4me3 recognition are still able to form a complex with the TIP60 HAT, and cell-based assays show that the ING3-H3K4me3 interaction is required for DNA damage-induced cell death. These data illustrate for the first time that histone recognition by the PHD finger region of ING3 is crucial for its activity as a candidate tumor suppressor protein. Interestingly, increased copy number of ING3 along with overexpression and deregulation of ING3 have recently been linked to poor outcomes in prostate cancer patients and castrate-resistant prostate cancer cell lines (34–36). This suggests that ING3 may suppress tumor formation in some cases while promoting cancer in others. Together, our results illustrate the structure and function of the ING3 PHD finger domain in histone recognition and in regulating the biological activity of the ING3 protein, which will be important for the development of new epigenetic therapies aimed a modulating the role of ING3 in disease. Results ING3PHD Recognizes Methylated Histone H3 The ING3 subunit of the TIP60 HAT complex contains a C-terminal PHD finger. PHD finger domains are generally known to recognize methylated lysine on histone tails, and the closely related ING4 and ING5 tumor suppressor proteins have been shown to recognize H3K4me3 through their PHD domains (13, 37, 38). To determine the binding specificity and affinity of the ING3PHD, we used a combination of biochemical and biophysical methods to screen the PHD finger against a variety of methylated and unmodified histone tail peptides. We carried out tryptophan fluorescence experiments to test the binding of ING3PHD to histone peptides, H3K4me3 (residues 1–12), H3K4me2 (residues 1–12), H3K4me1 (residues 1–12), H3 unmodified (residues 1–12), and H4 unmodified (residues 1–12). Tryptophan fluorescence is an ideal method to investigate the ING3PHD-histone interactions because the binding pocket of the ING3PHD contains two tryptophan residues involved in histone coordination. As seen with other ING PHD finger proteins, ING3PHD preferably recognized histone H3 that was tri-methylated on lysine 4 (KD = 0.63 μm), and the histone peptide binding affinity decreased in conjunction with the methylation status of lysine 4 (KD = 4.05 μm for H3K4me2 and 21.45 μm for H3K4me1) (Table 1). Unmodified histone H3 bound the weakest with a binding coefficient of 131.6 μm, and no binding was detected between the ING3PHD and histone H4. TABLE 1 Dissociation constants of the ING3 PHD finger with post-translationally modified histone peptides as measured by tryptophan fluorescence and ITC Histone peptide Sequence Trp fluorescence KD ITC KD μm μm H3K4me3(1–12) ARTKme3QTARKSTG 0.63 ± 0.11 0.93 ± 0.04 H3K4me2(1–12) ARTKme2QTARKSTG 4.05 ± 0.53 2.99 ± 0.33 H3K4me1(1–12) ARTKme1QTARKSTG 21.45 ± 3.51 23.24 ± 0.80 H3 unmodified(1–12) ARTKQTARKSTG 131.57 ± 13.48 180.62 ± 19.17 H4 unmodified(4–17) GKGGKGLGKGGAKR NAa No binding a NA, not available. The dissociation constants of the ING3PHD with the histone H3 and histone H4 peptides were also analyzed by ITC experiments (Fig. 1), and the results are included in Table 1. The KD values determined from the ITC titration data confirmed that the H3K4me3 bound to the ING3PHD with the highest affinity (0.93 μm), followed by H3K4me2 (2.99 μm), H3K4me2 (23.24 μm), and H3 unmodified (180.6 μm), consistent with the tryptophan fluorescence data shown in Table 1. No binding was observed between ING3PHD and the unmodified histone H4 peptide. Our results demonstrate that the ING3 PHD domain preferentially binds to H3K4me3, consistent with the other ING PHD finger proteins (13, 14, 27, 37, 39). FIGURE 1. ITC measurement of the interaction between the wild-type ING3 PHD finger and methylated or unmodified histone peptides. A–F, exothermic ITC enthalpy plots for the binding of the ING3 PHD finger to H3K4me3, H3K4me2, H3K4me1, H3 unmodified, H3K9me3, and H4 unmodified. The calculated binding constants are indicated. Chemical Shift Mapping of the ING3 Binding Pocket To outline the specific interactions between the histone peptide ligands and the ING3PHD binding pocket, we carried out nuclear magnetic resonance (NMR) experiments. The backbone assignments of the ING3PHD finger were obtained from the 15N,13C double-labeled ING3PHD using the ADAPT-NMR program at the NMRFAM facility in Madison, WI, which allowed for rapid data collection and assignment of the NMR spectra (Fig. 2A). To confirm recognition of the histone tail peptides observed by tryptophan fluorescence and ITC, the 1H-15N heteronuclear single quantum coherence (HSQC) spectra of the uniformly 15N-labeled ING3 PHD finger were recorded in the absence and presence of the following histone tail peptides: unmodified histone H3 (residues 1–12), unmodified histone H4 (residues 4–17), and H3K4me1, H3K4me2, and H3K4me2 (residues 1–12). Chemical shift perturbations were induced in the ING3PHD upon addition of unmodified H3, H3K4me1, H3K4me2, and H3K4me3 (Fig. 2B). No significant resonance shifts were observed upon the addition of unmodified histone H4 peptide. This pattern of histone recognition is similar to the recognition of the histone tails by other ING PHD finger proteins (13, 37) and confirms a strong interaction of the ING3PHD domain with methylated histone H3 tails. FIGURE 2. Interaction of the ING3 PHD finger with histone ligands. A, two-dimensional 1H-15N HSQC spectra of 15N-labeled ING3 PHD finger with the complete HSQC assignments labeled. B, superimposed 1H-15N HSQC spectra of the 0.5 mm ING3 PHD finger, collected during titrating in the indicated histone peptides. The spectra are color-coded according to the protein/peptide ratio. C, histogram of normalized 1H-15N chemical shift changes in backbone amides of the ING3 PHD finger upon addition of the H3K4me3 peptide. Chemical shift changes were from 0.2 to 0.3 ppm (yellow), from 0.3 to 0.4 ppm (orange), and >0.4 ppm (red). D, mapping of residues exhibiting significant resonance perturbations upon addition of the H3K4me3 ligand onto the surface of the NMR structure of the ligand-free ING3 PHD finger (PDB code 1X4I). The residues in the binding pocket are colored red, orange, and yellow depending on the magnitude of the chemical shift change upon ligand addition as in C. The chemical shift perturbations observed upon binding of the H3K4me3 histone peptide were used to map the binding pocket of the ING3 PHD finger. The normalized changes in chemical shift from the NMR HSQC spectrum were plotted as a bar graph to show the amino acid residues most affected by addition of the H3K4me3 histone tail peptide in a 1:2.61 ING3PHD-to-peptide ratio. The largest chemical shifts (changes greater than 0.4 ppm) are shown in red, and changes of >0.3 or >0.2 ppm are shown in orange and yellow, respectively (Fig. 2C). Nine amino acids within the ING3PHD showed large chemical shift changes upon addition of the H3K4me3 histone ligand, including Tyr-8, Tyr-16, Glu-18, Val-20, Cys-22, Ile-29, Glu-30, Trp-31, and Lys-46, indicating that these residues are directly or indirectly involved in ligand binding. Mapping the position of these amino acids onto the surface of the native ING3PHD structure (PDB code 1X4I) reveals that there are two adjacent ligand coordination regions on the surface of the PHD finger (Fig. 2D). The first group of residues with large chemical shift changes is clustered around Trp-31, which makes up the side of the aromatic cage responsible for coordination of the tri-methylated lysine in other PHD finger proteins, including ING4 and ING5 (13, 37, 40). The second group of residues is located near Trp-47 and comprises a potential binding site for the N terminus of the histone H3 tail, which in the ING5PHD (PDB code 3C6W) structure was a region critical for ligand recognition and coordination (37). Notably, the residues showing the largest chemical shift changes in the PHD binding pocket are conserved with the ligand-coordinating residues observed in the ING4 and ING5 crystal structures bound to H3K4me3 (13, 37, 40) (also see the sequence alignment in Fig. 3A). These include Trp-47, Tyr-8, and Ser-15, which make up the aromatic cage in the ING5PHD structure (37). Residue Cys-22, which shows the largest chemical shift change of any residue, is one of the conserved cysteines responsible for coordinating one of the two zinc ions in the ING3PHD finger. In the ING5PHD structure, the residue conserved with Cys-22 (Cys-202) makes an important hydrogen bond contact to Arg-2 in the H3K4me3 histone tail peptide, and our NMR chemical shift data (in combination with our MD simulation data, Fig. 3B) indicate that Cys-22 likely makes a large conformational change moving from a more buried position to the surface of the ING3PHD binding pocket upon ligand binding. FIGURE 3. Characterization of specific ING3 PHD finger-histone ligand interactions. A, domain architecture of the full-length human ING3 protein with the sequence alignment of the PHD finger domains of ING1–5. Residues 1–418 of human ING3 are shown with the N-terminal leucine zipper-like domain (LZL) and novel conserved region (NCR), as well as the nuclear localization signal (NLS) and C-terminal PHD. Sequence alignment of the PHD finger domains of ING1–5 were aligned to the ING3PHD construct used in this study (residues 4–57, corresponding to residues 360–409 in full-length ING3). The conserved Cys4-His-Cys3 PHD finger motif residues involved in zinc coordination are highlighted in blue; the aromatic cage residues are highlighted in orange, and residues mutated in this study are colored in red. B, time-averaged structure of the ING3 PHD finger in complex with the H3K4me3 peptide ligand obtained using MD simulation. PyMOL (32) was used to depict the ING3 PHD finger, shown in green, and the peptide ligand is colored in blue. Hydrogen bonds and salt bridges are indicated by red dotted lines. C, structural alignment of the ING1–5 PHD finger proteins in complex with the H3K4me3 histone peptide. Structures of the ING PHD complexes were taken from the Protein Data Bank and aligned in PyMOL (32). The structures include ING1 (2QIC, orange), ING2 (2G6Q, blue), ING3 (MD simulation, green), ING4 (2PNX, magenta), and ING5 (3C6W, yellow). D, surface representation of the time-averaged structure of the ING3PHD-H3K4me3 complex from the MD simulation showing specific point mutations introduced into the binding pocket by site-directed mutagenesis, with the H3K4me3 histone ligand in the binding pocket shown in blue. E, circular dichroism spectra in the far-UV region of the ING3 PHD finger wild-type and mutant proteins. The percent α-helical content of each protein is listed in the inset. Molecular Dynamic (MD) Simulations of Histone Ligand Binding Because our attempts to crystallize the H3K4me3 histone peptide in complex with the ING3 PHD finger were unsuccessful, we used MD simulations to further investigate the ING3PHD-histone ligand interactions. The simulations were carried out for the experimentally studied peptides bound to ING3 to correlate differences in the binding affinities measured among histone ligands and specific molecular interactions observed in the protein-ligand complexes. All the peptides used in MD simulations were protonated on their N terminus (NH3+) and amidated on the C terminus to represent the structures of the peptides used in the in vitro experiments. The initial MD simulation structures of the methylated histone H3 peptides in complex with the ING3PHD protein were obtained by flexible docking of each of the four histone peptides listed in Table 1 to the ING3PHD using the FlexPepDock program (41, 42) from Rosetta commons (43). The initial approximate binding poses for FlexPepDock were taken from the x-ray crystal structures of the ING4PHD and ING5PHD proteins in complex with H3K4me3 peptides (PDB codes 2PNX and 3C6W, respectively). Both ING4PHD and ING5PHD share very high sequence and structural identities with the ING3PHD, especially among the binding site residues (Fig. 3A) (40). Because of the absence of experimental structures for related PHD finger proteins bound to the unmodified histone H4 peptide, the initial FlexPepDock binding pose for this ligand was obtained by docking histone H4 using the Hex webserver (44), which does not require an initial approximate bound conformation of the peptide. Then docking of the histone H4 peptide was refined in FlexPepDock, so that the top scored complex of each peptide generated with FlexPepDock could be used as the starting point for MD simulations. The MD simulations, which were run using the Amber14 package for 10 ns under isobaric and isothermal conditions, provided detailed insights into the molecular interactions between the ING3PHD and its histone ligands (Tables 2 and 3). The ING3PHD domain is a typical PHD finger fold with two structural Zn2+ atoms that are coordinated by a conserved pattern of cysteine and histidine residues (Zn1-CCCC and Zn2-CCCH) with tetrahedral geometry. Coordination of the zinc ions by the Cys4-His-Cys3 consensus sequence stabilizes the three loop regions of the PHD finger and is essential for folding of secondary structures, including the double-stranded antiparallel β-sheet that forms the core of the PHD domain (Fig. 3B). The structural integrity and secondary structural elements of the protein were maintained well throughout the simulations, which was ascertained by the coordination geometries measured between the Zn2+ cations and interacting residue atoms, and the secondary structure analysis (data not shown). TABLE 2 ING3-histone peptides binding energies by free energy calculations No. Peptide Binding free energies and its components (kcal/mol) ΔSa MM/GBSAbb MM/PBSAbb GBSA ΔG ΔGc PBSA ΔG ΔGc kcal/mol 1 H3K4me3 −42.358 −70.332 −27.975 −65.248 −22.890 2 H3K4me2 −40.958 −62.682 −21.724 −61.206 −20.249 3 H3K4me1 −42.956 −60.704 −17.749 −61.068 −18.112 4 H3K4me0 −38.316 −56.285 −17.970 −54.131 −15.816 5 H4-unmodified −39.460 −42.744 −3.285 −43.705 −4.245 a Rotational, translational, and vibrational entropy was from normal mode analysis. b Molecular Mechanics/Generalized Born Surface Area and Molecular Mechanics/Poisson-Boltzmann Surface Area Methods are described in Ref. 45. c Difference between GBSA or PBSA ΔG and ΔS is shown. TABLE 3 Hydrogen bond interactions between the ING3 PHD finger and histone peptides from MD simulations Histone peptide Interacting residues % present Average bond length Average bond angle Peptide ING3 Å ° H3K4me3 K4me3@N-H Met-19@O 99.96 2.95 159.85 K4me3@O Met-19@N-H 98.44 2.93 153.88 Arg-8@O Gly-17@N-H 99.46 2.96 159.71 Thr-6@N-H Gly-17@O 84.08 3.09 150.95 Arg-2@O Gly-21@N-H 99.10 2.99 155.59 Arg-2@NH2-HH21 Cys-22@O 69.38 3.02 141.96 Arg-2@NH2-HH22 Glu-30@OE2 50.86 2.89 150.60 Arg-2@NH1-HH11 Glu-30@OE2 50.64 2.93 148.85 Arg-2@NH1-HH12 Glu-30@OE1 46.54 2.86 152.13 Arg-8@NE-HE Ser-15@O 85.44 2.97 145.35 H3K4me2 K4me2@N-H Met-19@O 99.86 3.01 157.29 K4me2@O Met-19@N-H 97.06 2.94 150.87 Arg-8@O Gly-17@N-H 99.46 2.95 159.54 Arg-2@O Gly-21@N-H 99.34 2.98 154.54 Arg-2@NE-HE Asp-23@OD1 77.36 2.88 149.88 Arg-2@NH2-HH21 Asp-23@OD1 77.10 2.84 148.06 Thr-6@N-H Gly-17@O 78.76 2.93 151.51 H3K4me1 K4me2@N-H Met-19@O 99.88 2.98 157.16 K4me2@O Met-19@N-H 98.24 2.93 154.48 Arg-2@NE-HE Asp-23@OD1 72.02 2.86 150.64 Arg-2@O Gly-21@N-H 99.22 3.02 155.08 Arg-2@NH1-HH12 Glu-30@OE1 77.96 2.83 153.52 Arg-2@NH2-HH22 Glu-30@OE1 76.02 2.98 143.27 Arg-2@NH2-HH21 Asp-23@OD1 64.78 2.94 142.04 Arg-8@O Gly-17@N-H 58.44 3.04 159.48 Thr-6@N-H Gly-17@O 56.62 3.01 149.52 H3 unmodified Lys-4@NH Met-19@O 99.36 2.95 155.64 Lys-4@O Met-19@N-H 92.94 2.98 155.10 Arg-2@O Gly-21@N-H 98.50 2.99 152.15 Arg-8@O Gly-17@N-H 97.54 2.99 158.64 Arg-2@NE-HE Asp-23@OD1 72.08 2.89 149.81 Arg-2@NH2-HH21 Asp-23@OD1 69.96 2.86 147.77 Thr-6@N-H Gly-17@O 59.72 3.12 151.08 H4 unmodified Gly-3@N-H Ser-6@O 83.62 3.06 151.85 Gly-4@O Tyr-8@N-H 49.70 3.21 142.92 Gly-3@N-H Ser-3@O 42.78 2.96 151.46 The MD simulation data were processed using molecular mechanics/Poisson-Boltzmann surface area (MM/PBSA) and Generalized Born surface area (MM/GBSA) formalisms (45), utilizing the normal-mode analysis to obtain rotational, translational, and vibrational entropy (46). The resulting binding free-energy values (ΔG) and their components are summarized for the histone peptides in Table 2. The relative energies calculated for the histone peptides (H3K4me3, H3K4me2, H3K4me1, H3 unmodified, and H4 unmodified) show good agreement with the experimental NMR, ITC, and Trp fluorescence binding data. The most favorable binding energy was observed for the H3K4me3 peptide, followed by the H3K4me2, H3K4me1, and H3 unmodified peptides, respectively. The non-binding histone H4 unmodified peptide correctly generated a relatively weak binding energy. The trajectories from each of 10-ns MD simulations (a total of 5000 snapshots or frames, each representing a 2-ps time interval) were analyzed for the presence of H-bond interactions between the ING3PHD protein and the histone peptide residues. The strength and geometries were quantitatively described in terms of bond length and donor-H-acceptor bond angles as well as the longevities of the H-bond as % of frames (5000 frames represent 100%), in which the bonds were seen. In the time-averaged structure of the ING3PHD-H3K4me3 complex from the MD simulation (Fig. 3B and supplemental data file ING3PHD-H3K4me3-MDsimulation.pdb), the histone peptide binds to the PHD domain forming a third antiparallel β-strand that pairs with the central β-sheet core described above (Fig. 3C). The most common interactions seen with all of the methylated histone H3 peptides (mono-, di-, and tri-methylated and non-substituted lysines) and the ING3PHD are intra-β-strand backbone contacts between the Lys-4 residues of the histone peptides and Met-19 in ING3PHD. Table 3 summarizes the hydrogen bond interactions seen at least 40% of the time throughout the production phase of MD simulation. The average bond lengths and bond angles of these H-bonds are also summarized in Table 3. For example, the H-bonds between backbone atoms of Met-19 and Gly-21 of ING3 and the histone H3 peptide residues Lys-4 and Arg-2 occur for nearly the entire 10-ns trajectory of all MD simulations (92.9–99.9%). These conserved intra-β-sheet hydrogen bonds facilitate the important cation-π interactions exhibited by the methylated histone peptides and residues Trp-31 and Tyr-8 in the ING3PHD. As in other ING PHD finger proteins, the H3K4me3 peptide binds in a deep and extensive pocket that consists of two large grooves connected by a narrow channel (37, 40). The tri-methylated Lys-4 side chain fits into one groove containing residues Trp-31, Tyr-8, and Ser-15 identified from the NMR chemical shift perturbation and MD simulation data (Figs. 2D and 3B). Fig. 3B, highlights the hydrogen bond interactions observed between amino acid side chains of the histone H3K4me3 peptide and the ING3PHD. For instance, the side chain amino group of Arg-2 in the histone peptide makes a constant H-bond interaction with the side chain carboxylic group of Glu-30 of ING3PHD, which was observed throughout the entire 10-ns simulation time (Table 3). Table 3 includes the details of the Arg-2 interaction from the MD simulations and also reveals several strong H-bond interactions made by the Thr-6 and Arg-8 residues of H3K4me3. Molecular Mechanism of Methyl-lysine Coordination Our NMR chemical shift perturbation experiments and MD simulation data of the ING3PHD-H3K4me3 ligand complex suggest that several amino acids in the ING3PHD binding pocket are crucial for histone recognition and coordination of the tri-methylated lysine moiety. To further investigate the contribution of specific amino acids to ligand binding, we carried out site-directed mutagenesis on ING3PHD residues and measured their effect on ligand binding affinity with ITC experiments. We selected ING3PHD residues Trp-47, Ser-15, Pro-43, Tyr-8, and Trp-31 for mutagenesis to alanine because of their proximity to the binding pocket and/or direct involvement in H3K4me3 coordination (Fig. 3D). As shown in Table 4 and Fig. 4, the W31A mutation had the largest effect on histone peptide binding, resulting in no recognition of this ligand. This is likely due to the importance of residue Trp-31 in formation of the aromatic cage in the ING3PHD as well as cation-π interactions between the side chain of this residue and the trimethylammonium group of Lys-4. The W47A and Y8A mutations reduced the ING3PHD binding affinity from 0.93 μm for the WT protein to 48 and 46 μm in the mutant proteins, respectively. This observation is not surprising because Trp-47 forms the bottom of the Arg-2 binding pocket and Tyr-8 comprises one side of the aromatic cage around Lys-4. Both Trp-47 and Tyr-8 likely contribute to the structural features of the binding pocket and make important hydrophobic interactions with the histone ligands. Additionally, mutation of Trp-47 to alanine resulted in significant structural changes in the ING3PHD causing a large increase in the α-helical content of the protein as observed by circular dichroism (Fig. 3E). These results indicate that Trp-47 also plays a role in the hydrophobic packing and folding of the ING3 PHD finger motif. Mutation of ING3PHD residues Pro-43 and Ser-15 had a less dramatic effect, resulting in a 2–3-fold drop in binding affinity for the H3K4me3 histone ligand (Table 4). Pro-43 was previously shown to be important for the coordination of Ala-1 of the histone H3 peptide for ING5PHD (37); however, there were no hydrogen bond interactions observed in any of our MD simulations (Table 3). Interestingly, the P43A mutation did change the α-helical content of the ING3PHD from 55 to 80%, indicating it does play an important structural role in protein folding/packing. Residue Ser-15 is highly conserved in all ING PHD finger proteins and makes up one side of the Lys-4 binding pocket. However, mutation of Ser-15 to alanine only produced a 2-fold drop in ligand binding affinity. The trajectory analysis of the MD data indicates that in the ING3PHD-H3K4me3 complex, Ser-15 does make hydrogen bond contacts to the Arg-8 side chain in the histone peptide over 85% of the simulation time (Table 3), but the modest 2-fold drop in binding affinity observed by ITC suggests that this interaction is less critical for histone ligand coordination than Tyr-8 and Trp-47 (Table 4). This is corroborated by only moderate chemical shift changes of the Ser-15 residue upon addition of the H3K4me3 peptide in NMR titration experiments (Fig. 2C). TABLE 4 Binding of the H3K4me3 peptide by ING3 PHD mutants by ITC Mutant ITC KD μm ING3 PHD S15A 1.50 ± 0.01 ING3 PHD D26H 1.93 ± 0.26 ING3 PHD P43A 2.45 ± 0.27 ING3 PHD Y8A 46 ± 2.99 ING3 PHD W47A 48.31 ± 0.23 ING3 PHD W31A No binding FIGURE 4. ITC measurements of interactions between mutant ING3 PHD finger proteins and the H3K4me3 histone tail ligand. A–F, exothermic ITC enthalpy plots for the binding of the ING3 PHD finger mutant proteins to the H3(1–12)K4me3 peptide. The inset lists the measured binding constants. Because the ING3 protein is a known tumor suppressor that is down-regulated in multiple cancers (8, 47, 48), we searched the cBioPortal for Cancer Genomics to see whether any mutations in the PHD finger of ING3 have been reported in cancer patients (33, 49). Our search uncovered three mutations in the ING3PHD, which include a frameshift insert at Cys-376 (Cys-22 in our MD simulation structure) found in cutaneous skin melanoma, a missense mutation H387P (His-33) associated with bladder carcinoma, and a missense mutation at D380H (Asp-26) found in head and neck carcinoma. Because the Cys-22 and His-33 residues are important for zinc ion coordination in the PHD finger motif, and their mutation would likely result in unfolding of the PHD domain, we focused on the effect of the D26H mutation on histone recognition by the ING3PHD. As with the S15A and P43A mutations, the D26H mutation resulted in a slightly weaker binding affinity for the H3K4me3 histone peptide as compared with the WT ING3PHD with a KD of 1.93 ± 0.26 μm (Table 4 and Fig. 4F). Interestingly, according to our circular dichroism analysis, this mutation increases the α-helical content slightly, from 56% in the WT ING3PHD to 60% in the ING3PHD-D26H protein (Fig. 3E and Table 5). As seen in Fig. 3D, residue Asp-26 (highlighted in cyan) is located peripheral to the histone binding pocket, so it is not surprising that this mutation has only a minor effect on histone recognition. TABLE 5 Analysis of ING3 PHD finger proteins by circular dichroism ING3 PHD protein α-Helix β-Strand % % WT 55.81 1.85 S15A 57.57 0.99 D26H 60.31 0.81 P43A 80.5 0.06 Y8A 50.58 2.57 W47A 90.16 0.04 W31A 32.08 5.44 Interaction of ING3 with the TIP60 Complex As described above, the ING3 protein is a functional subunit of the TIP60 HAT, and recognition of methylated histones by the ING3PHD is thought to recruit this multisubunit chromatin remodeling complex to histones to regulate gene expression (17). To determine whether ING3 mutants defective in H3K4me3 binding were retained within the TIP60 HAT complex (17, 19), FLAG-tagged full-length (residues 1–418) ING3WT, ING3Y362A, and ING3W385A forms were expressed in highly transfectable COS-7 monkey cells. Then, anti-FLAG immunoprecipitates were analyzed by immunoblotting with anti-TIP60 and anti-TRRAP antibodies. As observed with ING2 (14, 50) and ING4 (13), the defective H3K4me3-binding mutants, ING3Y362A and ING3W385A proteins do associate with the TIP60 and TRRAP subunits of the HAT complex (Fig. 5A). FIGURE 5. ING3 mutant proteins form a complex with TIP60, and the ING3PHD-H3K4me3 association is required for ING3-induced DNA damage-dependent cell death. A, full-length ING3 mutant proteins defective in H3K4me3 binding were retained within the TIP60 complex. COS-7 cells were transfected with pCMV-3×FLAG-ING3 wild-type (WT), Y362A (Y362A), or W385A (W385A) vectors for 72 h. Exogenous ING3 was immunoprecipitated with a FLAG antibody followed by immunoblotting. Empty pCMV-3×FLAG vector was used as a control in lane C. B and C, MCF7 cells were stably transduced with FLAG-ING3-expressing retroviral particles as indicated for 72 h, and for the final 20 h cells were treated with 400 ng/ml doxorubicin (doxo+). Subsequently, cell cycle status was determined by flow cytometry analysis of propidium iodide staining. Distribution of cells in cell cycle phases (B) and percentage of cells in sub-G1 phase (C) were determined. Data were plotted as the mean ± S.E. of three independent experiments; statistical analysis was undertaken with t test, and the * indicates p < 0.05. D, MCF7 cells were transduced with retroviral particles as in B for 72 h, and for the final 20 h the cells were treated with 400 ng/ml doxorubicin (doxo+). Subsequently, the percentage of apoptotic cells was determined by flow cytometry detection of annexin V-positive cells. Data were plotted as the mean ± S.E. of three independent experiments, and statistical analysis was undertaken with t test. The exogenous expression of ING proteins generally potentiates DNA damage responses and induces apoptosis (13). To determine whether the association with H3K4me3 is required for ING3 to induce apoptosis, human mammary carcinoma MCF7 cells were transduced with the ING3WT, ING3Y362A, or ING3W385A proteins (which correspond to residues Tyr-8 and Trp-31 highlighted in the time-averaged structure of the ING3PHD-H3K4me3 complex from the MD simulation, Fig. 3D), and DNA damage was induced by treating the cells with doxorubicin. Expression of ING3WT affected the cell cycle, and after treatment with doxorubicin we observed a decreased percentage of cells in both G1 and S phase (Fig. 5B). Interestingly, doxorubicin induced an increase in the sub-G1 population in the presence of ING3WT (p value 0.005), but not with the ING3Y362A or ING3W385A forms (Fig. 5, B and C), suggesting that the association of ING3 with H3K4me3 is required for the full-length ING3 protein to regulate apoptosis. Indeed, in parallel experiments, addition of doxorubicin induced an increase in annexin V staining in ING3WT-expressing cells (p value 0.006) but not in ING3Y362A- or ING3W385A-expressing cells (p values of 0.2 and 0.3, respectively) (Fig. 5D). These results are consistent with functional studies on other ING proteins, which suggest that H3K4me3 recognition tethers ING3 to its histone ligands and stimulates its biological activity (13, 14, 27, 37, 39). Discussion Our results demonstrate that the ING3 PHD finger preferentially binds to histone H3K4me3 over H3K4me2, H3K4me1, and the unmodified histone H3. Furthermore, this study confirms that selection of tri-methylated histones by the ING3PHD is a conserved function of the PHD finger domains in all ING family proteins, including ING1, ING2, ING4, and ING5, which underscores its importance in directing their biological function (13, 14, 27, 37, 39). The molecular mechanism of histone recognition by the ING3 PHD finger was evaluated using experimentally observed interactions between the ING3PHD and the histone H3K4me3 ligand with a combination of NMR chemical shift perturbation experiments and mutational analysis. The experimental results are supported by data obtained from the time-averaged structure of the ING3PHD-H3K4me3 complex using MD simulation. The Trp-31 and Tyr-8 residues help form an aromatic cage in the ING3PHD binding pocket, and they directly coordinate the tri-methylated lysine 4 in the histone H3 ligand. To recognize and specifically select for histone H3, ING3PHD also makes important contacts using residues Trp-47, Asp-23, and Cys-22 to create a secondary binding pocket and to form hydrogen bonds to Arg-2 in the histone peptide. As observed with other PHD fingers, the two grooves in the ING3PHD H3K4me3 binding pocket that coordinate Lys-4 and Arg-2 are separated by a narrow channel, which precludes binding of histone peptides with a large side chain at position 3. MD simulations revealed that the trimethylammonium group of Lys-4 is largely coordinated by hydrophobic and cation-π interactions with residues Tyr-8 and Trp-31 in the ING3PHD binding pocket. Additionally, mutation of Tyr-8 and Trp-31 prevented histone binding in our ITC assays. These results demonstrate that formation of the ING3-H3K4me3 complex is driven by a combination of hydrogen bonding, complementary surface interactions, and hydrophobic contacts. We used the time-averaged structure of the ING3PHD-H3K4me3 complex generated by MD simulation to compare coordination of the histone H3K4me3 peptide by ING3PHD with x-ray crystal structures of the other ING PHD fingers bound to H3K4me3. Superposition of the ING3PHD-H3K4me3 MD simulation onto the ING1PHD, ING2PHD, ING4PHD, and ING5PHD structures revealed that the overall structural fold and the histone binding mechanism among these PHD fingers is highly conserved (Fig. 3C). The time-averaged structure of the ING3PHD-H3K4me3 complex from the MD simulation is most closely related to the x-ray crystal structure of ING1PHD (PDB code 2QIC, r.m.s.d. of 0.86 Å over 53 Cα atoms), followed by ING2PHD (PDB code 2G6Q, r.m.s.d. of 0.87 Å), ΙNG5PHD (PDB code 3C6W, r.m.s.d. 1.094 Å), and ING4PHD (PDB code 2PNX, r.m.s.d. 1.098 Å). Interestingly, our ligand-bound time-averaged MD simulation structure of the ING3PHD-H3K4me3 complex superimposes more closely with the other ligand-bound ING PHD finger structures than it does with the ligand-free NMR structure of ING3PHD (PDB code 1X4I) (r.m.s.d., range from 1.3 o 2.0 Å over the Cα atoms) (40). As seen in Fig. 3C, the orientation of the H3K4me3 histone peptide in the binding pocket is conserved among all five of the INGPHD-H3K4me3 structures. In addition, the same specific molecular interactions observed in our MD simulation of ING3PHD-H3K4me3 complex are also found in the other INGPHD-H3K4me3 structures, revealing that all of the ING PHD fingers use a conserved binding mode for histone ligand recognition. For example, in the time-averaged structure of the ING3PHD-H3K4me3 complex from the MD simulation, the semi-aromatic cage is formed around the trimethylammonium group of Lys-4, in which Trp-31 and Tyr-8 make cation-π, hydrophobic, and van der Waals contacts with this group. Residues Met-19 and Ser-15 compose the remainder of the hydrophobic cage, and all four of these residues are conserved in the PHD finger binding pockets of ING1–5 (Fig. 3A) (40). Additionally, the coordination of Arg-2 in the H3K4me3 histone peptide is similar among the ING PHD finger proteins with hydrogen bond contacts between the side chains of Glu-30, Asp-23, and the backbone carbonyl of Cys-22 in ING3PHD, to the side chain amino groups of Arg-2 (Fig. 3B). The same interactions are observed between residues Glu-234, Asp-227, and Cys-226 of the ING1 PHD finger and the Arg-2 side chain in the H3K4me3 histone peptide (39), and this bonding pattern is also seen in the ING2, ING4, and ING5 PHD finger proteins (13, 27, 37). Thus, this structural comparison of the molecular recognition of H3K4me3 by the ING3 PHD finger with other INGPHD proteins reveals that the mechanism of histone ligand binding is universally conserved within the ING protein family. Finally, experiments with the full-length ING3 protein revealed that PHD finger mutations that inhibit H3K4me3 binding do not have an effect on the presence of ING3 in the TIP60 HAT complex, but they do prevent ING3 from up-regulating DNA damage-induced cell death. A recent study also showed that the chromodomain in TIP60 recognizes H3K9me3 and stimulates the acetylation of the ataxia telangiectasia mutated protein kinase in DNA double strand break repair (20). The recruitment of large enzymatic complexes involved in chromatin remodeling is often carried out by epigenetic “reader” domains that recognize post-translational modifications on the nucleosomes (37, 51–54). These data indicate that recognition of H3K4me3 by the ING3 PHD is necessary to target the TIP60 complex acetyltransferase activity to up-regulate apoptosis. Collectively, the structural and functional information presented here will be essential for further study of the biological activity of ING3 in the TIP60 HAT complex and the role of the ING3 PHD finger domain in epigenetic signaling by this complex. Experimental Procedures ING3 PHD Plasmid Construction The ING3 PHD finger region (residues 360–409) was amplified using PCR and cloned into the pDEST15 vector encoding an N-terminal glutathione transferase (GST) tag using the Gateway cloning technology (Invitrogen) as described previously (55). ING3PHD proteins with single mutations at W31A, W47A, S15A, P43A, Y8A, and D26H were generated using the QuikChange® mutagenesis procedure (Stratagene) as described previously (56). The DNA sequence for wild-type and mutant proteins was verified at the University of Vermont DNA facility, and the plasmids were transformed into in Escherichia coli RosettaTM 2(DE3)pLysS competent cells (Novagen) for protein expression. ING3 PHD Finger Expression and Purification E. coli cells containing the wild-type GST-tagged ING3PHD were grown in Terrific Broth (TB) or in 15NH4Cl-supplemented or 15NH4Cl/13C6 d-glucose-supplemented minimal media. The cultures were grown at 37 °C to an A600 of 1, induced with 0.25 mm isopropyl β-d-1-thiogalactopyranoside (IPTG), and incubated for an additional 16 h at 20 °C. Recombinant protein was purified by sonicating the harvested cell pellet resuspended in 200 ml of lysis buffer (50 mm Tris-HCl, pH 7.5, 150 mm NaCl, 0.05% Nonidet P-40, and 1 mm DTT), containing 0.1 mg/ml lysozyme, 50 units of DNase I (Thermo Scientific), and 1 tablet of Pierce Protease Inhibitor mixture (Thermo Scientific). After centrifugation at 10,000 rpm for 10 min, the cell supernatant was added to 12.5 ml of glutathione-agarose resin beads (Thermo Scientific) and incubated on ice (4 °C) for 2 h while agitating. The beads were added to a 25-ml Econo-Column® chromatography column (Bio-Rad) and washed with three column volumes of wash buffer (20 mm Tris-HCl, pH 7.5, 150 mm NaCl, and 1 mm DTT). The GST tag was cleaved overnight at 4 °C by addition of PreScission Protease (GE Healthcare), and the ING3 PHD finger was eluted in wash buffer and concentrated to about 3-ml total volume. Protein concentration was determined by absorbance measurement and the A280 extinction coefficient of ING3PHD (16960 m−1 cm−1). The purity of the ING3 PHD finger was verified by SDS-polyacrylamide gels stained with GelCode Blue Safe protein stain (Thermo Scientific). Tryptophan Fluorescence Spectroscopy Tryptophan fluorescence spectra of samples in tryptophan fluorescence buffer (100 mm NaPO4, pH 7.5, 150 mm NaCl, and 1 mm DTT) were collected at 25 °C using a Cary Eclipse fluorescence spectrometer (Varian). The samples contained 10 μm ING3 PHD finger protein and progressively increasing concentrations of histone H3 and H4 peptides. The 12- and 14-mer unlabeled histone tail peptides with an amidated C terminus and specific methylation modifications (H3K4me3, H3K4me2, H3K4me1, H3 unmodified, and H4 unmodified) were synthesized by the Peptide Core Facility at the University of Colorado at Denver. After excitation at 295 nm, emission spectra were recorded between 305 and 405 nm with 0.5-nm increments and at a 1-s integration time, averaging over three scans. Each titration experiment was repeated three times, and the average KD values were calculated using quadratic Equation 1, where Fi is the fluorescence change; Fs is the fluorescence change at saturation of XT (the total protein concentration); and YT is the peptide concentration. Isothermal Titration Calorimetry ITC measurements were recorded at 5 °C using a MicroCal iTC200 (GE Healthcare) as described previously (55). The wild-type and mutant ING3PHD proteins and histone peptide samples were prepared in a 20 mm NaH2PO4, pH 7.0, and 150 mm NaCl ITC buffer by dialysis for 24–48 h. Titration experiments were set up for optimal heat of binding reactions by using 100–200 μm ING3PHD in the sample cell, and between 1 and 5 mm histone peptide in the injection syringe. Control experiments were performed under identical conditions to determine the heat of dilution of the titrant peptides into the experimental buffer. This was subtracted from the experimental data as part of data analysis. Data were analyzed using the software ORIGIN 7.0 (OriginLab Corp.). All experiments where binding occurred were performed in triplicate, whereas non-binding experiments were performed in duplicate. HSQC-NMR Chemical shift perturbation experiments were conducted using 0.5 mm uniformly 15N-labeled ING3 PHD finger in buffer containing 20 mm Tris-HCl, pH 6.8, 150 mm NaCl, 10 mm DTT, and 10% D2O. Titration mixtures of the ING3PHD protein and each of the modified histone peptides were prepared at concentration ratios of 1:0, 1:0.13, 1:0.26, 1:0.52, 1:1.3, and 1:2.61 in a volume of 35 μl. These mixtures were then transferred into 1.7-mm NMR tubes (Bruker). Two-dimensional 15N HSQC (heteronuclear single quantum coherence) experiments for all samples were run at 25 °C on a 600 MHz Bruker AVANCE III spectrometer equipped with a z-gradient 1.7-mm TCI probe at the National Magnetic Resonance Facility at Madison (NMRFAM) using the NMRBot software (57). The NMR data were collected with 1024 × 128 complex data points along the 1H and 15N dimensions, with acquisition times of 104 and 81 ms, respectively, using eight scans per free induction decay. Normalized chemical shift changes were calculated using Equation 2, where ΔδH and ΔδN are the proton and nitrogen change in chemical shift in ppm, respectively. KD values were calculated by a nonlinear least squares analysis in KaleidaGraph using Equation 3, where [L] is the concentration of the peptide; [P] is the concentration of the protein; Δδ is the observed chemical shift change, and Δδmax is the normalized chemical shift change at saturation. To obtain the backbone resonance assignments, 1 mm of the 15N,13C double-labeled ING3 PHD finger was prepared in buffer containing 20 mm Tris-HCl, pH 6.8, 150 mm NaCl, 10 mm DTT, and 10% D2O. The Agilent version of ADAPT-NMR (Assignment-directed Data collection Algorithm utilizing a Probabilistic Toolkit in NMR) was used to optimize simultaneous fast data collection and fully automated NMR backbone assignments. In addition to a two-dimensional 1H-15N HSQC spectrum, ADAPT-NMR recorded six three-dimensional spectra as two-dimensional orthogonal and tilted planes at optimal projection angles as follows: HNCO, HN(CA)CO, HN(CO)CA, HNCA, CBCA(CO)NH, and HN(CA)CB (58). These experiments were collected at 25 °C on a 600 MHz Varian VNMRS spectrometer equipped with a z-gradient 5mm cryogenic probe. All two-dimensional planes were processed automatically by ADAPT-NMR with NMRpipe software (59). After less than 1 day, ADAPT-NMR was able to assign all backbone amides (100%). These fully automated assignments were visualized, validated, and further refined by using the ADAPT-NMR Enhancer package (60). This inspection led to the correction of backbone assignments for two residues. MD Simulations The MD simulations were used to investigate the molecular interactions of the ING3PHD with histone peptide ligands. The published NMR solution structure of the ING3 protein (PDB code 1X4I) was taken from the Protein Data Bank (61). The 1X4I structure was modified with Biopolymer, a structure preparation tool in the SybylX2.1 suite; the N- and the C-terminal residues were capped with N-acetyl and N-methylamide groups; and protonation types were set for His (ϵ-protonated), Glu (negatively charged), and Lys (positively charged) residues. The Cys and His residues making tetrahedral coordination geometry with structural zinc atoms were treated as special residues in their anionic forms as follows: CY1 residue in zinc-CCCC, CY2, and HIN in zinc-CCCH coordination structures, respectively. The force field parameters for these residues were adopted from ZAFF (62). The side chains of residues Cys-21, Cys-26, Cys-48, and Cys-51 coordinate Zn1 in zinc-CCCC fashion, and Cys-8, Cys-10, Cys-35, and the His-32 coordinate Zn2 in zinc-CCCH fashion; both exhibit tetrahedral geometry. The starting structures for MD simulations were obtained by flexible protein-peptide docking using the Rosetta FlexPepDock (41, 42) docking program. The docking protocol incorporates iterative cycles of optimization and energy minimization that include full flexibility and rigid body orientation for the peptide backbone, as well as side chain flexibility for both the peptide and the receptor protein. The resulting FlexPepDock complexes were ranked based on the Rosetta full-atom energy function (Rosetta score 12) available within the Rosetta modeling framework (43). The complexes with the highest scores were used as input for the MD simulations. The FlexPepDock program requires the approximate initial conformation of peptide close to its putative binding pocket. For the H3 peptides, the x-ray crystal structures of ING4 (PDB code 2PNX) and ING5 (PDB code 3C6W), in which the H3K4Me3 peptide is bound to the respective proteins, were used as starting templates. The ING4PHD-H3K4me3 and ING5PHD-H3K4me3 structures were aligned with the apo-ING3 structures (PDB code 1X4I) based on homology, using the Biopolymer module in Sybyl-X2.1. No structures are available for ING PHD fingers bound to the unmodified histone H4 peptide. Hence, the starting conformation was obtained by docking the unmodified H4 peptide to ING3PHD using the Hex docking program through its webserver (accessed on June 1, 2014) (44). The Hex docking algorithm is based on spherical polar correlations of protein surface shape and electrostatic representations. The initial constraints given to the program were the putative binding residues of the protein observed in our NMR titration studies. The three lowest energy poses of the unmodified histone H4 peptide obtained from the Hex program were fed into the FlexPepDock server to ascertain the binding conformations, and the best pose with the highest Rosetta 12 score was selected for MD simulation, as described above for the H3 peptides. The MD simulations were performed using the Amber 14 package (63) under isothermal/isobaric (NPT) conditions with Amber ff14SB force field (64) for protein and peptide molecules. The force field parameters for methylated lysine (mono-, di-, and tri-methylated) were obtained by following the standard procedure used in the AMBER force field development utilizing the Mulliken charges calculated as shown previously (55). The backbone torsion parameters are the same as those of the natural lysine residue in the AMBER ff14SB force field. To prepare the ING3PHD structure for MD simulations, the Leap program from Antechamber tools, AmberTools 14 suite (63, 65), was used to generate the parameter/topology (prmtop) and input coordinate (inpcrd) files. The net charge of the protein-peptide complexes varied from +3 to +4 depending upon the overall charge of the peptide, and was neutralized by adding Cl− ions at positions of high positive electron potential around the complexes. The system was immersed in a truncated octahedral box of pre-equilibrated TIP3P water molecules (66) in the way that no atoms in the protein-ligand complexes were closer than 16 Å to any of the sides of the water box. The counter ions and solvent molecules were briefly minimized (2500 steps) to remove any bad contacts with the complexes, whereby the protein and peptides were position-restrained using force constant of 100 kcal/(mol·Å2), followed by another 2500-step minimization of the whole solvated complex. To allow the readjustment of solvent molecules to the potential field of the protein-peptide complex, the solvent equilibration step was performed in three stages. The details of this equilibration step have been described previously (55). The production phase with the entire system was carried out under isothermal/isobaric conditions for 10 ns. The SHAKE algorithm (67) was used to constrain bonds involving hydrogen, allowing time steps of 2 fs, for a total of 5,000,000 steps. The trajectory file was written for every 1000 steps (2 ps) resulting in 5000 frames. The cutoff for non-bonded interactions was set to 12 Å in all steps. H-bond interaction analysis was carried out on the 5000 snapshots from the entire 10-ns production phase using the cpptraj program in the AmberTools 14 suite (63, 65), with the cutoff values for distance (3.2 Å) and angle (135°). The pairwise interactions were monitored between the ING3PHD-binding site residues (Tyr-8, Gln-13, Val-14, Ser-15, Tyr-16, Gly-17, Glu-18, Met-19, Val-20, Gly-21, Cys-22, Asp-23, Gln-25, Glu-30, and Trp-31) and all histone peptide residues. To calculate the free energy of binding for each peptide, we used the Molecular Mechanics/Poisson Boltzmann Surface Area (MM/PBSA) method (45), which combines the molecular mechanical energies with the continuum solvent approaches. The molecular mechanical energies represent the internal (bond, angle, and dihedral) energy and van der Waals and electrostatic interactions. The nonpolar contribution to the solvation free energy is determined with solvent-accessible surface area-dependent terms. The method separates nonpolar contribution into two terms as follows: the attractive (dispersion) and repulsive (cavity) interactions. The estimates of vibrational entropies were made using the frequencies from the normal mode analysis with the nmode or NAB module from Amber. The calculations were done using the recently published MMPBSA.py program (46) from Amber Suite. Full-length ING Plasmids The cDNA of human ING3 was cloned by PCR (Platinum PCR SuperMix High Fidelity (Invitrogen; catalog no. 12532-016) with fwd 5′-ggccAGATCTttgtacctagaagactatctgga-3′ and rev 5′-aggacTCGAGttatttgtgtctgctgcctct-3′ primers) on reverse-transcribed (SuperScript VILO Master Mix; Invitrogen; catalog no. 11755-050) total RNA isolated (TRIzol; Invitrogen; catalog no. 15596-026) from breast carcinoma MCF7 cells. The PCR product was gel-purified, digested with BglII (New England Biolabs) and XhoI (New England Biolabs), and inserted in pCMV 3×FLAG (Stratagene pCMV-3Tag-1A) or pMF retroviral vector (68). The Y362A and W385A H3K4me3-binding defective mutants were generated by site-directed mutagenesis using degenerate primers (Y362A fwd 5′-ccaaatgaacctcgaGCctgcatttgtaatcag-3′ and Y362A rev 5′-ctgattacaaatgcagGCtcgaggttcatttgg-3′; W385A fwd 5′-gattgccctatagaaGCgttccattatggctgc-3′ and W385A rev 5′-gcagccataatggaacGCttctatagggcaatc-3′) and Pfu Turbo (Stratagene). PCR products were digested with DpnI (New England Biolabs) and transformed. All constructs were sequence-verified (Beckman Coulter Genomics). Immunoprecipitation COS-7 cells were seeded at 500,000 cells per 100-mm dish and transfected with pCMV 3×FLAG-ING3 wild-type (ING3WT), Y362A (ING3Y362A), or W385A (ING3W385A) vectors using Mirus LT1 reagent. Cells were harvested 72 h post-transfection, resuspended in lysis buffer (50 mm Tris-Cl, pH 7.5, 150 mm NaCl, 0.2 mm Na3VO4, 1% Nonidet P-40, 1 mm PMSF, 1 mm DTT, and protease inhibitors (Roche Applied Science)), incubated with anti-FLAG antibody (M2 clone, Sigma) for 16 h at 4 °C, and immunoprecipitated using Protein G-Sepharose beads (GE Healthcare). Immunoprecipitates were analyzed by immunoblotting with anti-TRRAP (Abcam), anti-TIP60 (SCBT), or anti-FLAG (Sigma) antibodies. Retroviral Expression HEK293T cells (3 million cells per 100-mm dish) were transfected with 9 μg of indicated pMF constructs (empty, ING3WT, ING3Y362A, or ING3W385A) and 4.5 μg of each VSV-G and gag/pol expressing plasmids using 54 μl of TransIT-LT1 reagent (Mirus). The medium was changed 24 h post-transfection, and supernatants collected at 48 and 72 h post-transfection were filtered (Millipore Millex-HV 0.45-μm PVDF filters) and concentrated (Millipore Amicon Ultra-15). Flow Cytometry Analysis MCF7 cells were transduced with the indicated retroviral particles in the presence of Polybrene (8 μg/ml). The next day, the medium was refreshed. After 48 h, puromycin selection (1 μg/ml) was applied for 24 h. Cell cycle profiles were assessed as described previously (68). Briefly, MCF7 cells were re-suspended in a propidium iodide staining mixture (0.8% Triton X-100, 50 μg/ml propidium iodide, and 75 μg/ml RNase A) and incubated for 10 min at room temperature. Apoptosis was assessed by staining with annexin V Alexa Fluor488 and propidium iodide (Life Technologies, Inc., V13241). Stained cells were immediately injected into a FACSCalibur; 10,000 cells were analyzed per sample. Author Contributions S.K., S. N., G. C., S. C., M. T., and U. L. M. performed the experiments and together with O. B., C. N. R., J. L. M., S. B., and K. C. G. analyzed the data. K. C. G. and O. B. wrote the manuscript with input from all authors. Supplementary Material Supplemental Data * This work was supported in part by an Albany College of Pharmacy and Health Sciences Scholarship of Discovery grant (to K. C. G). The authors declare that they have no conflicts of interest with the contents of this article. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. This article contains supplemental File 1. 7 The abbreviations used are: PHDplant homeodomain ITCisothermal titration PDBProtein Data Bank r.m.s.d.root mean square deviation MDmolecular dynamics HAThistone acetyltransferase INGINhibitor of Growth HSQCheteronuclear single quantum coherence fwdforward revreverse. Acknowledgments Isolated cDNA from human ING3 (Q9NXR8) was kindly provided by Dr. Jacques Côté at Université Laval. This study made use of the National Magnetic Resonance Facility at Madison, WI, which is supported by National Institutes of Health Grants P41GM103399 (NIGMS), original Grant Number P41RR002301. NMRFAM equipment was purchased with funds from the University of Wisconsin-Madison; National Institutes of Health Grants P41GM103399, S10RR02781, S10RR08438, S10RR023438, S10RR025062, and S10RR029220; National Science Foundation Grants DMB-8415048, OIA-9977486, and BIR-9214394; and the United States Department of Agriculture. DNA sequencing was performed in the University of Vermont Cancer Center DNA Analysis Facility. ==== Refs References 1. Luger K. , Mäder A. W. , Richmond R. K. , Sargent D. F. , and Richmond T. J. (1997 ) Crystal structure of the nucleosome core particle at 2.8 A resolution . Nature 389 , 251 –260 9305837 2. Lachner M. , O'Carroll D. , Rea S. , Mechtler K. , and Jenuwein T. (2001 ) Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins . Nature 410 , 116 –120 11242053 3. Musselman C. A. , Khorasanizadeh S. , and Kutateladze T. G. (2014 ) Towards understanding methyllysine readout . Biochim. Biophys. Acta 1839 , 686 –693 24727128 4. Musselman C. A. , Lalonde M. E. , Côté J. , and Kutateladze T. G. (2012 ) Perceiving the epigenetic landscape through histone readers . Nat. Struct. Mol. Biol . 19 , 1218 –1227 23211769 5. Garkavtsev I. , Kazarov A. , Gudkov A. , and Riabowol K. (1996 ) Suppression of the novel growth inhibitor p33ING1 promotes neoplastic transformation . Nat. Genet . 14 , 415 –420 8944021 6. Nagashima M. , Shiseki M. , Miura K. , Hagiwara K. , Linke S. P. , Pedeux R. , Wang X. W. , Yokota J. , Riabowol K. , and Harris C. C. (2001 ) DNA damage-inducible gene p33ING2 negatively regulates cell proliferation through acetylation of p53 . Proc. Natl. Acad. Sci. U.S.A . 98 , 9671 –9676 11481424 7. Nagashima M. , Shiseki M. , Pedeux R. M. , Okamura S. , Kitahama-Shiseki M. , Miura K. , Yokota J. , and Harris C. C. (2003 ) A novel PHD-finger motif protein, p47ING3, modulates p53-mediated transcription, cell cycle control, and apoptosis . Oncogene 22 , 343 –350 12545155 8. Gunduz M. , Ouchida M. , Fukushima K. , Ito S. , Jitsumori Y. , Nakashima T. , Nagai N. , Nishizaki K. , and Shimizu K. (2002 ) Allelic loss and reduced expression of the ING3, a candidate tumor suppressor gene at 7q31, in human head and neck cancers . Oncogene 21 , 4462 –4470 12080476 9. Shiseki M. , Nagashima M. , Pedeux R. M. , Kitahama-Shiseki M. , Miura K. , Okamura S. , Onogi H. , Higashimoto Y. , Appella E. , Yokota J. , and Harris C. C. (2003 ) p29ING4 and p28ING5 bind to p53 and p300, and enhance p53 activity . Cancer Res . 63 , 2373 –2378 12750254 10. Garkavtsev I. , Kozin S. V. , Chernova O. , Xu L. , Winkler F. , Brown E. , Barnett G. H. , and Jain R. K. (2004 ) The candidate tumour suppressor protein ING4 regulates brain tumour growth and angiogenesis . Nature 428 , 328 –332 15029197 11. Kim S. , Chin K. , Gray J. W. , and Bishop J. M. (2004 ) A screen for genes that suppress loss of contact inhibition: identification of ING4 as a candidate tumor suppressor gene in human cancer . Proc. Natl. Acad. Sci. U.S.A . 101 , 16251 –16256 15528276 12. Bua D. J. , and Binda O. (2009 ) The return of the INGs, histone mark sensors and phospholipid signaling effectors . Curr. Drug Targets 10 , 418 –431 19442114 13. Hung T. , Binda O. , Champagne K. S. , Kuo A. J. , Johnson K. , Chang H. Y. , Simon M. D. , Kutateladze T. G. , and Gozani O. (2009 ) ING4 mediates crosstalk between histone H3 K4 trimethylation and H3 acetylation to attenuate cellular transformation . Mol. Cell 33 , 248 –256 19187765 14. Shi X. , Hong T. , Walter K. L. , Ewalt M. , Michishita E. , Hung T. , Carney D. , Peña P. , Lan F. , Kaadige M. R. , Lacoste N. , Cayrou C. , Davrazou F. , Saha A. , Cairns B. R. , Ayer D. E. , Kutateladze T. G. , Shi Y. , Côté J. , Chua K. F. , and Gozani O. (2006 ) ING2 PHD domain links histone H3 lysine 4 methylation to active gene repression . Nature 442 , 96 –99 16728974 15. Guérillon C. , Bigot N. , and Pedeux R. (2014 ) The ING tumor suppressor genes: status in human tumors . Cancer Lett . 345 , 1 –16 24333729 16. Tallen G. , and Riabowol K. (2014 ) Keep-ING balance: tumor suppression by epigenetic regulation . FEBS Lett . 588 , 2728 –2742 24632289 17. Doyon Y. , Cayrou C. , Ullah M. , Landry A. J. , Côté V. , Selleck W. , Lane W. S. , Tan S. , Yang X. J. , and Côté J. (2006 ) ING tumor suppressor proteins are critical regulators of chromatin acetylation required for genome expression and perpetuation . Mol. Cell 21 , 51 –64 16387653 18. Kuzmichev A. , Zhang Y. , Erdjument-Bromage H. , Tempst P. , and Reinberg D. (2002 ) Role of the Sin3-histone deacetylase complex in growth regulation by the candidate tumor suppressor p33(ING1) . Mol. Cell. Biol . 22 , 835 –848 11784859 19. Doyon Y. , Selleck W. , Lane W. S. , Tan S. , and Côté J. (2004 ) Structural and functional conservation of the NuA4 histone acetyltransferase complex from yeast to humans . Mol. Cell. Biol . 24 , 1884 –1896 14966270 20. Sun Y. , Jiang X. , and Price B. D. (2010 ) Tip60: connecting chromatin to DNA damage signaling . Cell Cycle 9 , 930 –936 20160506 21. Grigoletto A. , Lestienne P. , and Rosenbaum J. (2011 ) The multifaceted proteins Reptin and Pontin as major players in cancer . Biochim. Biophys. Acta 1815 , 147 –157 21111787 22. Lee J. S. , Kim Y. , Bhin J. , Shin H. J. , Nam H. J. , Lee S. H. , Yoon J. B. , Binda O. , Gozani O. , Hwang D. , and Baek S. H. (2011 ) Hypoxia-induced methylation of a pontin chromatin remodeling factor . Proc. Natl. Acad. Sci. U.S.A . 108 , 13510 –13515 21825155 23. Lee J. S. , Kim Y. , Kim I. S. , Kim B. , Choi H. J. , Lee J. M. , Shin H. J. , Kim J. H. , Kim J. Y. , Seo S. B. , Lee H. , Binda O. , Gozani O. , Semenza G. L. , Kim M. , et al (2010 ) Negative regulation of hypoxic responses via induced Reptin methylation . Mol. Cell 39 , 71 –85 20603076 24. Colla S. , Tagliaferri S. , Morandi F. , Lunghi P. , Donofrio G. , Martorana D. , Mancini C. , Lazzaretti M. , Mazzera L. , Ravanetti L. , Bonomini S. , Ferrari L. , Miranda C. , Ladetto M. , Neri T. M. , et al (2007 ) The new tumor-suppressor gene inhibitor of growth family member 4 (ING4) regulates the production of proangiogenic molecules by myeloma cells and suppresses hypoxia-inducible factor-1α (HIF-1α) activity: involvement in myeloma-induced angiogenesis . Blood 110 , 4464 –4475 17848618 25. Ozer A. , Wu L. C. , and Bruick R. K. (2005 ) The candidate tumor suppressor ING4 represses activation of the hypoxia inducible factor (HIF) . Proc. Natl. Acad. Sci. U.S.A . 102 , 7481 –7486 15897452 26. Obri A. , Ouararhni K. , Papin C. , Diebold M. L. , Padmanabhan K. , Marek M. , Stoll I. , Roy L. , Reilly P. T. , Mak T. W. , Dimitrov S. , Romier C. , and Hamiche A. (2014 ) ANP32E is a histone chaperone that removes H2A.Z from chromatin . Nature 505 , 648 –653 24463511 27. Peña P. V. , Davrazou F. , Shi X. , Walter K. L. , Verkhusha V. V. , Gozani O. , Zhao R. , and Kutateladze T. G. (2006 ) Molecular mechanism of histone H3K4me3 recognition by plant homeodomain of ING2 . Nature 442 , 100 –103 16728977 28. Huang H. , Reed C. P. , Mordi A. , Lomberk G. , Wang L. , Shridhar V. , Hartmann L. , Jenkins R. , and Smith D. I. (1999 ) Frequent deletions within FRA7G at 7q31.2 in invasive epithelial ovarian cancer . Genes Chromosomes Cancer 24 , 48 –55 9892108 29. Edelson M. I. , Scherer S. W. , Tsui L. C. , Welch W. R. , Bell D. A. , Berkowitz R. S. , and Mok S. C. (1997 ) Identification of a 1300 kilobase deletion unit on chromosome 7q31.3 in invasive epithelial ovarian carcinomas . Oncogene 14 , 2979 –2984 9205105 30. Latil A. , Cussenot O. , Fournier G. , Baron J. C. , and Lidereau R. (1995 ) Loss of heterozygosity at 7q31 is a frequent and early event in prostate cancer . Clin. Cancer Res . 1 , 1385 –1389 9815935 31. Neklason D. W. , Kerber R. A. , Nilson D. B. , Anton-Culver H. , Schwartz A. G. , Griffin C. A. , Lowery J. T. , Schildkraut J. M. , Evans J. P. , Tomlinson G. E. , Strong L. C. , Miller A. R. , Stopfer J. E. , Finkelstein D. M. , Nadkarni P. M. , et al (2008 ) Common familial colorectal cancer linked to chromosome 7q31: a genome-wide analysis . Cancer Res . 68 , 8993 –8997 18974144 32. DeLano W. L. (2002 ) The PyMOL Molecular Graphics System . Version 1.5.0.5 DeLano Scientific , Palo Alto, CA 33. Gao J. , Aksoy B. A. , Dogrusoz U. , Dresdner G. , Gross B. , Sumer S. O. , Sun Y. , Jacobsen A. , Sinha R. , Larsson E. , Cerami E. , Sander C. , and Schultz N. (2013 ) Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal . Sci. Signal . 6 , pl1 23550210 34. Bismar T. A. , Alshalalfa M. , Petersen L. F. , Teng L. H. , Gerke T. , Bakkar A. , Al-Mami A. , Liu S. , Dolph M. , Mucci L. A. , and Alhajj R. (2014 ) Interrogation of ERG gene rearrangements in prostate cancer identifies a prognostic 10-gene signature with relevant implication to patients' clinical outcome . BJU Int . 113 , 309 –319 24006850 35. Nabbi A. , Almami A. , Thakur S. , Suzuki K. , Boland D. , Bismar T. A. , and Riabowol K. (2015 ) ING3 protein expression profiling in normal human tissues suggest its role in cellular growth and self-renewal . Eur. J. Cell Biol . 94 , 214 –222 25819753 36. Almami A. , Hegazy S. A. , Nabbi A. , Alshalalfa M. , Salman A. , Abou-Ouf H. , Riabowol K. , and Bismar T. A. (2016 ) ING3 is associated with increased cell invasion and lethal outcome in ERG-negative prostate cancer patients . Tumour Biol ., in press 37. Champagne K. S. , Saksouk N. , Peña P. V. , Johnson K. , Ullah M. , Yang X. J. , Côté J. , and Kutateladze T. G. (2008 ) The crystal structure of the ING5 PHD finger in complex with an H3K4me3 histone peptide . Proteins 72 , 1371 –1376 18623064 38. Palacios A. , Garcia P. , Padró D. , López-Hernández E. , Martín I. , and Blanco F. J. (2006 ) Solution structure and NMR characterization of the binding to methylated histone tails of the plant homeodomain finger of the tumour suppressor ING4 . FEBS Lett . 580 , 6903 –6908 17157298 39. Peña P. V. , Hom R. A. , Hung T. , Lin H. , Kuo A. J. , Wong R. P. , Subach O. M. , Champagne K. S. , Zhao R. , Verkhusha V. V. , Li G. , Gozani O. , and Kutateladze T. G. (2008 ) Histone H3K4me3 binding is required for the DNA repair and apoptotic activities of ING1 tumor suppressor . J. Mol. Biol . 380 , 303 –312 18533182 40. Champagne K. S. , and Kutateladze T. G. (2009 ) Structural insight into histone recognition by the ING PHD fingers . Curr. Drug Targets 10 , 432 –441 19442115 41. London N. , Raveh B. , Cohen E. , Fathi G. , and Schueler-Furman O. (2011 ) Rosetta FlexPepDock web server–high resolution modeling of peptide-protein interactions . Nucleic Acids Res . 39 , W249 –W253 21622962 42. Raveh B. , London N. , and Schueler-Furman O. (2010 ) Sub-angstrom modeling of complexes between flexible peptides and globular proteins . Proteins 78 , 2029 –2040 20455260 43. Das R. , and Baker D. (2008 ) Macromolecular modeling with Rosetta . Annu. Rev. Biochem . 77 , 363 –382 18410248 44. Ritchie D. W. , and Kemp G. J. (2000 ) Protein docking using spherical polar Fourier correlations . Proteins 39 , 178 –194 10737939 45. Srinivasan J. , Cheatham T. E. 3rd., Cieplak P. , Kollman P. A. , and Case D. A. (1998 ) Continuum solvent studies of the stability of DNA, RNA, and phosphoramidate-DNA helices . J. Am. Chem. Soc . 120 , 9401 –9409 46. Miller B. R. 3rd., McGee T. D. Jr., Swails J. M. , Homeyer N. , Gohlke H. , and Roitberg A. E. (2012 ) MMPBSA.py: an efficient program for end-state free energy calculations . J. Chem. Theory Comput . 8 , 3314 –3321 26605738 47. Gou W. F. , Sun H. Z. , Zhao S. , Niu Z. F. , Mao X. Y. , Takano Y. , and Zheng H. C. (2014 ) Downregulated inhibitor of growth 3 (ING3) expression during colorectal carcinogenesis . Indian J. Med. Res . 139 , 561 –567 24927342 48. Gunduz M. , Beder L. B. , Gunduz E. , Nagatsuka H. , Fukushima K. , Pehlivan D. , Cetin E. , Yamanaka N. , Nishizaki K. , Shimizu K. , and Nagai N. (2008 ) Downregulation of ING3 mRNA expression predicts poor prognosis in head and neck cancer . Cancer Sci . 99 , 531 –538 18081876 49. Cerami E. , Gao J. , Dogrusoz U. , Gross B. E. , Sumer S. O. , Aksoy B. A. , Jacobsen A. , Byrne C. J. , Heuer M. L. , Larsson E. , Antipin Y. , Reva B. , Goldberg A. P. , Sander C. , and Schultz N. (2012 ) The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data . Cancer Discov . 2 , 401 –404 22588877 50. Bua D. J. , Martin G. M. , Binda O. , and Gozani O. (2013 ) Nuclear phosphatidylinositol-5-phosphate regulates ING2 stability at discrete chromatin targets in response to DNA damage . Sci. Rep . 3 , 2137 23823870 51. Saksouk N. , Avvakumov N. , Champagne K. S. , Hung T. , Doyon Y. , Cayrou C. , Paquet E. , Ullah M. , Landry A. J. , Côté V. , Yang X. J. , Gozani O. , Kutateladze T. G. , and Côté J. (2009 ) HBO1 HAT complexes target chromatin throughout gene coding regions via multiple PHD finger interactions with histone H3 tail . Mol. Cell 33 , 257 –265 19187766 52. Carlson S. , and Glass K. C. (2014 ) The MOZ histone acetyltransferase in epigenetic signaling and disease . J. Cell. Physiol . 229 , 1571 –1574 24633655 53. Wysocka J. , Swigut T. , Xiao H. , Milne T. A. , Kwon S. Y. , Landry J. , Kauer M. , Tackett A. J. , Chait B. T. , Badenhorst P. , Wu C. , and Allis C. D. (2006 ) A PHD finger of NURF couples histone H3 lysine 4 trimethylation with chromatin remodelling . Nature 442 , 86 –90 16728976 54. Rajakumara E. , Wang Z. , Ma H. , Hu L. , Chen H. , Lin Y. , Guo R. , Wu F. , Li H. , Lan F. , Shi Y. G. , Xu Y. , Patel D. J. , and Shi Y. (2011 ) PHD finger recognition of unmodified histone H3R2 links UHRF1 to regulation of euchromatic gene expression . Mol. Cell 43 , 275 –284 21777816 55. Poplawski A. , Hu K. , Lee W. , Natesan S. , Peng D. , Carlson S. , Shi X. , Balaz S. , Markley J. L. , and Glass K. C. (2014 ) Molecular insights into the recognition of N-terminal histone modifications by the BRPF1 bromodomain . J. Mol. Biol . 426 , 1661 –1676 24333487 56. Champagne K. S. , Piscitelli E. , and Francklyn C. S. (2006 ) Substrate recognition by the hetero-octameric ATP phosphoribosyltransferase from Lactococcus lactis . Biochemistry 45 , 14933 –14943 17154531 57. Clos L. J. 2nd, Jofre M. F. , Ellinger J. J. , Westler W. M. , and Markley J. L. (2013 ) NMRbot: Python scripts enable high-throughput data collection on current Bruker BioSpin NMR spectrometers . Metabolomics 9 , 558 –563 23678341 58. Bahrami A. , Tonelli M. , Sahu S. C. , Singarapu K. K. , Eghbalnia H. R. , and Markley J. L. (2012 ) Robust, integrated computational control of NMR experiments to achieve optimal assignment by ADAPT-NMR . PLoS ONE 7 , e33173 22427982 59. Delaglio F. , Grzesiek S. , Vuister G. W. , Zhu G. , Pfeifer J. , and Bax A. (1995 ) NMRPipe: a multidimensional spectral processing system based on UNIX pipes . J. Biomol NMR 6 , 277 –293 8520220 60. Lee W. , Bahrami A. , and Markley J. L. (2013 ) ADAPT-NMR Enhancer: complete package for reduced dimensionality in protein NMR spectroscopy . Bioinformatics 29 , 515 –517 23220573 61. Berman H. M. , Westbrook J. , Feng Z. , Gilliland G. , Bhat T. N. , Weissig H. , Shindyalov I. N. , and Bourne P. E. (2000 ) The Protein Data Bank . Nucleic Acids Res . 28 , 235 –242 10592235 62. Peters M. B. , Yang Y. , Wang B. , Füsti-Molnár L. , Weaver M. N. , and Merz K. M. Jr. (2010 ) Structural survey of zinc containing proteins and the development of the zinc AMBER force field (ZAFF) . J. Chem. Theory Comput . 6 , 2935 –2947 20856692 63. Case D. A. , Berryman J. T. , Betz R. M. , Cai Q. , Cerutti D. S. , Cheatham T. E. 3rd.,Darden T. A. A. , Duke R. E. , Gohlke H. , Goetz A. W. , Gusarov S. Homeyer N. , Janowski P. , Kaus J. , Kolossváry I. , Kovalenko A. , et al (2014 ) Amber14 and AmberTools14. in AMBER 14 , University of California , San Francisco 64. Hornak V. , Abel R. , Okur A. , Strockbine B. , Roitberg A. , and Simmerling C. (2006 ) Comparison of multiple Amber force fields and development of improved protein backbone parameters . Proteins 65 , 712 –725 16981200 65. Case D. A. , Cheatham T. E. 3rd, Darden T. , Gohlke H. , Luo R. , Merz K. M. Jr., Onufriev A. , Simmerling C. , Wang B. , and Woods R. J. (2005 ) The Amber biomolecular simulation programs . J. Comput. Chem . 26 , 1668 –1688 16200636 66. Jorgensen W. L. , Chandrasekhar J. , Madura J. D. , Impey R. W. , and Klein M. L. (1983 ) Comparison of simple potential functions for simulating liquid water . J. Chem. Phys . 79 , 926 –935 67. Ryckaert J. P. , Ciccotti G. , and Berendsen H. J. C. (1977 ) Numerical integration of the Cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes . J. Comp. Phys . 23 , 327 –341 68. O'Neill D. J. , Williamson S. C. , Alkharaif D. , Monteiro I. C. , Goudreault M. , Gaughan L. , Robson C. N. , Gingras A. C. , and Binda O. (2014 ) SETD6 controls the expression of estrogen-responsive genes and proliferation of breast carcinoma cells . Epigenetics 9 , 942 –950 24751716

PMC005xxxxxx/PMC5000131.txt

==== Front Front Cardiovasc MedFront Cardiovasc MedFront. Cardiovasc. Med.Frontiers in Cardiovascular Medicine2297-055XFrontiers Media S.A. 10.3389/fcvm.2016.00028Cardiovascular MedicineReviewThe Role of Genetic Testing in the Identification of Young Athletes with Inherited Primitive Cardiac Disorders at Risk of Exercise Sudden Death Tiziano Francesco Danilo 1Palmieri Vincenzo 2Genuardi Maurizio 1*Zeppilli Paolo 21Istituto di Medicina Genomica, Università Cattolica del Sacro Cuore, Roma, Italy2Unità di Medicina dello Sport, Fondazione Policlinico “A. Gemelli”, Università Cattolica del Sacro Cuore, Roma, ItalyEdited by: Matteo Vatta, Indiana University Bloomington, USA Reviewed by: Francesca Girolami, Azienda Ospedaliero-Universitaria Careggi, Italy; Massimo Zecchin, Azienda Sanitaria Universitaria Integrata di Trieste, Italy *Correspondence: Maurizio Genuardi, maurizio.genuardi@unicatt.itSpecialty section: This article was submitted to Cardiovascular Genetics and Systems Medicine, a section of the journal Frontiers in Cardiovascular Medicine 26 8 2016 2016 3 2823 5 2016 16 8 2016 Copyright © 2016 Tiziano, Palmieri, Genuardi and Zeppilli.2016Tiziano, Palmieri, Genuardi and ZeppilliThis is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.Although relatively rare, inherited primitive cardiac disorders (IPCDs) in athletes have a deep social impact since they often present as sudden cardiac death (SCD) of young and otherwise healthy persons. The diagnosis of these conditions is likely underestimated due to the lack of shared clinical criteria and to the existence of several borderline clinical pictures. We will focus on the clinical and molecular diagnosis of the most common IPCDs, namely hypertrophic cardiomyopathies, long QT syndrome, arrhythmogenic right ventricular cardiomyopathy, and left ventricular non-compaction. Collectively, these conditions account for the majority of SCD episodes and/or cardiologic clinical problems in athletes. In addition to the clinical and instrumental tools for the diagnosis of IPCD, the viral technological advances in genetic testing have facilitated the molecular confirmation of these conditions. However, genetic testing presents several issues: the limited sensitivity (globally, around 50%), the low prognostic predictive value, the probability to find pathogenic variants in different genes in the same patient, and the risk of non-interpretable results. In this review, we will analyze the pros and cons of the different clinical approaches for the presymptomatic identification, the diagnosis and management of IPCD athletes, and we will discuss the indications to the genetic testing for patients and their relatives, particularly focusing on the most complex scenarios, such as presymptomatic tests, uncertain results, and unexpected findings. athletessudden cardiac deathgeneticsmedicalhypertrophic cardiomyopathylong QT syndromearrhythmogenic right ventricular displasiaisolated non-compact myocardium ==== Body Introduction Inherited primitive cardiac disorders (IPCDs) comprise a wide and heterogeneous group of conditions. Two major subcategories of IPCDs are universally recognized: primitive cardiomyopathies and primitive electric disorders of the heart. Primitive cardiomyopathies can be defined as disorders characterized by morphologically and functionally abnormal myocardium, in the absence of other diseases that can cause the observed cardiac phenotype (1). This definition is aimed at distinguishing primitive cardiomyopathies from conditions in which the cardiac involvement is secondary to a systemic disorder. Primitive electric disorders are conditions characterized by the presence heart electric conduction disturbances with a morphologically normal myocardium (2). However, while the pathophysiological mechanisms are different, the two groups of IPCDs encompass a continuous spectrum of diseases: rhythm disturbances occur in cardiomyopathies secondary to myocardial disarray and are the leading cause of death. Most IPCDs are Mendelian conditions, most commonly transmitted as autosomal dominant traits with incomplete penetrance, show familial recurrence and have a high degree of genetic heterogeneity (many genes causing the same or similar phenotypes). First level diagnostic tools are standard cardiologic investigations, such as electrocardiography (ECG) and/or echocardiography (EchoCG), which can be supplemented with cardiac magnetic resonance and/or Holter ECG when needed (3). These will not be discussed in this review, since papers more focused on clinical aspects have been recently published (3). The prevalence of IPCDs in general is likely underestimated, due to the reduced penetrance (see below), and to the existence of a wide gray zone between normal and definitely pathological instrumental findings. In the setting of sports medicine, the identification and clinical management of IPCD becomes even more complex: the younger age of the population at risk and the presence of common features between IPCD and athlete heart may delay or prevent a timely diagnosis of these conditions; additionally, the extreme exertion may exacerbate underlying cardiac defects. The onset of IPCD symptoms in athletes is often dramatic, since these are the leading cause of sudden cardiac death (SCD). Thus, presymptomatic identification of affected individuals is of paramount relevance. Following an IPCD diagnosis, it is crucial to evaluate the eligibility for competitive or recreational sports (that implies a lesser cardiac impact), to establish the prognosis, to prevent the occurrence of fatal events, and, last but not least, to evaluate the risk for the athlete’s offspring and sibship. The main relevance of genetic testing for IPCDs is in identifying at risk subjects or to solve diagnostic uncertainties. The most commonly altered genes involved in IPCDs are listed in Table 1. The introduction of next-generation sequencing (NGS) platforms in diagnostic laboratories and the consequential reduction of costs of molecular analysis per patient have improved the diagnostic yield and reduced the time interval between sampling and final reports. However, the data accrued have revealed the complexity of the genetics of IPCD (4–10). With the traditional Sanger sequencing diagnostic approach, single genes were investigated sequentially, one after another, and often testing was interrupted when a pathogenic or likely pathogenic variant was found. NGS-based approaches allow to test several genes simultaneously. As a consequence, the finding of individuals carrying ≥2 rare pathogenic or potentially pathogenic variants in the same or different genes has become not uncommon. At the same time, there has been a surge in the numbers of variants of uncertain significance (VUS) detected. Moreover, it has become clear that allelic variants in the same gene can be associated with different phenotypes, increasing the difficulties inherent to the interpretation of genetic test results. These findings suggest that the variable phenotypic spectrum of IPCDs cannot be only accounted for by classical Mendelian mechanisms, and point toward the involvement of an oligogenic model with strong environmental influences. Table 1 The genes most commonly altered in cardiomyopathies. Gene name Gene symbol Frequency (%) Hypertrophic cardiomyopathy Beta-myosin heavy chain MYH7 15–25 Cardiac myosin-binding protein C MYBPC3 15–25 Cardiac troponin T TNNT2 <5 Cardiac troponin I TNNI3 <5 Alpha-tropomyosin TPM1 <5 LIM-binding domain 3 LBD3 1–5 Ventricular regulatory myosin light chain MYL2 <2 Myosin light chain 3 MYL3 <1 Cardiac muscle alpha actin ACTC1 <1 Long QT syndrome Potassium channel, voltage gated, member 1 (Kv7.1; LQT1) KCNQ1 40–55 Potassium channel, voltage gated, member 2 (Kv11.1; LQT2) KCNH2 30–45 Sodium channel, voltage gated, type V, α subunit (Nav1.5; LQT3) SCN5A 5–10 Arrhythmogenic right ventricular cardiomyopathy Plakophilin 2 PKP2 10–45 Desmoplakin DSP 10–15 Desmoglein DSG2 7–10 Desmocollin 2 DSC-2 2 Junction plakoglobin JUP <1 Genes that are rarely mutated have not been included [adapted from Bos et al. (11), Mizusawa et al. (12), and Iyer and Chin (13)]. In this review, we will focus on the clinical and molecular diagnosis of the most common IPCDs in athletes, namely hypertrophic cardiomyopathies (HCMs), long QT syndrome (LQTS), arrhythmogenic right ventricular cardiomyopathy (ARVC), and left ventricular non-compaction (LVNC). We will also discuss about the TTN gene, one of the largest genes in our genome encoding for the giant protein Titin, which is often altered in different clinical conditions. We will finally discuss the prognostic utility of genetic testing, and the counseling approaches to IPCD patients and their families. Hypertrophic Cardiomyopathies Hypertrophic cardiomyopathies belong to the wider spectrum of cardiomyopathies that include also the dilative and restrictive phenotypes. HCM is diagnosed on the basis of left ventricular hypertrophy, in the absence of abnormal loading conditions. The estimated prevalence in the young adults is about 0.1–0.2%, which does not likely reflect the prevalence in the general population that is expected to be higher (14): available data have been obtained through clinical studies, and thus do not take into account the ascertainment biases related to later onset of symptoms and to the presence of borderline patients (14). According to studies performed in the US, HCM is the most common cause of SCD in young athletes (15–17); it is noteworthy that in countries where the preparticipation screening by ECG is mandatory by law, the incidence of HCM as cause of SCD among athletes is dramatically lower (18). About half of HCM cases are familial with an autosomal dominant pattern of inheritance. More than 20 genes have been related to HCM: b-myosin heavy chain (MYH7) and cardiac myosin-binding protein C (MYBPC3) account for about 50% of the cases; the other genes are rarely affected, with some involved in a single family so far. Incomplete penetrance is an important issue for HCM management. Environmental factors (such as intense training) and/or modifier genes may increase the risk of clinical manifestations especially during exercise or sport. If this was the case, one should expect to find more frequently clinical/instrumental signs of HCM among athletes compared with the general population. However, to the best of our knowledge, there are no available data on the true prevalence of HCM among professional athletes. Since the main complication of this condition is sudden death, the development of primary prevention programs aimed at identifying at risk subjects is very important, despite the relatively low frequency of HCM (16–18). There are not yet enough sensitive clinical markers that may help to identify HCM patients at risk of sudden death. The most reliable predictors are family history of sudden death related to HCM, syncope or presyncope events, ventricular tachyarrhythmia, marked hypotension during training, extreme left ventricle hypertrophy, and extended late enhancement at cardiac MRI (19–22), but their performance is far from satisfactory. A quantitative approach for the assessment of the risk of sudden death in HCM has been reported by O’Mahony et al. (23), the so-called HCM risk SCD. In this case, the risk is estimated on the basis of data collected in a retrospective longitudinal study by taking into account different variables. Rather than for patients with clear HCM phenotypes, genetic testing may be useful for the proper interpretation of borderline patients falling in the gray zone. However, reduced penetrance, genetic heterogeneity, and high VUS frequency make the interpretation of the clinical significance of genetic variants challenging. Furthermore, a preliminary NGS-based study reported the occurrence of double heterozygosity in a high proportion of HCM patients, 2 of the 11 patients with pathogenic variants. These patients were reported as having a more severe phenotype compared with patients with a single disease causing variant (24). With regard to practical implications for molecular diagnosis, according to the guidelines of the European Society of Cardiology (ESC), genetic testing could be offered to all patients fulfilling the HCM diagnostic criteria. Irrespective of the sequencing methodology employed, genetic analysis should include the most commonly implicated sarcomere protein genes (1). Following the identification of a definite pathogenic variant in the proband, genetic testing can be offered to all relatives on a voluntary basis. If no causative variants are found in the proband, relatives should be advised to undergo clinical reassessment should symptoms of HCM manifest. Long QT Syndrome Long QT syndrome is defined by the finding of a prolonged QT interval in standard ECG recording. It is generally accepted that the normal duration of the QT interval is 0.37–0.44 s. Based on this criterion, the diagnosis of LQTS is apparently easy, but 15% of subjects in the general population have a QT interval >0.44 s (0.44–0.47 s) and 25–35% of individuals with a pathogenic variant in one of the LQTS genes has a normal QT interval (25, 26). This latter observation deserves some additional comments: at this stage, it is very difficult to establish whether the finding of a variant considered disease causing in asymptomatic patients is due to reduced penetrance or if, in the light of more recent concepts of molecular genetics, it is the consequence of a wrong interpretation, and the observed DNA change in a VUS, or even a rare benign, not clinically relevant, variation. Since the diagnostic value of QT interval measurement on its own is not sufficient, a scoring method based on multiple parameters is currently used [Table 2; (27)]. LQTS belongs to the wider nosologic group of the channelopathies, and its cumulative prevalence is about 1/2,500: as in the case of HCM, this is likely an underestimate, due to the wide phenotypic heterogeneity. The majority of cases are familial (about 90%). As in the case of HCM, few genes account for the vast majority of cases: specifically, defects in KCNQ1, KCNH2, and SCN5A are found in about 80% of patients. Double heterozygotes are not uncommon: two pathogenic or likely pathogenic variants in different genes are observed in about 10% of patients, and these often display more severe phenotypes (28). Table 2 LQTS diagnostic criteria [from Schwartz and Crotti (25)]. Points Electrocardiographic findingsa A. QTcb   ≥480 ms 3   460–479 ms 2   450–459 ms (in males) 1 B. QTcb fourth minute of recovery from exercise stress   Test ≥480 ms 1 C. Torsade de pointesc 2 D. T wave alternans 1 E. Notched T wave in 3 leads 1 F. Low heart rate for aged 0.5 Clinical history A. Syncopec   With stress 2   Without stress 1 B. Congenital deafness 0.5 Family history A. Family members with definite LQTSe 1 B. Unexplained sudden cardiac death below 30 years of age among immediate family memberse 0.5 aIn the absence of medications or disorders known to affect these electrocardiographic features. bQTc calculated by Bazett’s formula where QTc = QT/√RR. cMutually exclusive. dResting heart rate below the second percentile for age. eThe same family member cannot be counted in A and B. Score: ≤1 point: low probability of LQTS; 1.5–3 points: intermediate probability of LQTS; ≥3.5 points high probability. The diagnosis of LQTS in athletes is complicated by the correlation between duration of the QTc interval and exercise, and the extreme variation of heart rate reached by athletes. In two studies, the prevalence of LQT was 0.6 and 0.4% in an Italian and a British athlete population, respectively (29, 30), that is about 10- to 15-fold higher than in the general population. In the British study, molecular screening of KCNQ1, KCNH2, and SCN5A was performed in five of the seven patients with LQT, three of whom had a QT interval >0.50 s and additional signs reinforcing the suspicion of LQTS; a pathogenic variant was found only in one patient. However, the apparently low yield of genetic testing in this cohort of patients could be related to technical limitations. A proper diagnosis of LQTS in athletes is of particular relevance: besides the obvious implications for the patient and the family, it entails also important career implications, since it is suggested that it may represent a contraindication to competitive sport disciplines involving moderate- and high-intensity strenuous exertion (31–33). Arrhythmogenic Right Ventricular Cardiomyopathy Arrhythmogenic right ventricular cardiomyopathy is a cardiac muscle disease characterized by life-threatening ventricular arrhythmias. The estimated prevalence is about 1:2,500–5,000. ARVC is considered one of the major causes of sudden death in young individuals and in athletes (18). ARVC is generally associated with ECG alterations, including negative T wave in right precordial leads, ventricular arrhythmias with a left bundle branch block morphology, epsilon waves, and others. However, some of these abnormalities are not specific and may be found in other pathological conditions with a different prognosis, such as myocarditis (34). The extensive use of ECG screening may help the sports physician to suspect the diagnosis, while genetic testing may be very useful for the differential diagnosis with more benign conditions. Cardiac pathology shows dystrophy of the right ventricular myocardium with fibrofatty replacement. The clinical picture may include a subclinical asymptomatic phase; ventricular fibrillation, or an electrical disorder with palpitations and syncope, due to tachyarrhythmias of right ventricular origin, may be the first presentation. Most ARVC genes encode for proteins of mechanical cell junctions (DSC2, DSG2, DSP, PKP2, JUP, and DES), while others encode for structural proteins of the nuclear membrane (LMNA and TMEM43) or membrane channels (RYR2). ARVC has an autosomal dominant pattern of inheritance with incomplete penetrance. Pathogenic variations in the nine ARVC genes identified so far account for about 50% of cases. Double heterozygotes have been reported also in this condition. Clinical diagnosis may be achieved by demonstrating functional and structural alterations of the right ventricle, depolarization and repolarization abnormalities, arrhythmias with left bundle branch block morphology, and fibrofatty replacement upon endomyocardial biopsy [see Basso et al. (35) for a review]. Albeit rare, the diagnosis of ARVC is of crucial importance for athletes, due to the risk of sudden death: the condition was originally described as the most common cause of death in sportsmen. However, it is now evident that the condition has a wide phenotypic variability, including very mild asymptomatic cases: sport activity may increase the risk of ventricular arrhythmias in asymptomatic subjects with pathogenic variants in desmosomal genes (36). Left Ventricular Non-Compaction Left ventricular non-compaction is due to the precocious arrest of myocardial compaction during the first weeks of the embryonic development. This causes persistence of prominent trabeculae in the ventricular cavity. The disease spectrum is very wide: the first reported cases were of patients with a marked dilation of the left ventricle and high risk of death (37–40), but asymptomatic and barely progressive segmental forms have also been reported, in which the lack of compaction involves only part of the left ventricle. LVNC is a rare disorder with prevalence <0.1%, although it has been increasingly diagnosed over the last few years. Similar to many other rare disorders, with increasing knowledge its diagnosis has become more common, and among newly identified cases, there is an increasing proportion of asymptomatic subjects, including athletes, with mild phenotypic expression. Indeed, heart hypertrabeculation has been observed in up to 18.3% of athletes (41), about 8% of which fulfill the diagnostic criteria of LVNC. In our experience, a multiparametric evaluation, based on morphological and functional parameters, such as the thickness of the residual compact layer and the presence of major conduction defects and arrhythmias, may help to discriminate between true cardiomyopathies and “benign” forms of LVNC (42). In the latter group of patients, the risk of sudden death, heart failure, and life-threatening arrhythmias is likely low. In any case, close follow-up is still recommended. Genetic testing is not particularly useful for the molecular confirmation of LVNC for at least two reasons: the detection rate of pathogenic variants is relatively low, about 40%, and the genes involved in LVNC are also responsible also for other cardiomyopathies, complicating the interpretation of positive test results (43). Based on these findings, it has been proposed that LVNC may be a phenotypic variant of other cardiomyopathies, characterized by impaired general development of the sarcomeric proteins: this pathogenic model is supported by the cooccurrence in the same family of LVNC and different cardiomyopathies. It is conceivable that genetic background and environmental factors may play a relevant role in the onset of LVNC (44). Regarding the relationship with sport activity, the number of incidental diagnoses has increased over time, often in asymptomatic athletes. Of note, hypertrabeculation of the left ventricle may physiologically occur in athletes, particularly in elite and black sportsmen. Thus, it is crucial to distinguish between the true cardiomyopathy and the benign segmental LVNC for the assessment of the risk of serious life-threatening events (45). Titin: A Titan or A Giant with Clay Feet? Although if it is only one of the genes involved in cardiomyopathies, the titin gene (TTN) deserves a separate discussion due to its peculiarities. TTN is one of the largest genes in our genome and encodes for the largest human protein. The titin protein has several functions in both cardiac and skeletal muscle. Due to the size, prior to the advent of NGS, the mutational analysis of TTN was limited to few exons. The exact number of isoforms is unknown, although it has been estimated that at least one-third of TTN exons may give raise to alternative splicing events (46, 47). TTN has been associated with both dominant and recessive disorders and is currently considered one of the most commonly altered genes in human disease (48), causing at least 10 different conditions, involving skeletal muscle, heart, or both. However, accruing data on genomic variations in the general population have shown that rare TTN variants overall are common, with at least 2–3% of healthy individuals bearing monoallelic truncating mutations. Rare and private missense variants are extremely common as well (47, 49). We should then expect a prevalence of recessive pathogenic variants of at least 1/4,000–10,000, much higher than the cumulative prevalence of titinopathies. Thus, it seems that at least a part, if not the majority, of truncating TTN variants is benign and does not cause pathological phenotypes on their own. The lack of pathogenicity of TTN alleles potentially causing complete loss of function could be explained by alternative splicing events rescuing the gene function. Based on these observations, the pathogenic role of TTN variants should be assessed cautiously, especially for the potential application to presymptomatic-predictive testing in healthy relatives of patients in whom TTN alterations have been detected. It is likely that in the few next years, with accruing genomic data in the general population and the spreading of NGS platforms for the diagnosis of cardiomyopathies, the pathogenicity of TTN variants will be largely elucidated. Discussion In this review, we have highlighted critical aspects associated with the clinical and genetic diagnosis of the IPCDs. The most recent findings on the variability of the human genome are quickly changing the approach to DNA variant interpretation. Indeed, a systematic assessment of variants, including those previously interpreted as pathogenic, is ongoing for several genes associated with inherited conditions. Overall, the following issues are associated with all types of IPCD and complicate their diagnosis and management: (1) wide genetic heterogeneity, (2) incomplete penetrance, (3) relatively high frequency of double heterozygotes, and (4) effect of environmental factors (largely unknown as well, besides sport activity). In the light of these characteristics, IPCDs could be considered as complex traits determined by the predominant effect of single gene variants, rather than as monogenic disorders. Considering IPCDs as oligogenic multifactorial disorders has three main implications: (1) genotype–phenotype correlations are unclear, (2) difficulty in establishing prognosis and risks for patients, and consequently, (3) genetic testing has a limited predictive power both in affected patients and in asymptomatic relatives at risk. Therefore, in our opinion, the use of the risk figures estimated according to Mendelian inheritance is not fully appropriate for predictive purposes. It could be useful to develop a risk assessment model similar to those applied for the familial predisposition to breast cancer, taking into account the presence of multiple factors (e.g., family history, level of exposure to physical activity, presence of multiple gene variants) (50). Indeed, with few exceptions (51–53), due to poor genotype–phenotype correlations, results of clinical investigations provide better prognostic information than those of genetic testing. These problems become even more complicated in athletes, who can display some features resembling those of IPCD as a consequence of physiological rearrangements of the myocardium with training. These subjects, who are mostly young, present some additional issues: the ascertainment of a variant known to be the cause of an IPCD phenotype may have strong implications for the prosecution of their sport career, for reproductive choices, and for their families. The offer of a predictive test for relatives should be considered with caution, and only when the pathogenicity of the variant detected in the proband has been clearly established according to consensus criteria (54). The opportunity of testing underage relatives of athletes should also be carefully scrutinized, especially when presymptomatic diagnosis may be beneficial, such as in the case of LQTS, which can manifest as infant sudden death. Variant interpretation is the main issue in molecular diagnosis of IPCDs, and it is further exacerbated by the small size of many pedigrees. This hampers analysis of variant segregation with respect to the phenotype, one of the most useful points of evidence for clinical interpretation of genetic variants, as well the estimation of penetrance values. Ideally, novel or unclassified variants should be validated by functional studies; however, with few exceptions, these are not performed in clinical diagnostic laboratories and are associated with several issues, such as the choice of the cellular model and feasibility, since most sarcomeric mRNAs are quite large and not easily manageable. In conclusion, unless validated functional tests have been performed, the main hints suggesting pathogenicity of a variant are its identification in different patients or de novo occurrence. Another open issue is the significance of double heterozygosity and related counseling. So far, it is common opinion that these subjects may in general display a more severe phenotype but the cohorts published so far are too small to draw definite conclusions. TTN deserves separate considerations. In particular, given the difficulties inherent with TTN molecular testing and interpretation, one might wonder whether it is appropriate to include this gene in diagnostic panels and in genetic reports for patients, or rather, whether it should still be investigated in research settings for epidemiological purposes. These might shed light on the pathogenicity of TTN truncating and missense variants, as well as and on their clinical relevance; indeed, it remains to be established if TTN variants act as main phenotypic drivers and/or as a risk factor for the appearance of the clinical manifestation of some IPCDs, insufficient alone to determine a phenotype. Similar to TTN, also in HCM the large amount of whole exome data that are accumulating in the different databases is disclosing that presumptive pathogenic variants can be found in “controls” at a higher rate than expected for the prevalence of the condition (55–57). On the one hand, this excess of pathogenic variants could be accounted for by reduced penetrance: carriers identified in the general population may or may not develop signs of HCM over time, but should be considered as asymptomatic subjects. On the other hand, these findings may indicate that the effects of these gene variants are too weak to cause appearance of the phenotype on their own. In conclusion, the refinement of clinical diagnosis of IPCD, coupled with the new technological tools available in molecular genetics, has opened the Pandora box of cardiac primitive defects. Now, the pieces of this puzzle need to be reconstructed in order to provide patients and athletes with more accurate information and best care. But, there is still a long way to go. Author Contributions All authors listed, have made substantial, direct and intellectual contribution to the work, and approved it for publication. Conflict of Interest Statement The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. ==== Refs References 1 Elliott PM Anastasakis A Borger MA Borggrefe M Cecchi F Charron P 2014 ESC Guidelines on diagnosis and management of hypertrophic cardiomyopathy: the Task Force for the Diagnosis and Management of Hypertrophic Cardiomyopathy of the European Society of Cardiology (ESC) . Eur Heart J (2014 ) 35 :2733 –79 .10.1093/eurheartj/ehu284 25173338 2 Bezzina CR Lahrouchi N Priori SG . Genetics of sudden cardiac death . Circ Res (2015 ) 116 :1919 –36 .10.1161/CIRCRESAHA.116.304030 26044248 3 Myerson M Sanchez-Ross M Sherrid MV . Preparticipation athletic screening for genetic heart disease . Prog Cardiovasc Dis (2012 ) 54 :543 –52 .10.1016/j.pcad.2012.03.003 22687598 4 Bortot B Athanasakis E Brun F Rizzotti D Mestroni L Sinagra G High-throughput genotyping robot-assisted method for mutation detection in patients with hypertrophic cardiomyopathy . Diagn Mol Pathol (2011 ) 20 :175 –9 .10.1097/PDM.0b013e31820b34fb 21817903 5 Fokstuen S Munoz A Melacini P Iliceto S Perrot A Ozcelik C Rapid detection of genetic variants in hypertrophic cardiomyopathy by custom DNA resequencing array in clinical practice . J Med Genet (2011 ) 48 :572 –6 .10.1136/jmg.2010.083345 21239446 6 Meder B Haas J Keller A Heid C Just S Borries A Targeted next-generation sequencing for the molecular genetic diagnostics of cardiomyopathies . Circ Cardiovasc Genet (2011 ) 4 :110 –22 .10.1161/CIRCGENETICS.110.958322 21252143 7 Faita F Vecoli C Foffa I Andreassi MG Next generation sequencing in cardiovascular diseases . World J Cardiol (2012 ) 4 :288 –95 .10.4330/wjc.v4.i10.288 23110245 8 Lopes LR Rahman MS Elliott PM . A systematic review and meta-analysis of genotype-phenotype associations in patients with hypertrophic cardiomyopathy caused by sarcomeric protein mutations . Heart (2013 ) 99 :1800 –11 .10.1136/heartjnl-2013-303939 23674365 9 Loar RW Bos JM Will ML Ommen SR Ackerman MJ Genotype-phenotype correlations of hypertrophic cardiomyopathy when diagnosed in children, adolescents, and young adults . Congenit Heart Dis (2015 ) 10 :529 –36 .10.1111/chd.12280 26061417 10 Lopes LR Zekavati A Syrris P Hubank M Giambartolomei C Dalageorgou C Genetic complexity in hypertrophic cardiomyopathy revealed by high-throughput sequencing . J Med Genet (2013 ) 50 :228 –39 .10.1136/jmedgenet-2012-101270 23396983 11 Bos JM Towbin JA Ackerman MJ . Diagnostic, prognostic, and therapeutic implications of genetic testing for hypertrophic cardiomyopathy . J Am Coll Cardiol (2009 ) 54 (3 ):201 –11 .10.1016/j.jacc.2009.02.075 19589432 12 Mizusawa Y Horie M Wilde AA Genetic and clinical advances in congenital long QT syndrome . Circ J (2014 ) 78 :2827 –33 .10.1253/circj.CJ-14-0905 25274057 13 Iyer VR Chin AJ . Arrhythmogenic right ventricular cardiomyopathy/dysplasia (ARVC/D) . Am J Med Genet C Semin Med Genet (2013 ) 163C :185 –97 .10.1002/ajmg.c.31368 23824749 14 Maron BJ Gardin JM Flack JM Gidding SS Kurosaki TT Bild DE Prevalence of hypertrophic cardiomyopathy in a general population of young adults. Echocardiographic analysis of 4111 subjects in the CARDIA Study. Coronary Artery Risk Development in (Young) Adults . Circulation (1995 ) 92 :785 –9 .10.1161/01.CIR.92.4.785 7641357 15 Maron BJ Haas TS Doerer JJ Thompson PD Hodges JS Comparison of U.S. and Italian experiences with sudden cardiac deaths in young competitive athletes and implications for preparticipation screening strategies . Am J Cardiol (2009 ) 104 :276 –80 .10.1016/j.amjcard.2009.03.037 19576360 16 Maron BJ Distinguishing hypertrophic cardiomyopathy from athlete’s heart physiological remodelling: clinical significance, diagnostic strategies and implications for preparticipation screening . Br J Sports Med (2009 ) 43 :649 –56 .10.1136/bjsm.2008.054726 19734498 17 Maron BJ Sudden death in hypertrophic cardiomyopathy . J Cardiovasc Transl Res (2009 ) 2 :368 –80 .10.1007/s12265-009-9147-0 20559995 18 Corrado D Basso C Rizzoli G Schiavon M Thiene G . Does sports activity enhance the risk of sudden death in adolescents and young adults? J Am Coll Cardiol (2003 ) 42 :1959 –63 .10.1016/j.jacc.2003.03.002 14662259 19 Jarcho JA McKenna W Pare JA Solomon SD Holcombe RF Dickie S Mapping a gene for familial hypertrophic cardiomyopathy to chromosome 14q1 . N Engl J Med (1989 ) 321 :1372 –8 .10.1056/NEJM198911163212005 2811944 20 Elliott PM Gimeno JR Thaman R Shah J Ward D Dickie S Historical trends in reported survival rates in patients with hypertrophic cardiomyopathy . Heart (2006 ) 92 :785 –91 .10.1136/hrt.2005.068577 16216855 21 Spirito P Bellone P Harris KM Bernabo P Bruzzi P Maron BJ . Magnitude of left ventricular hypertrophy and risk of sudden death in hypertrophic cardiomyopathy . N Engl J Med (2000 ) 342 :1778 –85 .10.1056/NEJM200006153422403 10853000 22 Delcrè SD Di Donna P Leuzzi S Miceli S Bisi M Scaglione M Relationship of ECG findings to phenotypic expression in patients with hypertrophic cardiomyopathy: a cardiac magnetic resonance study . Int J Cardiol (2013 ) 167 :1038 –45 .10.1016/j.ijcard.2012.03.074 22464482 23 O’Mahony C Jichi F Pavlou M Monserrat L Anastasakis A Rapezzi C A novel clinical risk prediction model for sudden cardiac death in hypertrophic cardiomyopathy (HCM risk-SCD) . Eur Heart J (2014 ) 35 :2010 –20 .10.1093/eurheartj/eht439 24126876 24 Golbus JR Puckelwartz MJ Dellefave-Castillo L Fahrenbach JP Nelakuditi V Pesce LL Targeted analysis of whole genome sequence data to diagnose genetic cardiomyopathy . Circ Cardiovasc Genet (2014 ) 7 :751 –9 .10.1161/CIRCGENETICS.113.000578 25179549 25 Vyas H Hejlik J Ackerman MJ . Epinephrine QT stress testing in the evaluation of congenital long-QT syndrome: diagnostic accuracy of the paradoxical QT response . Circulation (2006 ) 113 :1385 –92 .10.1161/CIRCULATIONAHA.105.600445 16534005 26 Priori SG Napolitano C Schwartz PJ . Low penetrance in the long-QT syndrome: clinical impact . Circulation (1999 ) 99 :529 –33 .10.1161/01.CIR.99.4.529 9927399 27 Schwartz PJ Crotti L QTc behavior during exercise and genetic testing for the long-QT syndrome . Circulation (2011 ) 124 :2181 –4 .10.1161/CIRCULATIONAHA.111.062182 22083145 28 Tester DJ Will ML Haglund CM Ackerman MJ . Compendium of cardiac channel mutations in 541 consecutive unrelated patients referred for long QT syndrome genetic testing . Heart Rhythm (2005 ) 2 :507 –17 .10.1016/j.hrthm.2005.07.012 15840476 29 Corrado D Basso C Pavei A Michieli P Schiavon M Thiene G . Trends in sudden cardiovascular death in young competitive athletes after implementation of a preparticipation screening program . JAMA (2006 ) 296 :1593 –601 .10.1001/jama.296.13.1593 17018804 30 Basavarajaiah S Wilson M Whyte G Shah A Behr E Sharma S . Prevalence and significance of an isolated long QT interval in elite athletes . Eur Heart J (2007 ) 28 :2944 –9 .10.1093/eurheartj/ehm404 17947213 31 Zipes DP Ackerman MJ Estes NA IIIGrant AO Myerburg RJ Van Hare G Task force 7: arrhythmias . J Am Coll Cardiol (2005 ) 45 :1354 –63 .10.1016/j.jacc.2005.02.014 15837287 32 Pelliccia A Corrado D Bjørnstad HH Panhuyzen-Goedkoop N Urhausen A Carre F Recommendations for participation in competitive sport and leisure-time physical activity in individuals with cardiomyopathies, myocarditis and pericarditis . Eur J Cardiovasc Prev Rehabil (2006 ) 13 :876 –85 .10.1097/01.hjr.0000238393.96975.32 17143118 33 Biffi A Delise P Zeppilli P Giada F Pelliccia A Penco M Italian Society of Sports Cardiology and Italian Sports Medicine Federation. Italian cardiological guidelines for sports eligibility in athletes with heart disease: part 1 . J Cardiovasc Med (Hagerstown) (2013 ) 14 :477 –99 .10.2459/JCM.0b013e32835f6a21 23615077 34 Dello Russo A Pieroni M Santangeli P Bartoletti S Casella M Pelargonio G Concealed cardiomyopathies in competitive athletes with ventricular arrhythmias and an apparently normal heart: role of cardiac electroanatomical mapping and biopsy . Heart Rhythm (2011 ) 8 :1915 –22 .10.1016/j.hrthm.2011.07.021 21782775 35 Basso C Corrado D Marcus FI Nava A Thiene G . Arrhythmogenic right ventricular cardiomyopathy . Lancet (2009 ) 373 :1289 –300 .10.1016/S0140-6736(09)60256-7 19362677 36 James CA Bhonsale A Tichnell C Murray B Russell SD Tandri H Exercise increases age-related penetrance and arrhythmic risk in arrhythmogenic right ventricular dysplasia/cardiomyopathy-associated desmosomal mutation carriers . J Am Coll Cardiol (2013 ) 62 :1290 –7 .10.1016/j.jacc.2013.06.033 23871885 37 Kohli SK Pantazis AA Shah JS Adeyemi B Jackson G McKenna WJ Diagnosis of left-ventricular non-compaction in patients with left-ventricular systolic dysfunction: time for a reappraisal of diagnostic criteria? Eur Heart J (2008 ) 29 :89 –95 .10.1093/eurheartj/ehm481 17993472 38 Bellavia D Michelena HI Martinez M Pellikka PA Bruce CJ Connolly HM Speckle myocardial imaging modalities for early detection of myocardial impairment in isolated left ventricular non-compaction . Heart (2010 ) 96 :440 –7 .10.1136/hrt.2009.182170 19966109 39 Nucifora G Aquaro GD Pingitore A Masci PG Lombardi M . Myocardial fibrosis in isolated left ventricular non-compaction and its relation to disease severity . Eur J Heart Fail (2011 ) 13 :170 –6 .10.1093/eurjhf/hfq222 21208941 40 Melendez-Ramirez G Castillo-Castellon F Espinola-Zavaleta N Meave A Kimura-Hayama ET . Left ventricular noncompaction: a proposal of new diagnostic criteria by multidetector computed tomography . J Cardiovasc Comput Tomogr (2012 ) 6 :346 –54 .10.1016/j.jcct.2012.07.001 22981859 41 Gati S Chandra N Bennett RL Reed M Kervio G Panoulas VF Increased left ventricular trabeculation in highly trained athletes: do we need more stringent criteria for the diagnosis of left ventricular non-compaction in athletes? Heart (2013 ) 99 :401 –8 .10.1136/heartjnl-2012-303418 23393084 42 Poscolieri B Bianco M Vessella T Gervasi S Palmieri V Zeppilli P Identification of benign form of ventricular non-compaction in competitive athletes by multiparametric evaluation . Int J Cardiol (2014 ) 176 :1134 –6 .10.1016/j.ijcard.2014.07.288 25156849 43 Hoedemaekers YM Caliskan K Michels M Frohn-Mulder I van der Smagt JJ Phefferkorn JE The importance of genetic counseling, DNA diagnostics, and cardiologic family screening in left ventricular noncompaction cardiomyopathy . Circ Cardiovasc Genet (2010 ) 3 :232 –9 .10.1161/CIRCGENETICS.109.903898 20530761 44 McNally EM Barefield DY Puckelwartz MJ . The genetic landscape of cardiomyopathy and its role in heart failure . Cell Metab (2015 ) 21 :174 –82 .10.1016/j.cmet.2015.01.013 25651172 45 Caselli S Attenhofer Jost CH Jenni R Pelliccia A . Left ventricular noncompaction diagnosis and management relevant to pre-participation screening of athletes . Am J Cardiol (2015 ) 116 :801 –8 .10.1016/j.amjcard.2015.05.055 26141199 46 Neiva-Sousa M Almeida-Coelho J Falcão-Pires I Leite-Moreira AF . Titin mutations: the fall of Goliath . Heart Fail Rev (2015 ) 20 :579 –88 .10.1007/s10741-015-9495-6 26024954 47 Gerull B . The rapidly evolving role of titin in cardiac physiology and cardiomyopathy . Can J Cardiol (2015 ) 31 :1351 –9 .10.1016/j.cjca.2015.08.016 26518445 48 Chauveau C Rowell J Ferreiro A . A rising titan: TTN review and mutation update . Hum Mutat (2014 ) 35 :1046 –59 .10.1002/humu.22611 24980681 49 Herman DS Lam L Taylor MR Wang L Teekakirikul P Christodoulou D Truncations of titin causing dilated cardiomyopathy . N Engl J Med (2012 ) 366 :619 –28 .10.1056/NEJMoa1110186 22335739 50 Mavaddat N Peock S Frost D Ellis S Platte R Fineberg E Cancer risks for BRCA1 and BRCA2 mutation carriers: results from prospective analysis of EMBRACE . J Natl Cancer Inst (2013 ) 105 :812 –22 .10.1093/jnci/djt095 23628597 51 Semsarian C Healey MJ Fatkin D Giewat M Duffy C Seidman CE A polymorphic modifier gene alters the hypertrophic response in a murine model of familial hypertrophic cardiomyopathy . J Mol Cell Cardiol (2001 ) 33 :2055 –60 .10.1006/jmcc.2001.1466 11708849 52 Suzuki M Carlson KM Marchuk DA Rockman HA . Genetic modifier loci affecting survival and cardiac function in murine dilated cardiomyopathy . Circulation (2002 ) 105 :1824 –9 .10.1161/01.CIR.0000014926.32463.89 11956126 53 Wheeler M Pavlovic A DeGoma E Salisbury H Brown C Ashley EA . A new era in clinical genetic testing for hypertrophic cardiomyopathy . J Cardiovasc Transl Res (2009 ) 2 :381 –91 .10.1007/s12265-009-9139-0 20559996 54 Richards S Aziz N Bale S Bick D Das S Gastier-Foster J Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology . Genet Med (2015 ) 17 :405 –24 .10.1038/gim.2015.30 25741868 55 Andreasen C Nielsen JB Refsgaard L Holst AG Christensen AH Andreasen L New population-based exome data are questioning the pathogenicity of previously cardiomyopathy-associated genetic variants . Eur J Hum Genet (2013 ) 21 :918 –28 .10.1038/ejhg.2012.283 23299917 56 Golbus JR Puckelwartz MJ Fahrenbach JP Dellefave-Castillo LM Wolfgeher D McNally EM . Population-based variation in cardiomyopathy genes . Circ Cardiovasc Genet (2012 ) 5 :391 –9 .10.1161/CIRCGENETICS.112.962928 22763267 57 Pan S Caleshu CA Dunn KE Foti MJ Moran MK Soyinka O Cardiac structural and sarcomere genes associated with cardiomyopathy exhibit marked intolerance of genetic variation . Circ Cardiovasc Genet (2012 ) 5 :602 –10 .10.1161/CIRCGENETICS.112.963421 23074333

PMC005xxxxxx/PMC5000294.txt

==== Front Sci RepSci RepScientific Reports2045-2322Nature Publishing Group srep3239110.1038/srep32391ArticleThe optical measurement of large cluster tracks in a gas jet Chen Zhiyuan 1Liu Dong 1Han Jifeng a1Bai Lixin 11 Key Laboratory of Radiation Physics and Technology of the Ministry of Education, Institute of Nuclear Science and Technology, Sichuan University, Chengdu 610064, Chinaa hanjf@scu.edu.cn26 08 2016 2016 6 3239125 04 2016 04 08 2016 Copyright © 2016, The Author(s)2016The Author(s)This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/We propose an optical method based on Rayleigh scattering for the direct measurement of cluster tracks produced by a high-pressure gas jet. The tracks of the argon and methane clusters are acquired by a high-speed camera. It is found that the cluster sizes of these tracks are within the range of 7E + 03~1E + 07 for argon and 2E + 06~4E + 08 for methane. Most argon tracks are continuous and their intensity changes gradually, while the majority of the methane tracks are separated into discrete fractions and their intensity alters periodically along the flight path, which may indicate the methane clusters are more unstable and easily to break up. Special methane clusters which may fly at an axial velocity of less than 2.5m/s are also found. This method is very sensitive to large gas cluster and has broad application prospects in cluster physics. ==== Body Clusters are made of atoms or molecules by physical or chemical bonding forces. The interaction of an intense short-pulsed laser with Nano-sized clusters has been a popular topic of study1 for its potential application in neutron sources2, X-ray sources3, fast electrons and ions4, etc. The supersonic gas jet is produced by opening one nozzle to inject the high pressured gas into the vacuum, the temperature of the gas decreases greatly during the free expansion process and the clusters are formed by Van der Waals forces. The clusters are produced by gas jet technique and its size and density distribution are required to analyze the laser-cluster interaction dynamics or calculate the yield rates. The cluster size is distributed over a broad range from two (dimer) up to many thousands, which can be well approximated by the log-normal function56. Typically the average cluster size and density7 are used to represent the distribution and normally the average parameters are investigated by the Rayleigh scattering method combining other techniques such as mass spectroscopy8, time of flight mass spectrometer9, and electron diffraction10, etc. Although large clusters are preferred for ensuring higher efficiency and yield rate, there are few studies of large clusters611 which are rare and hard to detect. In this study, an optical method based on Rayleigh scattering is presented to measure the cluster tracks in the gas jet and large argon and methane clusters were discovered. This imaging method can be used to capture the images of cluster tracks and study the cluster interactions with materials, thereby demonstrating broad prospects for application in cluster physics. Traditionally for the axisymmetric gas expansion, the average cluster size Nc (the average number of atoms per cluster) has been characterized by the semi-empirical parameter introduced by Hagena12131415, and it is revised in ref. 18 for very big clusters, which is expressed as Formula (1). Г* is a semi-empirical parameter that takes into account all the factors1216 affecting the average cluster size, such as the gas property (k), the nozzle geometry (deq) and the gas parameters (P0, T0). k is a gas-dependent constant (k~1650 for Ar and 2360 for CH4), the equivalent nozzle diameter deq is assumed to be 0.74d/tanα for monatomic clusters and 0.87d/tanα for diatomic clusters where d is the throat diameter of the nozzle in μm and α is the half opening angle of the nozzle17. T0 is the gas stagnation temperature in Kelvin, and P0 is the stagnation pressure in mbar. The parameter q has commonly been fixed at 0.85 in the scaling law of Hagena1317. The exponent r is the function of q and γ, where γ is the specific heat ratio of the gas which is 5/3 for monatomic gas and 7/5 for diatomic gas respectively1819. However, it must be declared that the Hagena scaling law is not suitable for methane cluster. According to the results from Akiyoshi Murakami et al.20, the relation between Nc and Г* for methane is described in formula (2). The exponent value used to calculate Nc in formula (2) is quite different from that of Hagena, and it is found that the value increases from 3.8 to 7.6 for higher stagnation pressures which is contradictory to Hagena’s formula whose exponent value decreases from 2.35 to 1.8 for higher stagnation pressure P0. Song Li et al.21 calculated the average methane cluster size by measuring the Coulomb explosion method, and it is reported that the average methane cluster size Nc = 1235 when Г* = 520 (deq = 3.85mm, P0 = 30bar, T0 = 296K, Γ* ∝ P0(T)−3.3) and Nc = 6230 when Г* = 1104(P0 = 30bar, T0 = 240 K), and it can be calculated that the exponent value is about 2.15 for these data. It is evident that the above relations between Nc and Г* for methane are not accordant and it is different from the relation of Nc ∝ (Γ*)5.78 which is measured by our experiment and is described in the following “Results and discussion” section. So it can be concluded that the Hagena scaling law is unsuitable for methane cluster, and the exact formula is expecting for further research. In addition, the results of methane cluster size given here are subject to the date from our experiment. The Rayleigh scattering method has been widely applied for cluster-size evaluations22 because it is nondestructive and relatively easy to operate. The scattering signal (S) is proportional to the cluster number density (nc) and the square of the average cluster size (Nc) if the clusters are supposed to have a spherical structure23. The cluster number density (nc) multiply the average cluster size (Nc) is the total number density of clusters, which equals to the product of the monomer density before clustering (n0) and the proportion of the molecules that are formed to clusters (η). The monomer density before clustering (n0) is proportional to the stagnation pressure (P0) before the gas jet, thereby we can obtain where n0 is the monomer density before clustering, and η is the proportion of the monomers that can form clusters. If we suppose η remains constant throughout the duration of one gas jet, then the scattering signal S is proportional to the average cluster size Nc. Experimental set-up The experimental set-up is depicted in Fig. 1. In this study, Z represents the gas jet direction (axial), R represents the laser beam direction (radial), and the axial distance is defined as the distance between the nozzle and the center of the laser beam. And the results are discussed in RZ coordinate system. The gas jet is generated along the axial direction and the continuous laser beam perpendicular to the gas jet is focused into the vacuum chamber along the radial direction to interact with clusters and generate the Rayleigh scattering photons. The power of the laser is fixed at 10 W, the wavelength is 445 nm and the spot diameter is about 5 mm. The laser beam and the scattered light beam are shown by thick black arrows in Fig. 1. A UX50 high-speed camera from Photron Corp is used to capture the 90o Rayleigh scattering lights on the top side of the vacuum chamber. The reflected lights from the wall of vacuum chamber have been disposed to minimize the background noise. The resolution of the camera is 1280*1024, the imaging speed is fixed at 1000 fps (frame per second), and the shutter speed is set to 1/1000 second to acquire entire evolution process. The argon and methane gas with 99.999% purity are employed to produce pure clusters and minimize the influence of impurities. The stagnation pressure of the gas can be adjusted from 10 to 48 bar. A 1500 L/s turbomolecular pump is used to supply a vacuum pressure of approximately 2E-4 Pa. Moreover, the stagnation temperature is maintained at approximately 293 K in the entire test duration. The conical nozzle can be moved along the axial direction in the range of 10–180 mm. The orifice diameter and half cone angle of the nozzle is 0.5 mm and 45o, and the total length of the nozzle is about 3 mm. In addition, the nozzle is opened by a high-speed pulsed solenoid valve from Parker Hannifin Corp. In this study, the open time of the nozzle is fixed to 20 milliseconds (ms) and 20 Rayleigh scattering images are acquired by the high-speed camera for each gas jet. Results and Discussion Large cluster tracks For each gas jet 20 scattering lights images are acquired by the high-speed camera, normally 1~6 of the 20 images contain cluster tracks, which are used as “effective images” to analyze the properties of large cluster. And the rest images without cluster tracks are used as “reference images” to represent the average clusters. The typical “effective image” that contains cluster tracks and corresponding reference image for argon are shown in Fig. 2(a,b). The intensity of the “reference image” is proportional to the average cluster size, and the intensity of the track is proportional to the size of the clusters inside the track. Since the intensity of the track is much bigger than that in the “reference image”, we can infer the cluster size of the track is much larger than the corresponding average, which means the tracks are composed of large clusters. As shown in Fig. 2(a), most argon tracks are continuous and the intensity of them changes gradually, while the majority of the methane tracks are separated into discrete fractions and the intensity of them alters periodically along the flight path as shown in Fig. 3(a). The broken and discrete process is clearly imaged for methane clusters, indicating that methane clusters may be very unstable to fragment into small pieces. It is hard to explain why the intensity alters periodically. We thought possibly it is caused by the asymmetric structure of the large methane cluster, but more studies are needed to verify. Three consecutive tracks of the same methane clusters are acquired at axial position about 15 mm and radial position about 21 mm during one gas jet. The three images are combined together, and the combined tracks are shown in Fig. 3(b). The two boundaries of the three tracks are at the axial positions of 15 mm and 17 mm, which are shown by two black lines in Fig. 3(b). The length of the middle track is about 2.5 mm, as the shutter time for one image is 1 millisecond, then the axial velocity of the clusters can be estimated to be less than 2.5 m/s, which is far smaller than the limit supersonic velocity of methane which is 1140 m/s at room temperature (, where Cp is the specific heat ratio at constant pressure (J. K−1. Kg−1), T0 is stagnation temperature (Klevin)) before the gas jet. The opening angle of this track to the center streamline is about 54o, which is much bigger than the half opening angle of the nozzle which is only 45o. The argon track images at different axial position are combined together into one image, which is shown in Fig. 4. It is found that the cluster tracks are distributed within the wide axial range of 10 to 180 mm and the extended line of the tracks on both sides are intersected to the same point. This verifies the results that the gas jet expends out the nozzle conically and the tracks are large clusters. The full width at half maximum (FWHM) of the tracks are calculated, which is acquired by 1D summation of the pixels along the flight path of the track and get the intensity variation curve perpendicular to the flight path. The curve is fitted with the Gaussian function, and FWHM is calculated from the fitted Gaussian function. The FWHM was used to represent the width of tracks and the peak value was chosen to show intensity of tracks. However no distinct relationship between the width and intensity was found, hence the width of the track cannot be used to represent the cluster size. The FWHM distribution for argon and methane is shown in Fig. 5(a,b). The Gaussian function is used to fit the distribution, and it is found that the mean FWHM is about 0.38 ± 0.12 for argon and 0.32 ± 0.07 for methane. In this test one pixel represents 0.1 mm, which means the spatial resolution of the camera is about 0.1 mm, which is in agreement with the RMS (Root Mean Square) value of the FWHM distribution. However, we believe the width of the track gives one hint about the number of clusters inside the track, and we can imagine there are fewer clusters for the methane track than the argon track because the FWHM is smaller for methane than argon. The cluster sizes of these large clusters can be estimated by comparing the intensity of the tracks in the “effective image” and that in the “reference image” at the same position, whereas the intensity of the “reference image” can be calibrated by the average cluster size of the gas jet. Average cluster size The average cluster size is required to calculate the size of the track. One simple method to estimate the average cluster size is introduced by refs 24,25, which is used broadly by many groups262728. When the relationship between the Rayleigh scattering signal and the stagnation pressure are tested, the average cluster size at different stagnation pressure can be easily calculated by supposing that the minimum detectable cluster size is 100 for the Rayleigh scattering method, which means the average cluster size is 100 when the Rayleigh scattering signal is just above the noise level. In ref. 22, the average cluster size is assumed to be about 100 at the “reasonable” onset point of clustering (when the signal-to-noise ratio ≈2) when the wavelength of the scattering light is 532 nm. The detectable threshold for argon cluster is supposed to be 100 when 532 nm and 526 nm laser are used in refs 25 and 26 respectively. In this work, the average cluster size is measured by using one 532 nm quasi-continuous laser with 5 kHz frequency and one PMT from Hamamatsu. And we suppose the cluster size is 100 when the signal to noise ratio is two. The PMT was set on the top of the vacuum chamber to replace the high-speed camera. The high speed camera is not suitable for this method for it is not sensitive to weak lights and normally the photomultiplier tube (PMT) is used instead. The typical Rayleigh scattering signal at the axial position of 20 mm and stagnation pressure of 35 bar is shown in Fig. 6. The valve is opened during 10 ms and 30 ms. It is found that the Rayleigh scattering signal rises up very quickly once the valve is opened and keeps almost constant during the open period of 20 ms, this means the clusters keeps almost the same average size during the whole gas jet period, which is consistent with the result of the high speed camera. From formula (3) we obtain S ∝ ηPNc, if we suppose η remains constant for different stagnation pressures, then the average cluster size is given by: When the average cluster size Nc1 at stagnation pressure P1 is known, then the average cluster size at other stagnation pressure can be calculated easily from formula (4). In this experiment we suppose Nc of argon is 100 when the scattering signal is about two times of the background noise, which happens when the stagnation pressure is about 10 bar. Since the molecular diameter of methane is larger than that of argon, the minimum detectable cluster size of methane should be smaller than argon for the same Rayleigh scattering system. The same minimum detectable cluster size of 100 is used for methane in this work for conservative purpose. And Nc of methane is supposed to be 100 when the stagnation pressure is about 13 bar. When the scattering signal versus stagnation pressure is tested, the average cluster size for different stagnation pressures can be calculated from formula (4). The Rayleigh scattering signal and the average cluster size at different stagnation pressure for argon are shown in Fig. 7(a,b), and the power function is used to fit the curve. The fitted coefficient of average cluster size versus pressure curve is a little bigger than that of Hagena. The Hagena curve is shown in black in Fig. 7(b) and it is found that the tested average argon cluster size is smaller than that calculated by the scaling law of Hagena (formula 1) at lower stagnation pressures but consistent with each other at higher pressures. This means the scaling law might overestimate the cluster size at lower stagnation pressure, and verifies that this method is effective at higher pressures. The Rayleigh scattering signal and the average cluster size versus stagnation pressure for methane are shown in Fig. 8(a,b). The fitted coefficient of Nc is much bigger than that of argon29, indicating that the average size of methane cluster will increase much heavily for higher stagnation pressure. The fitted coefficient is 5.78, which happens to be the mean value of the two coefficients of 3.8 and 7.6 in ref. 20. The average cluster size of argon and methane at all axial positions can be calculated from the formula (3), at the stagnation pressure of 25~45 bar. The results of argon are in agreement with that in the scaling law of Hagena (1), which indicates the Rayleigh scattering system works properly and the results are reliable. Large cluster size distribution of the tracks The size ratio (RSR) is defined as the ratio of large cluster size of the track in effective image over the average cluster size in reference image. If the average cluster sizes (Nc) are acquired then the large cluster sizes (Nbig) can be calculated from formula (5). For each cluster track, the size ratio can be given by comparing the intensity of the cluster track in effective image and that at the same position in reference image, and the average cluster size can be calculated from the formula (4), therefore the large cluster size can be calculated out. According to the formula (5) the large cluster sizes of argon and methane lie in the range of 7E + 03 ~ 1E + 07 and 2E + 06 ~ 4E + 08, respectively. By combining the error of the average cluster size and the size ratio, the relative error of large cluster size is calculated to be about 20% for argon and 40% for methane. The large cluster size at different axial position and stagnation pressure for argon are shown in Fig. 9(a,b), the Y axis is shown in logarithmic coordinate. There are lots of dots for each axial position in Fig. 9(a) which are corresponding to cluster tracks at different radial position. While a series of dots for each stagnation pressure in Fig. 9(b) corresponding to the tracks at different axial position. The size of large argon cluster exhibits a weak dependence both on the axial positions and stagnation pressures. This means large argon clusters can be found even at lower stagnation pressure or bigger axial positions where the gas density is lower. As for large methane clusters, the sizes are irrelevant to the axial position but have a strong growing trend with the stagnation pressure as shown in Fig. 10(a,b). But different from the argon cluster, there is hardly any methane cluster track when the axial position is bigger than 45 mm. This is possibly because the methane cluster is easier to fragment after long flight trip. Different from the size distribution, the number of tracks is strongly dependent on the axial position or stagnation pressure. The number of tracks versus axial position and stagnation pressure curve for argon and methane are shown in Figs 11 and 12. It is found from Fig. 11(a) that most of the large argon clusters are formed within the axial range of 15 mm to 35 mm, and the number starts to decrease sharply after that, which means these large clusters are unstable and easily to fragment into small pieces. It is found from Fig. 11(b) that more large argon clusters are formed at higher stagnation pressures when the average cluster size is higher. However no distinct relationship is found between the number of methane tracks and the stagnation pressure in Fig. 12(b). More tracks are available at the center of the gas jet both for argon and methane. The total number of methane tracks is only 1/6 of argon tracks, and it is found in Fig. 12(a) that the maximum tracks are available at axial distance of 20 mm, which decreases sharply after 40 mm and none is found after 45 mm. Different from the fact that the number of argon tracks increases linearly with the stagnation pressure, the number variation is heavier for methane tracks and no distinct trend can be found. Discussion Considering the large cluster size and number of tracks, we can conclude that at higher stagnation pressure, much more large argon clusters are formed but the size only increase slightly, while much larger methane clusters are found but the number doesn’t change greatly. This means the argon cluster is inclined to increase the number, while the methane cluster tends to increase the size. The total number of argon tracks is about six times more than methane, while the cluster size of argon is about three times less than methane. Since the scattering intensity of argon is larger than methane at the same stagnation pressure, we can infer that the “number effect” is stronger than the “size effect”. Most of the methane cluster tracks are separated into discrete fractions and the intensity of them alters along the flight path. This phenomenon indicates that these methane clusters may be very unstable to fragment into small pieces. It is hard to explain why the intensity alters periodically, and more studies are needed to verify that. Conclusion In conclusion, we have demonstrated an optical method based on Rayleigh scattering for measuring the cluster tracks in gas jet. The tracks are composed of large clusters, and the sizes of large clusters for argon and methane lie in the range of 7E + 03 ~ 1E + 07 and 2E + 06 ~ 4E + 08, respectively. The sizes of large argon clusters exhibit a weak dependence both on the axial positions and stagnation pressures, whereas the sizes of large methane clusters are irrelevant to the axial position but have a strong growing trend with the stagnation pressure. The methane clusters are very unstable and easily to fragment into small pieces in comparison with argon clusters. Special methane clusters at the axial position of 15 mm and radial position of 21 mm which may fly at the axial velocity of less than 2.5 m/s are also found. The method can be setup simply and operated easily, which is very suitable for laser-cluster interaction experiments and has broad application prospects in the cluster physics and nuclear fusion fueling field. Additional Information How to cite this article: Chen, Z. et al. The optical measurement of large cluster tracks in a gas jet. Sci. Rep. 6, 32391; doi: 10.1038/srep32391 (2016). This work is supported by the National Magnetic Confinement Fusion Program of China (2014GB125004) and the National Natural Science Foundation of China (11275133). Author Contributions Z.C., D.L. and J.H. conceived and conducted the experiment, Z.C., D.L., J.H. and L.B. analysed the results. All authors reviewed the manuscript. Figure 1 The schematic diagram of this experiment. Figure 2 An effective image (a) and the corresponding reference image (b) of argon clusters. Figure 3 An image of methane tracks (a) and the combined methane track of three consecutive images (b). Figure 4 The argon cluster tracks at different axial positions. Figure 5 The histogram of FWHM of argon (a) and methane (b). Figure 6 Typical Rayleigh scattering signal acquired by PMT and quasi-continuous laser. Figure 7 The Rayleigh scattering signal (a) and average cluster size (b) versus stagnation pressure for argon. Figure 8 The Rayleigh scattering signal (a) and average cluster size (b) versus stagnation pressure for methane. Figure 9 The large cluster size of argon versus the axial position (a) and stagnation pressure (b). Figure 10 The large cluster size of methane versus the axial position (a) and stagnation pressure (b). Figure 11 The number of argon tracks versus the axial position (a) and stagnation pressure (b). Figure 12 The number of methane tracks versus the axial position (a) and stagnation pressure (b). ==== Refs Krainov V. P. & Smirnov. M. B. Cluster beams in the super-intense femtosecond laser pulse . Phys. Rep 370 , 237 –331 (2002 ). Chatterjee A. . Remote access and display of neutron data . Appl. Phys. A 74 , 194 –197 (2002 ). Parra E. . X-ray and extreme ultraviolet emission induced by variable pulse-width irradiation of Ar and Kr clusters and droplets . Phys. Rev. E 62 , R5931 (2000 ). Ditmire T. . Interaction of intense laser pulses with atomic clusters . Phys. Rev. A 53 , 3379 (1996 ).9913282 Chen G. L. . Pure Coulomb explosions of highly charged methane clusters investigated by a simple electrostatic model . J. Phys. B: At. Mol. Opt. Phys 41 , 105601 (2008) . Schutte S. & Buck. U. Strong fragmentation of large rare gas clusters by high energy electron impact . Int. J. Mass Spectrom 220 , 183 –192 (2002 ). Kim K. Y. , Kumarappan V. & Milchberg. H. M. Measurement of the average size and density of clusters in a gas jet . Appl. Phys. Lett 83 , 3210 (2003 ). Saxena A. . Dependence of ion kinetic energy and charge on cluster size in multi-photon ionization of xenon clusters . Int. J. Mass Spec 357 , 58 –62 (2014 ). Carado A. . Cluster SIMS with a hybrid quadrupole time-of-flight mass spectrometer . Mass. Spec 255 , 1610 –1613 (2008 ). Torchet G. , de Feraudy M.-F. , Boutin A. & Fuchs A. H. . Structural transformation in (CO2) N clusters, N < 100 . J. Chem. Phys 105 , 3671 (1996 ). Fedor J. . Velocity map imaging of HBr photodissociation in large rara gas cluasters . J. Chem. Phys 134 , 154303 (2011 ).21513383 Hagena O. F. & Obert. W. Cluster Formation in Expanding Supersonic Jets: Effect of Pressure, Temperature, Nozzle Size, and Test Gas . J. Chem. Phys 56, 1793 (1972 ). Hagena. O. F. Condensation in Free Jets: Comparison of Rare Gases and Metals . Z. Phys. D: At., Mol. Clusters 4 , 291 –299 (1987 ). Hagena. O. F. Cluster ion sources (invited) . Rev. Sci. Instrum 63 , 2374 (1992 ). Hagena O. F. . Nucleation and growth of clusters in expanding nozzle flows . Surf. Sci 106 , 101 –116 (1981 ). Chen G. L. , Kim B. , Ahn B. & Kim D. E. . Experimental investigation on argon cluster sizes for conical nozzles with different opening angles . J. Appl. Phys 108 , 064329 (2010 ). Knuth E. L. Size correlations for condensation clusters produced in free-jet expansions . J. Appl. Phys 107 , 9125 (1997 ). Dorchies F. . Gasilov. Spatial distribution of cluster size and density in supersonic jets as targets for intense laser pulses . Phys. Rev. A 68 , 023201 (2003 ). Gupta K. C. . Determining the mean size and density of clusters formed in supersonic jets by Rayleigh scattering and Mach-Zehnder interferometer . J. Appl. Phys 118 , 114308 (2015 ). Murakami A. . Investigation of the clustering condition for various gases ejected from a fast solenoid valve for supersonic cluster beam injection . Plasma. Fusion. Res 5 , S1032 (2010 ). Li S. . Cooling-induced increase of methane cluster size investigated under a Coulomb explosion scheme . Chin. Phys. Lett 11 , S20201 (2013 ). Liu C. B. . Size estimation and time evolution of large size rare gas clusters by Rayleigh scattering techniques . Chin. Phys. Lett 19 , 659 (2002 ). Lei A. L. . Size determination of argon clusters from a Rayleigh scattering experiment . Chin. Phys. Lett 17 , 661 (2000 ). Bell A. J. . Mean cluster size by Rayleigh scattering . J. Phys. D 26 , 994 –996 (1993 ). DeArmond. F. M. , Suelzer J. & Mastersa. M. F. Rayleigh scattering from argon clusters in a planar expansion . J. Appl. Phys. 103 , 093509 (2008 ). Smith. R. A. Ditmire T. & Tisch J. W. G. . Characterization of a cryogenically cooled high-pressure gas jet for laser and cluster interaction experiments . Rev. Sci. Instrum 69 , 3798 (1998 ). Han J. F. . The spatial distribution of argon clusters in gas jet . Euro. Phys. J. D 56 , 347 –352 (2010 ). Wormer J. , Joppien M. & Moller. T. Mass determination of free van der Waals clusters from absorption and scattering measurements . Chem. Phys. Lett 182 , 632 –636 (1991 ). Han J. F. . An experimental observation of axial variation of average size of methane clusters in a gas jet . Chin. Phys. Lett 27 , 043601 (2010 ).

PMC005xxxxxx/PMC5000295.txt

==== Front Sci RepSci RepScientific Reports2045-2322Nature Publishing Group srep3242410.1038/srep32424ArticleFSTL1 as a Potential Mediator of Exercise-Induced Cardioprotection in Post-Myocardial Infarction Rats Xi Yue 1Gong Da-Wei 23Tian Zhenjun a11 Institute of Sports and Exercise Biology, Shaanxi Normal University, Xi’an, Shaanxi, 710119, P. R. China2 Division of Endocrinology, Diabetes and Nutrition, Department of Medicine, University of Maryland School of Medicine, Baltimore, MD 21201, USA3 VA Research Service, Geriatric Research, Education and Clinical Center, Baltimore Veterans Administration Medical Center, Baltimore, MD 21201, USA.a tianzj2013@hotmail.com26 08 2016 2016 6 3242414 03 2016 05 08 2016 Copyright © 2016, The Author(s)2016The Author(s)This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/Exercise training has been reported to ameliorate heart dysfunction in both humans and animals after myocardial infarction (MI), but the underlying mechanisms are poorly understood. Follistatin-like1 (FSTL1) is a cardioprotective factor against ischemic injury and is induced in cardiomyocytes and skeletal muscle in ischemic and hypoxic conditions. To test the hypothesis that FSTL1 may be a molecular link between exercise and improved heart function post MI, we subjected MI-rats, induced by left coronary artery ligation, to two modes of exercise: intermittent aerobic exercise (IAE) or mechanical vibration training (MVT), for four weeks and examined the relevance of FSTL1 to exercise-mediated cardiac effects. Exercise improved the functional performance, reduced fibrosis of MI-hearts and induced FSTL1 expression, the TGFβ-Smad2/3 signaling and angiogenesis in myocardium. In gastrocnemius, exercise increased the cross-sectional area of myocytes and FSTL1 expression. Importantly, exercise increased circulating FSTL1 levels, which were positively correlated with the skeletal muscle FSTL1 expression and negatively correlated with heart fibrosis. Overall, the IAE was more effective than that of MVT in cardioprotection. Finally, exogenous FSTL1 administration directly improved angiogenesis as well as functionality of post-MI hearts. Taken together, we have demonstrated that FSTL1 is a potential mediator of exercise-induced cardioprotection in post-MI rats. ==== Body Myocardial infarction (MI) is a leading cause of mortality and morbidity in the world123. Pathologically, MI results in immediate tissue damage due to myocardial ischemia, followed by biochemical changes triggered by reperfusion and pathological remodeling, leading to left ventricular (LV) heart failure and mortality45. However, despite greater understanding of the pathological processes of MI and the use of pharmacological interventions made in recent decades, post-MI mortality remains high; a 5-year survival rate is about 66.70%67. Therefore, novel interventional strategies to prevent ischemia/reperfusion (I/R) injury and pathological remodeling are called for to improve the post-MI survival rate. Besides pharmacological interventions, exercise has shown cardioprotective effects against I/R injury and facilitates post-MI recovery. But how exercise mediates this beneficial effect is not well understood. One possible explanation is that skeletal muscle secretes some heart-protective factors89 and MI results in muscle atrophy and decrease in secretion of those factors. Conversely, exercise would counteract the muscle atrophy10 and hence, improve post-MI recovery. Moreover, exercise may act directly on the myocardium111213 to improve the microenvironment of infarcted hearts. Recently, the significance of follistatin-like1 (FSTL1), an angiogenic factor14, in cardiovascular system has been increasingly recognized151617. Mice with cardiac-specific fstl1 knock-out (cFstl1-KO) develop cardiac hypertrophy and ventricular dysfunction in response to transverse aortic constriction (TAC)18. FSTL1 is reported to suppress cardiac hypertrophy caused by pressure overload18 and to improve endothelial cells (EC) and vascular remodeling in hypoxic-ischemic regions14. Intriguingly, FSTL1 is secreted from both skeletal muscle19 and myocardium1520 and the muscle-derived FSTL1 can function as an endocrine hormone to modulate vascular remodeling in response to wire-induced artery injury21. However, whether and how FSTL1 is regulated by exercise has not been studied. Different modes of exercise have been reported to affect post-MI recovery differently: Intermittent aerobic exercise (IAE) is effective in diminishing pathological myocardial transformation in the post-infarction failing rat heart22, increasing peak oxygen uptake23 and improving functional capacity and life quality in patients with chronic heart failure (CHF)24, whereas mechanical vibration training (MVT) accelerates the reperfusion of vessels25 and elevates circulating levels of angiogenic regulators such as VEGF and MMP-2/9 in humans26. This study aimed to address the questions of whether FSTL1 is involved in exercise-mediated protection of post-MI hearts and which exercise mode, IAE or MVT, is more effective in cardioprotection. We found that exercise stimulated FSTL1 expression in skeletal muscle and myocardium after acute MI, concurrently with enhanced TGFβ-Smad2/3 signaling, increased myocardium angiogenesis and improved heart functional performance. Significantly, IAE was more effective than MVT in cardioprotection after MI. Results Exercise mitigates heart dysfunction and reduces heart fibrosis in MI rats We studied the effect of exercise on post-MI hearts in four groups of animals: MI control (MI) by left coronary artery (LAD) ligation, MI with exercises (IAE vs MVT), and sham operation control (C). As expected, heart functional parameters of the left ventricular systolic pressure (LVSP) and the contractility index, absolute value of ±dP/dt(max), in the group of MI were greatly decreased whereas the left ventricular end-diastolic pressure (LVEDP) was increased, compared to the control group, indicating a heart dysfunction (Fig. 1a). Significantly, exercise fully or partially restored these indexes, and IAE appeared more effective in lowering LVEDP than MVT (Fig. 1a). Thus, exercise improved the functionalities of the post-MI heart. Masson staining indicated strong blue collagen staining in myocardium of the MI group (Fig. 1b). Both the IAE and MVT modes of exercise reduced the staining area, but the degree of reduction is greater in the IAE group (Fig. 1c). These results demonstrate that exercise mitigated dysfunction in the post-MI heart and reduced heart fibrosis, with IAE being more effective in these respects. Exercise induces myocardial FSTL1 expression and angiogenesis and activates TGFβ-Smad2/3 signaling Next, we investigated whether exercise regulates the expression of FSTL1 by immunohistochemistry (IHC). As shown in Fig. 2a, FSTL1 was detected in the cytoplasm of cardiomyocytes and appeared to be increased in the non-infarction area in the MI group. IAE further increased the staining intensity. Quantitatively, compared to the control group, MI induced FSTL1 expression by 1.96 fold, which was further induced by 4.04 fold by IAE but not by MVT (Fig. 2b). Further western analyses (Fig. 2c) of FSTL1 confirmed the IHC findings. FSTL1 is a known angiogenic factor1427 and induction of FSTL1 would promote angiogenesis. Hence, we examined endothelial cell proliferation by co-staining of PCNA+, a cell proliferation marker and vWF+, an endothelial cell marker. Compared with the control group, MI induced more PCNA+/vWF+ cells, whose number was increased and distribution expanded by exercise (Fig 3a). Notably in the IAE group, some double stained cells appeared to form new small vessel-like structures whereas in the MVT group, the double-stained cells were more diffuse (Fig. 3a). Further co-staining of FSTL1 with CD31, an endothelial cell marker, revealed a significant increase in the number of CD31+/FSTL1+ vessels (8.50 ± 1.12 counts) in the MI group, compared to the control (Fig. 3b,c). The number of double stained vessels was further increased in the IAE (28.67 ± 2.06 counts) and MVT groups (20.50 ± 1.69 counts). We speculated that FSTL1 might act like FST through the TGFβ-Smad2/3 signaling pathway28 to induce angiogenesis29 and investigated the effect of exercise on the pathway. As a result, TGFβ1 protein expression was found to be up-regulated after MI, and further induced by IAE (Fig. 3d), but MVT had no effect on the protein expression. Downstreamly, Smad2/3 levels were significantly increased in MI and further up-regulated in IAE and MVT, but the former was more effective than MVT (Fig. 3e). Together, these results show that exercise, especially IAE, induced FSTL1 expression in myocardium of the infarct heart, promoted angiogenesis and enhanced angiogenesis-related signaling. Exercise reverses the skeletal muscle atrophy and increases the FSTL1 expression in skeletal muscle and serum after MI Skeletal muscle is a main source of the circulating FSTL11930, and we next investigated the effect of exercise on skeletal muscle mass and FSTL1 expression. We labeled the cell membrane by using derivatives of indocarbocyanine iodide (DiI) to measure changes of the gastrocnemius cell cross-sectional area (CSA) as a representative parameter for skeletal muscle mass3132. In consistent with previous reports3334, compared with the control group (481.73 ± 18.60 μm2/cell), CSA was significantly decreased after MI (340.50 ± 13.87 μm2/cell, p < 0.01), indicating myoatrophy. After 4 weeks of exercise, CSA increased by 2.23 fold and 1.81 fold in IAE and MVT groups (both p < 0.01 vs. the MI, Fig. 4a,b), respectively. IHC showed that FSTL1+ myocytes were sporadically detected in the control group, and were sparse in MI (p < 0.01). IAE and MVT significantly increased the number and intensity of the FSTL1+ cells (Fig. 4c,d). Western quantification showed that IAE and MVT significantly increased the FSTL1 expression by 36% and 10% respectively vs. the MI group (Fig. 4e). Furthermore, we measured circulating FSTL1 levels. Compared to the control group, circulating FSTL1 levels were increased post-MI (Control 6.22 ± 0.11 vs. MI 7.09 ± 0.22 ng/ml; p < 0.05), which were further increased by IAE (7.94 ± 0.33 ng/ml), but not by MVT (7.35 ± 0.24 ng/ml, Fig. 4f). Correlation analyses revealed significant positive correlations between the gastrocnemius CSA and the skeletal muscle FSTL1 expression (r = 0.83, p < 0.01; Fig. 5a), and between the skeletal muscle FSTL1 expression and the serum FSTL1 concentration (r = 0.76, p < 0.01; Fig. 5b), with a negative correlation between the serum FSTL1 and the heart collagen volume fraction (CVF) (r = −0.69, Fig. 5c). Collectively, these findings show a significant increase of gastrocnemius CSA, skeletal muscle FSTL1 expression and serum FSTL1 levels, and a remarkable correlation of the expression of skeletal muscle FSTL1 with the level of serum FSTL1. Exogenous FSTL1 improves angiogenesis, enhances TGFβ-Smad2/3 signaling and heart function after MI To determine whether FSTL1 exerts a direct effect on post-MI hearts, we administrated FSTL1 at the dosage of 100 μg/kg body weight/day17 from 1wk to 5wk post-MI and serum FSTL1 was measured at several time points as illustrated in Fig. 6a. Compared to the pre-MI level, serum FSTL1 decreased by 45.57% on day 1 after MI and remained lower during the acute phase of MI (1 day-1week) and gradually increased at week 2 and week 5. Administration of FSTL1 from week 1 resulted in a small increase at week 2 and a statistically significant increase at week 5, compared to the control group. Serum FSTL1 levels were elevated in the FSTL1 administration group (Fig. 6b). IHC of myocardium at week 5 revealed more vWF+ cells and small vessels in the FSTL1 group than in the PBS control group (Fig. 6c,d). In addition, at the protein level, exogenous FSTL1 administration significantly up-regulated heart FSTL1 and TGFβ1 protein expressions at week 5 by 95.81% and 54.14% vs. the PBS control, respectively. Interestingly, Smad2/3 and p-Smad2/3 expressions were also increased in FSTL1 group, while the p-Smad2/3 vs. Smad2/3 ratio had risen about 76.70% (p < 0.01 vs. the control group, Fig. 6e). Additionally, exogenous FSTL1 administration enhanced signaling pathways of Akt, Erk1/2 and AMPK but not of Smad1/5/8 in MI hearts (Fig. S2). Importantly, FSTL1-treated animals showed a significant improvement of heart function with an increase of LVSP by 31.71%, +dP/dt(max) by 33.17%, −dP/dt(max) by 41.97% and a decrease of LVEDP by 64.49% at week 5 (p < 0.01, Fig. 6f). These results demonstrate that FSTL1 exerts direct cardioprotective action by promoting angiogenesis, increasing FSTL1 protein content and activating TGFβ-Smad2/3 signaling in the post-MI heart, all of which mitigate heart dysfunction. Discussion In this study, we have replicated exercise’s cardioprotective effect in post-MI hearts3536 and demonstrated that exercise can significantly induce skeletal muscle and cardiac expression of FSTL1 and increase its circulation levels. Importantly, the level of skeletal muscle FSTL1 expression is significantly correlated with its circulating levels and associated with heart functional performance, providing a piece of evidence that exercise may exert its cardioprotection through increased production and secretion of FSTL1 in the skeletal muscle and myocardium. Which tissue of FSTL1 expression (skeletal muscle or myocardium) is crucial to the exercise-mediated cardiac benefits is an intriguing question but difficult to answer for the time being. Cardiac FSTL1 expression is increased in response to heart ischemic injury1720. Mice with cardiomyocyte-specific deletion of FSTL1 are susceptible to, whereas mice with FSTL1 overexpression are resistant to, heart dysfunction following transverse aortic constriction (TAC)18. Thus, myocardial FSTL1 appears to be cardioprotective again heart injury. Recently, Wei et al. has reported that epicardial FSTL1 can promote immature cardiac myocyte proliferation and diminishes infarct size post-MI15, and that myocardial FSTL1 is insufficient for long-term recovery from MI. Nevertheless, this study does not negate a possible significance of endogenous myocardial FSTL1 in cardioprotection. Angiogenesis is an essential part of heart repair post-MI; it helps to salvage ischemic myocardium at the early stages and is also essential for long-term left ventricular remodeling to prevent the transition to heart failure3738. Angiogenesis by TGFβ is known to be partly mediated by Smad2/3 activation3940. We found that the increased myocardial FSTL1 expression was associated with enhanced TGFβ-Smad2/3 signaling and angiogenesis, suggesting that FSTL1 may exert its cardiac benefits through angiogenesis. FSTL1 can activate and increase the expression of TGFβ41. In the present study, we observed that TGFβ, total Smad2/3 and p-Smad2/3 protein levels were increased by exogenous FSTL1 administration. This finding is consistent with prior studies4243 where co-regulations among TGFβ and total Smad2/3 protein expression p-Smad2/3 level are reported, suggesting a possible autoregulation mechanism42. Interestingly, Akt, Erk1/2 and AMPK signaling pathways were also enhanced in myocardium of rats treated with FSTL1 after MI, and Smad1/5/8 was significantly activated after exercise (Figs S1 and S2). These results suggested that multiple signaling mechanisms may participate in the FSTL1-mediated cardioprotection in MI, which deserves further elucidation. The role of skeletal muscle in cardioprotection in MI animal models has been previously recognized. For example, a brief pre-MI ischemic insult of skeletal muscle is reported to decrease infarct size434445 and chronic ischemia of skeletal muscle can increase left ventricle coronary vessel density46. These results suggest that ischemic muscle may exert its cardioprotection through some neurohumoral mechanism47. However, such a factor(s) has not been fully defined. Since exercise has muscle ischemia-like effects through hypoxia48, it may exert cardiac protective action through a similar mechanism to ischemia. Exercise is a strong stimulator of muscle growth, metabolism and endocrine function49. Thus, both the increased muscle mass and ischemia-mimic effect through exercise may contribute to the increased circulating levels of FSTL1. The functional significance of circulating FSTL1 was demonstrated by our animal study where exogenous FSTL1 administration activated the TGFβ-Smad2/3 signaling, promoted angiogenesis and ameliorated heart dysfunctions (Fig. 6). FSTL1 has been increasingly recognized as a potent cardiac protection factor151617. Whether exercise increases the heart FSTL1 expression through a peripheral, neuronal or cardiac mechanism will be a subject of future study. Interestingly, we found that exogenous FSTL1 administration increased FSTL1 expression (Fig. 6), suggesting that there might be a positive feedback regulation of circulating FSTL1 on myocardial FSTL1, probably through modulating the local microenvironment, e.g. angiogenesis. In this study, we trained post-MI rats with IAE and MVT out of the consideration that the exercise mode and intensity are reported to affect the endocrine function of skeletal muscle and post-MI cardiac outcome differently505152. For instance, twice a week IAE for 12 weeks at the intensity of 4 × 4 minute intervals at 85–95% of peak heart rate can diminish myocardial damage through increasing peak oxygen uptake23 and improving heart functional capacity and quality of life. LVEDP, left ventricular mass/body mass ratio (LVM:BM) and total CVF were decreased while LVSP and +dP/dt(max) were increased by aerobic interval training (40 min/day with 8 min of warm-up at 10 m/min and exercise at 15 m/min 4 × 4 min interspersed with 4 × 4 min at 23 m/min) in rats with chronic heart failure53. These results are consistent with our findings. Mechanical vibration mode by using the vibration platform with a peak-to-peak amplitude of 4 mm and a frequency of 30 Hz has been reported to improve microvessel circulation2526. But no comparative studies of IAE vs. MVT have been reported previously. Our study shows that IAE has more protective benefits for post-MI hearts than MVT, resulting in more myocardial angiogenesis, less fibrosis and improved heart functionality. These results may be instructive for rehabilitation of post-MI patients. Importantly, IAE increases more skeletal muscle and myocardial expression of FSTL1, providing another piece of evidence that FSTL1 may be a mediator of exercise’s cardiac benefits. In conclusion, we have demonstrated that exercise may exert its beneficial effect on post-MI hearts through the induction of cardiac and muscle FSTL1 expression. Importantly, exercise-regulated skeletal muscle FSTL1 expression is highly correlated with serum levels as well as increased heart functional performance and decreased heart fibrosis. Moreover, FSTL1’s signaling pathway is activated during exercise and exogenous FSTL1 administration. Thus, the induction of FSTL1 provides an explanation of exercise-mediated cardioprotection in post-MI hearts. Future studies using muscle- or heart-specific FSTL1 over-expression or knockout models will help to address the question of the relative significance of FSTL1 in cardioprotection between the two tissues by exercise. Materials and Methods Animals Male Sprague-Dawley rats(200 ± 20 g, 8-weeks old)were from the Laboratory Animal Centre of Xi’an Jiaotong University. Animal studies were performed in accordance with the “Guiding principles for research involving animals and human beings”54. All experimental protocols were approved by the Review Committee for the Use of Human or Animal Subjects of Shaanxi Normal University. MI Surgical Procedure Rats were anesthetized by pentobarbital sodium (30 mg/kg body weight). MI was induced by using the established method of left anterior descending (LAD) coronary artery ligation55. In brief, the coronary artery was ligated 2.0 mm from its origin using a 6.0 suture silk. An ST-segment elevation was observed by electrocardiogram after suture. Sham-operation control rats underwent the operation procedure without LAD ligation. Exercise Protocol One week after infarction, rats underwent one week of adaptive training before four weeks of normal exercise. The intermittent aerobic exercise (IAE) group was subjected to a training protocol using a motorized rodent treadmill (DSPT-202, Li Tai Technology, Hangzhou, China). For daily training, animals would first start with a warm-up for 10 min and then the exercise alternated between 7 min at 25 meter/min (85–90% VO2max) and 3 min at 15 meter/min (50–60% VO2max) for 1 hr. This protocol was performed once a day, 5 days a week for 4 weeks56. The mechanical vibration training (MVT) group was subjected to a custom-made small animal mechanical vibration platform. Rats were vibrated in the shaker at a frequency of 25 Hz with amplitude of 2 mm for everyday training. This protocol was performed for twice a day, 38 mins per day and 5 days per week for 4 weeks57. No rats died by the end of these two protocols. Hemodynamic Measurement At the end of the 4 weeks of training or ad-lib activity, rats were anesthetized with pentobarbital sodium (30 mg/kg body weight). A pressure transducer was inserted retrograde from the right carotid artery to the LV cavity, and intraventricular catheter recordings were performed by using Powerlab 8/30 (ML 870, AD Instruments, Castle Hill, Australia) to evaluate cardiac function. LV systolic pressure (LVSP, mmHg), LV end-diastolic pressure (LVEDP, mmHg), heart rate, and maximal positive and negative first derivative of LV pressure (±dP/dt max) were measured and calculated. Tissue Histology Heart and gastrocnemius tissue samples were fixed in ice-cold 4% paraformaldehyde for 24–48 hrs, dehydrated in a concentration gradient of ethanol, embedded in paraffin and sectioned (5 μm) for histopathologic examination. To evaluate the degree of skeletal muscle atrophy, the gastrocnemius slices were stained with DiI (Sigma-Aldrich, St. Louis, MO, USA), and the cross-sectional area of skeletal muscle cells was calculated. To evaluate the degree of myocardium fibrosis, heart tissue slices were stained with Masson’s trichrome, and the collagen volume fraction (CVF) was measured. Western Blotting Tissues from the LV infarct border area (5 mm) and gastrocnemius were homogenized. Total proteins extraction and SDS-PAGE were performed as described before58. The dilution of primary antibodies as follows: FSTL1 (1:1000, GeneTex, Irvine, CA, USA), TGFβ1 (1:500, Bioworld Technology, St. Louis, MN, USA), Smad2/3 (1:1000, Bioworld), p-Smad2/3 (1:500, Bioworld), Akt (1:2000, Cell Signaling, Danvers, MA, USA), p-Akt (1:2000, Cell Signaling), Erk1/2 (1:500, Signalway Antibody, College Park, MD, USA), p-Erk1/2 (1:500, Signalway Antibody), AMPKα1 (1:500, Signalway Antibody), p-AMPKα1 (1:500, Signalway Antibody), Smad1/5/8(1:500, Signalway Antibody), p-Smad1/5/8(1:1000, Cell Signaling) and GAPDH (1:10000, Bioworld). Following incubating with horseradish peroxidase (HRP)-conjugated secondary antibody (1:5000 dilution, Jackson ImmunoResearch, West Grove, PA, USA), protein bands were subsequently developed with enhanced chemiluminescence. Western quantification was performed by Image Processing and Analysis in Java 1.48 (Wayne Rasband, National Institutes of Health, USA). Immunohistochemistry Heart sections were incubated with rabbit polyclonal antibody vWF (1:50 dilution, Millipore) over night at 4 °C after deparaffination, 3% H2O2 treatment and antigen retrieval. HRP-conjugated goat anti-rabbit IgG (R&D, Minneapolis, MN 55413) was used as a secondary antibody and diaminobenzidine (R&D) for color development. Nuclei were stained by hematoxylin. For immunofluorescent staining, sections were incubated with primary antibody over night at 4 °C after deparaffination and antigen retrieval. The dilutions of primary antibodies were as follows: rabbit polyclonal antibody FSTL1 (1:100 dilution, GeneTex), rabbit polyclonal antibody vWF (1:50 dilution, Millipore), mouse monoclonal antibody CD31 (1:25 dilution, GeneTex) and mouse monoclonal antibody PCNA (1:50 dilution, Cell Signaling). TRITC/FITC-conjugated goat anti-rabbit/mouse IgG (1:100 dilution, Jackson ImmunoResearch) were used as secondary antibodies. Nuclei were stained by 4′6-diamidino-2-phenylindole (DAPI) dye (1:800 dilution, Sigma, St. Louis, MO, USA). Results were observed with a fluorescence microscope (Nikon Eclipse 55i, Nikon, Tokyo, Japan). FSTL1 Administration and ELISA One week post-operation, recombinant FSTL1 protein (ProSpec, East Brunswick, NJ, USA) was administrated by intraperitoneal injection at the dosage of 100 μg/kg bodyweight/day17 for 4 weeks. Blood was taken as indicated in Fig. 6a. ELISA was performed according to the manufacturer’s instruction (BD Biosciences). Statistical Analysis Density mean of random views were calculated by Image-Pro Plus 6.0 (Media Cybernetics, Bethesda, MD, USA). All data were presented as mean ± SE in this study. Statistical analysis was performed using SPSS 17.0. One-way analysis of variance (ANOVA) was used for group comparison and the Pearson correlation method was used to assess possible pairwise relationships. p < 0.05 was considered significant. Histograms were plotted by Prism 5.01 (GraphPad Software, La Jolla, CA, USA). Additional Information How to cite this article: Xi, Y. et al. FSTL1 as a Potential Mediator of Exercise-Induced Cardioprotection in Post-Myocardial Infarction Rats. Sci. Rep. 6, 32424; doi: 10.1038/srep32424 (2016). Supplementary Material Supplementary Information This work was supported by National Natural Science Foundation of China (Grant No. 31371199, 31171141) and the grants from “211 Project” of Shaanxi Normal University (2014-1). Author Contributions Y.X. performed the experiments. Y.X. and Z.T. prepared the figures. Y.X., D.-W.G. and Z.T. participated in design of the experiments. All authors contributed to the writing of the manuscript. Figure 1 Exercise improves heart function and reduces myocardial fibrosis. (a) Rats were subjected to sham operation (C) or myocardial infarction (MI). The MI rats were trained without or with intermittent aerobic exercise (IAE) or mechanical vibration training (MVT) for 4 weeks when heart hemodynamics were measured. LVSP, left ventricular systolic pressure; LVEDP, left ventricular end-diastolic pressure; ±dP/dt, maximum pressure increasing/decreasing rate. (b) Masson staining of heart tissue section for fibrosis evaluation. Collagen is stained blue, indicating fibrosis. The lower image is an amplification of the upper black box. (c) Quantification of collagen volume of fraction (CVF). Data are mean + SE (n = 6); *p < 0.05; **p < 0.01. Figure 2 Effect of exercise on myocardium expression of FSTL1. (a) Immunohistochemistry (IHC) of FSTL1 in cardiomyocytes. IHC staining (upper row) of FSTL1 in groups of control (C), MI, MI with IAE and MI with MVT, was counterstained with DAPI (blue) for nuclei (second raw). The third row is a merged image of FSTL1 with DAPI, indicating that FSTL1 (representatively indicated by small arrows in a small square) is expressed in the non-infarction boarder area of myocardium. The fourth row is the amplification of the small square area. (b) Quantification of the immunofluorescent staining. (c) Western blot analysis of FSTL1 and quantification (lower) with GAPGH as the loading control. Data are mean + SE (n = 6); *p < 0.05; **p < 0.01. Figure 3 Exercise stimulates myocardium angiogenesis, induces CD31+/FSTL1+ cell proliferation and activates TGFβ-Smad2/3 signaling. (a) Double staining of PCNA and vWF markers, with DAPI nuclei staining in myocardium of the sham-op control (C), MI, MI with IAE and MI with MVT. Arrows indicate the PCNA+/vWF+ cells and the farthest right image is an amplification of the white box within the merged image. More vessel-like structures are seen in IAE and MVT. (b) Co-staining of CD31+/FSTL1+ shows that endothelial cells form capillary-like structure indicated by arrows in IAE and MVT. Inserts are amplification of white boxes. (c) Quantification of the number of CD31+/FSTL1+ capillaries per slice. (d) Western analysis and quantification of TGFβ1. (e) Western blot analysis and quantification of Smad2/3 and phosphorylated Smad2/3 (p-Smad). Data are mean + SE (n = 6); *p < 0.05; **p < 0.01. Figure 4 Exercise reverses MI-induced myoatrophy and increases muscle FSTL1 expression and blood FSTL1 levels. (a) Staining of gastrocnemius myocyte membrane by DiI in sham-op control (C), MI and MI with exercise (IAE or MVT) for 4 weeks, with quantification (b) of cell cross-sectional area (CSA) of the average of 100 cells. Immunofluorescent staining (c) and quantification (d) and Western analysis (e) of FSTL1 in gastrocnemius muscle. Stronger staining is observed in IAE and MVT groups. f: Serum levels of FSTL1 in the groups of rats. Data are mean + SE (n = 6); *p < 0.05; **p < 0.01. Figure 5 Correlation of serum FSTL1 levels with muscle FSTL1 expression and with heart fibrosis. Positive correlations of skeletal muscle FSTL1 expression with gastrocnemius myocytes cross-section area (CSA) (a) and with serum FSTL1 levels (b). (c) Negative correlation of serum FSTL1 with heart collagen volume of fraction (CVF). Pearson correlations were conducted for the above analyses (n = 15, pooled samples of C, MI, IAE and MVT). Figure 6 Exogenous FSTL1 directly stimulates heart angiogenesis, activates TGFβ-Smad2/3 signaling and improves functional performance. (a) Scheme of the animal study. Normal rats underwent MI by LAD ligation on Day 0. Blood collection (BC) took places at the indicated time points for FSTL1 measurements. Recombinant FSTL1 was administered one week post-MI and continued daily for 4 weeks. Immunohistochemistry and functional assessment were conducted at the end of week 5. (b) Changes of blood FSTL1 levels of rats treated with or without FSTL1 during the animal protocol. (c,d) Immunohistochemistry staining shows more vWF+ vascular structure (yellow arrow) and vWF+ cells (red arrow) in rats treated with FSTL1 than the PBS control, and quantification of mean optical density per view. (e) Enhanced TGFβ1/p-Smad2/3 signaling in myocardium of rats treated with FSTL1 vs. PBS control, as analyzed by Western blot and quantification. (f) Functional assessment of hemodynamic parameters in rats treated with PBS or FSTL1. LVSP: left ventricular systolic pressure; LVEDP: left ventricular end-diastolic pressure.; and ±dP/dt: maximum pressure increasing/decreasing rate. Data are expressed as mean + SE (n = 6); *p < 0.05; **p < 0.01. ==== Refs Shroff G. R. , Li S. & Herzog C. A. Trends in Mortality Following Acute Myocardial Infarction Among Dialysis Patients in the United States Over 15 Years . J Am Heart Assoc 4 , e002460 (2015 ).26459933 Plakht Y. , Shiyovich A. & Gilutz H. Predictors of long-term (10-year) mortality postmyocardial infarction: age-related differences. Soroka Acute Myocardial Infarction (SAMI) Project . J Cardiol 65 , 216 –223 (2015 ).24994020 Aguero F. . New myocardial infarction definition affects incidence, mortality, hospitalization rates and prognosis . Eur J Prev Cardiol 22 , 1272 –1280 (2015 ).25139771 Wu Y. , Yin X. , Wijaya C. , Huang M. H. & McConnell B. K. Acute myocardial infarction in rats . J Vis Expp 48 (2011 ). Lujan H. L. & DiCarlo S. E. Mimicking the endogenous current of injury improves post-infarct cardiac remodeling . Med Hypotheses 81 , 521 –523 (2013 ).23871584 Icks A. . Mortality after first myocardial infarction in diabetic and non-diabetic people between 1985 and 2009. The MONICA/KORA registry . Eur J Epidemiol 29 , 899 –909 (2014 ).25366554 Hosseini S. H. , Ghaemian A. , Mehdizadeh E. & Ashraf H. Levels of anxiety and depression as predictors of mortality following myocardial infarction: A 5-year follow-up . Cardiol J 21 , 370 –377 (2014 ).24671898 Martinez P. F. . Chronic heart failure-induced skeletal muscle atrophy, necrosis, and changes in myogenic regulatory factors . Med Sci Monit 16 , BR374 –BR383 (2010 ).21119570 Forman D. E. . Analysis of skeletal muscle gene expression patterns and the impact of functional capacity in patients with systolic heart failure . J Card Fail 20 , 422 –430 (2014 ).24704539 Pedersen B. K. & Febbraio M. A. Muscles, exercise and obesity: skeletal muscle as a secretory organ . Nat Rev Endocrinol 8 , 457 –465 (2012 ).22473333 Lu K. . Effects of high-intensity interval versus continuous moderate-intensity aerobic exercise on apoptosis, oxidative stress and metabolism of the infarcted myocardium in a rat model . Mol Med Rep 12 , 2374 –2382 (2015 ).25936391 Pieri B. L. . Effects of physical exercise on the P38MAPK/REDD1/14-3-3 pathways in the myocardium of diet-induced obesity rats . Horm Metab Res 46 , 621 –627 (2014 ).24691733 Muthusamy V. R. . Acute exercise stress activates Nrf2/ARE signaling and promotes antioxidant mechanisms in the myocardium . Free Radic Biol Med 52 , 366 –376 (2012 ).22051043 Ouchi N. . Follistatin-like 1, a secreted muscle protein, promotes endothelial cell function and revascularization in ischemic tissue through a nitric-oxide synthase-dependent mechanism . J Biol Chem 283 , 32802 –32811 (2008 ).18718903 Wei K. . Epicardial FSTL1 reconstitution regenerates the adult mammalian heart . Nature 525 , 479 –485 (2015 ).26375005 Altekoester A. K. & Harvey R. P. Bioengineered FSTL1 Patches Restore Cardiac Function Following Myocardial Infarction . Trends Mol Med 21 , 731 –733 (2015 ).26596868 Ogura Y. . Therapeutic impact of follistatin-like 1 on myocardial ischemic injury in preclinical models . Circulation 126 , 1728 –1738 (2012 ).22929303 Shimano M. . Cardiac myocyte follistatin-like 1 functions to attenuate hypertrophy following pressure overload . Proc Natl Acad Sci USA 108 , E899 –E906 (2011 ).21987816 Gorgens S. W. . Regulation of follistatin-like protein 1 expression and secretion in primary human skeletal muscle cells . Arch Physiol Biochem 119 , 75 –80 (2013 ).23419164 Oshima Y. . Follistatin-like 1 is an Akt-regulated cardioprotective factor that is secreted by the heart . Circulation 117 , 3099 –3108 (2008 ).18519848 Miyabe M. . Muscle-derived follistatin-like 1 functions to reduce neointimal formation after vascular injury . Cardiovasc Res 103 , 111 –120 (2014 ).24743592 Kraljevic J. . Aerobic interval training attenuates remodelling and mitochondrial dysfunction in the post-infarction failing rat heart . Cardiovasc Res 99 , 55 –64 (2013 ).23554460 Moholdt T. . Aerobic interval training increases peak oxygen uptake more than usual care exercise training in myocardial infarction patients: a randomized controlled study . Clin Rehabil 26 , 33 –44 (2012 ).21937520 Nilsson B. B. , Westheim A. & Risberg M. A. Effects of group-based high-intensity aerobic interval training in patients with chronic heart failure . Am J Cardiol 102 , 1361 –1365 (2008 ).18993156 Marzencki M. . Low frequency mechanical actuation accelerates reperfusion in-vitro . Biomed Eng Online 12 , 121 (2013 ).24257116 Suhr F. . Effects of short-term vibration and hypoxia during high-intensity cycling exercise on circulating levels of angiogenic regulators in humans . J Appl Physiol (1985) 103 , 474 –483 (2007 ).17446405 Javerzat S. . Correlating global gene regulation to angiogenesis in the developing chick extra-embryonic vascular system . PloS one 4 , e7856 (2009 ).19924294 Wu J. . Follistatin-like 1 attenuates differentiation and survival of erythroid cells through Smad2/3 signaling . Biochem Biophys Res Commun 466 , 711 –716 (2015 ).26365350 Munjal C. . TGF-beta mediates early angiogenesis and latent fibrosis in an Emilin1-deficient mouse model of aortic valve disease . Dis Model Mech 7 , 987 –996 (2014 ).25056700 Stranska Z. & Svacina S. [Myokines - muscle tissue hormones] . Vnitr Lek 61 , 365 –368 (2015 ).25894270 Trombetti A. . Age-associated declines in muscle mass, strength, power, and physical performance: impact on fear of falling and quality of life . Osteoporos Int 27 , 463 –471 (2015 ).26194491 Sung B. . Loquat leaf extract enhances myogenic differentiation, improves muscle function and attenuates muscle loss in aged rats . Int J Mol Med 36 , 792 –800 (2015 ).26178971 Bechara L. R. . NADPH oxidase hyperactivity induces plantaris atrophy in heart failure rats . Int J Cardiol 175 , 499 –507 (2014 ).25023789 Lima A. R. . Myostatin and follistatin expression in skeletal muscles of rats with chronic heart failure . Int J Exp Pathol 91 , 54 –62 (2010 ).20002838 Rodrigues F. . Cardioprotection afforded by exercise training prior to myocardial infarction is associated with autonomic function improvement . BMC Cardiovasc Disord 14 , 84 (2014 ).25022361 Barbosa V. A. . Exercise training plays cardioprotection through the oxidative stress reduction in obese rats submitted to myocardial infarction . Int J Cardiol 157 , 422 –424 (2012 ).22498422 Shah A. M. & Mann D. L. In search of new therapeutic targets and strategies for heart failure: recent advances in basic science . Lancet 378 , 704 –712 (2011 ).21856484 van der Laan A. M. , Piek J. J. & van Royen N. Targeting angiogenesis to restore the microcirculation after reperfused MI . Nat Rev Cardiol 6 , 515 –523 (2009 ).19528962 Serrati S. . TGFbeta1 antagonistic peptides inhibit TGFbeta1-dependent angiogenesis . Biochem Pharmacol 77 , 813 –825 (2009 ).19041849 del Valle-Perez B. , Martinez-Estrada O. M. , Vilaro S. , Ventura F. & Vinals F. cAMP inhibits TGFbeta1-induced in vitro angiogenesis . FEBS letters 569 , 105 –111 (2004 ).15225617 Dong Y. . Blocking follistatin-like 1 attenuates bleomycin-induced pulmonary fibrosis in mice . J Exp Med 212 , 235 –252 (2015 ).25584011 Xie J. L. . Expression of Smad protein by normal skin fibroblasts and hypertrophic scar fibroblasts in response to transforming growth factor beta1 . Dermatol Surg 34 , 1216 –1225 (2008 ).18616749 Konstantinov I. E. . The remote ischemic preconditioning stimulus modifies gene expression in mouse myocardium . J Thorac Cardiovasc Surg 130 , 1326 –1332 (2005 ).16256785 Konstantinov I. E. . Remote ischemic preconditioning of the recipient reduces myocardial ischemia-reperfusion injury of the denervated donor heart via a Katp channel-dependent mechanism . Transplantation 79 , 1691 –1695 (2005 ).15973170 Kharbanda R. K. . Transient limb ischemia induces remote ischemic preconditioning in vivo . Circulation 106 , 2881 –2883 (2002 ).12460865 Varnavas V. C. . Chronic skeletal muscle ischemia preserves coronary flow in the ischemic rat heart . Am J Physiol Heart Circ Physiol 301 , H1229 –H1235 (2011 ).21765054 Loncar G. , Fulster S. , von Haehling S. & Popovic V. Metabolism and the heart: an overview of muscle, fat, and bone metabolism in heart failure . Int J Cardiol 162 , 77 –85 (2013 ).21982619 Zhang Y. . Dual AAV therapy ameliorates exercise-induced muscle injury and functional ischemia in murine models of Duchenne muscular dystrophy . Hum Mol Genet 22 , 3720 –3729 (2013 ).23681067 Gonzalez A. M. , Hoffman J. R. , Stout J. R. , Fukuda D. H. & Willoughby D. S. Intramuscular Anabolic Signaling and Endocrine Response Following Resistance Exercise: Implications for Muscle Hypertrophy . Sports Med 46 , 671 –685 (2015 ).26666743 Schulze P. C. , Gielen S. , Schuler G. & Hambrecht R. Chronic heart failure and skeletal muscle catabolism: effects of exercise training . Int J Cardiol 85 , 141 –149 (2002 ).12163219 Moreira J. B. . High- versus moderate-intensity aerobic exercise training effects on skeletal muscle of infarcted rats . J Appl Physiol (1985) 114 , 1029 –1041 (2013 ).23429866 Kraemer R. R. & Castracane V. D. Endocrine alterations from concentric vs. eccentric muscle actions: a brief review . Metabolism 64 , 190 –201 (2015 ).25467839 Nunes R. B. . Interval and continuous exercise enhances aerobic capacity and hemodynamic function in CHF rats . Braz J Phys s Ther 19 , 257 –263 (2015 ). Guiding principles for research involving animals and human beings. Am J Physiol Regul Integr Comp Physiol 283 , R281 –R283 (2002 ).12121837 Muthuramu I. , Lox M. , Jacobs F. & De Geest B. Permanent ligation of the left anterior descending coronary artery in mice: a model of post-myocardial infarction remodelling and heart failure . J Vis Exp 94 (2014 ). Wisløff U. , Loennechen J. P. , Currie S. , Smith G. L. & Ellingsen O. Aerobic exercise reduces cardiomyocyte hypertrophy and increases contractility, Ca2+ sensitivity and SERCA-2 in rat after myocardial infarction . Cardiovasc Res 54 , 162 –174 (2002 ).12062372 Oxlund B. S. , Ortoft G. , Andreassen T. T. & Oxlund H. Low-intensity, high-frequency vibration appears to prevent the decrease in strength of the femur and tibia associated with ovariectomy of adult rats . Bone 32 , 69 –77 (2003 ).12584038 Chen T. . Aerobic exercise inhibits sympathetic nerve sprouting and restores beta-adrenergic receptor balance in rats with myocardial infarction . PLoS One 9 , e97810 (2014 ).24842290

PMC005xxxxxx/PMC5000296.txt

==== Front Sci RepSci RepScientific Reports2045-2322Nature Publishing Group srep3245910.1038/srep32459ArticleDeep sexual dimorphism in adult medaka fish liver highlighted by multi-omic approach Qiao Qin 1Le Manach Séverine 1Sotton Benoit 1Huet Hélène 12Duvernois-Berthet Evelyne 3Paris Alain 1Duval Charlotte 1Ponger Loïc 4Marie Arul 1Blond Alain 1Mathéron Lucrèce 5Vinh Joelle 6Bolbach Gérard 5Djediat Chakib 1Bernard Cécile 1Edery Marc 1Marie Benjamin a11 UMR 7245 MNHN/CNRS Molécules de Communication et Adaptation des Micro-organismes, Sorbonne Universités, Muséum National d’Histoire Naturelle, Paris, France2 Université Paris-Est, Ecole Nationale Vétérinaire d’Alfort, BioPôle Alfort, Maisons-Alfort, France3 UMR 7221 CNRS/MNHN, Évolution des Régulations Endocriniennes, Sorbonne Universités, Muséum Nationale d’Histoire Naturelle, Paris, France4 UMR 7196 MNHN/CNRS, INSERM U1154, Sorbonne Universités, Museum National d’Histoire Naturelle, Paris, France5 Institut de Biologie Paris Seine/FR 3631, Plateforme Spectrométrie de masse et Protéomique, Institut de Biologie Intégrative IFR 83, Sorbonne Universités, Université Pierre et Marie Curie, Paris, France6 USR 3149 ESPCI/CNRS SMPB, Laboratory of Biological Mass Spectrometry and Proteomics, ESPCI Paris, PSL Research University, Paris, Francea bmarie@mnhn.fr26 08 2016 2016 6 3245913 06 2016 09 08 2016 Copyright © 2016, The Author(s)2016The Author(s)This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/Sexual dimorphism describes the features that discriminate between the two sexes at various biological levels. Especially, during the reproductive phase, the liver is one of the most sexually dimorphic organs, because of different metabolic demands between the two sexes. The liver is a key organ that plays fundamental roles in various physiological processes, including digestion, energetic metabolism, xenobiotic detoxification, biosynthesis of serum proteins, and also in endocrine or immune response. The sex-dimorphism of the liver is particularly obvious in oviparous animals, as the female liver is the main organ for the synthesis of oocyte constituents. In this work, we are interested in identifying molecular sexual dimorphism in the liver of adult medaka fish and their sex-variation in response to hepatotoxic exposures. By developing an integrative approach combining histology and different high-throughput omic investigations (metabolomics, proteomics and transcriptomics), we were able to globally depict the strong sexual dimorphism that concerns various cellular and molecular processes of hepatocytes comprising protein synthesis, amino acid, lipid and polysaccharide metabolism, along with steroidogenesis and detoxification. The results of this work imply noticeable repercussions on the biology of oviparous organisms environmentally exposed to chemical or toxin issues. ==== Body Sexual dimorphism terminology is widely used for organisms that perform sexual reproduction to describe physiological differences between two sexes at various biological levels. Although sexual dimorphism is generally considered at the anatomical or the behavioural level, it can also be extended to differences in the physiology of functions not directly involved in reproductive processes. According to different strategies for survival fitness of the two sexes and the consideration of their respective sexual specificities, exogenous perturbations could induce noticeable dissimilarities of endogenous response of individuals of the two sexes. Recent research observations have shown that a large number of genes exhibited deep sexual differences at the transcriptomic level in various tissues, suggesting that, in fact, sex-dependent genetic and hormonal regulations could also affect non-gonadal organs such as brain or liver123. This assumption is supported by numerous reports that have addressed that males and females may differ in their susceptibility to environmental or biological stresses, as well as in the differential responsiveness of the liver to various xenobiotics4. The liver is a key organ that plays fundamental roles in various physiological processes, including digestion, energetic metabolism, xenobiotic detoxification, biosynthesis of serum proteins, and also in endocrine or immune responses. Because of the different metabolic needs between sexes, especially during the reproductive phase, the liver is one of the most sexually dimorphic organs in terms of gene expression5. The first evidence of a sex-related difference in the rat hepatic steroid metabolism was published in 19536. Based on these initial observations, five decades of research have since then established the existence of a gonadal-hypothalamo-pituitary-liver axis that determines the differences between male and female liver. Moreover, the importance of hormone secretion patterns has been revealed and the understanding of hepatic gene regulation at the molecular level has advanced in mammals7. For example, various studies have shown that many hepatic genes associated with xenobiotic metabolisms, such as cytochrome P450, are expressed in a sex-dependent manner during the detoxification process4. Particularly, the sex-dimorphism of the liver is obvious in oviparous animals, as the female liver is the main organ for the synthesis of oocyte constituents, such as the yolk protein precursors (vitellogenins) and the zona pellucida proteins (choriogenins)8. As continental aquatic environments are threatened by a large spectrum of xenobiotics and pollutants, freshwater oviparous organisms such as fish are especially impacted by these potential toxicants, and their liver detoxification capabilities constitute essential defences for the fitness of these organisms. In this context, one can suppose that the various sexual dimorphisms of oviparous organisms, concerning energetic metabolism, detoxification and reproduction processes may drastically influence the hepatic responses of different sexes. In this work, we were interested in identifying the molecular sexual dimorphism in the liver of adult medaka fish and illustrating its implication in response to hepatotoxic exposures. Small fish such as the Japanese medaka (Oryzias latipes) have emerged as useful vertebrate model organisms, suitable for studying various physiological processes9, toxicological mechanisms10 and also ecotoxicological effects11. Medaka fish presents the advantages of small size, established models produced from inbred lines, rapid development and growth, high fecundity, omnivorousness, and also shows sugar and lipid metabolic profiles similar to those of mammals9. By developing an integrative approach comprising histology and different high-throughput omic investigations (i.e. metabolomics, proteomics and transcriptomics), we are now able to globally describe the sex-dimorphism in the medaka liver. To our knowledge, this constitutes the first systematic investigation of the liver sex-dimorphism in this model organism. Furthermore, under hepatotoxin perturbed conditions, sex-specific variation in molecular responses was investigated using quantitative proteomic analyses, implying potential different repercussions on the biology of fish environmentally exposed to chemical issues. Results Histology The liver of organisms under undisturbed lab condition presents in both sexes a characteristic architectural organization with polyhedral hepatocytes organized around the capillary sinusoids and the bile canaliculi, appearing in characteristic cord-like parenchymal structures. As shown in Fig. 1, medaka fish liver presents a sexual dimorphism at the cellular level based on histological observations of the hepatocytes. Indeed, male and female hepatocytes present obvious differences in their global cytoplasm appearance with a distinct distribution of vesicles that are revealed using hematoxylin-eosin-saffron (HES), periodic acid-Schiff/alcian blue (PAS) or toluidin blue staining. Whereas female hepatocytes present large isolated reserve vesicles (mostly one per hepatocyte) with dense contents of glycoprotein and/or glycogen (Fig. 1C,D,G,H), male hepatocytes exhibit more diffuse small vesicles (Fig. 1A,B,E,F). NMR metabolomics The hydrophilic fraction of the liver of medaka bred under undisturbed lab condition was investigated on 18 males and 18 females by 1H NMR analysis as shown in Fig. 2. Up to 237 different potential metabolites have been detected and relatively quantified according to the Batman R package analysis. The global analysis of the molecular pathway involved in liver metabolism reveals that the medaka liver metabolome presents a significant enrichment in a very wide diversity of processes, comprising principally glutathione, taurine, amino acid, carbohydrate, lipid, steroid hormone and tricarboxylic acid (TCA) cycle metabolisms (supplementary Table S1). Although our metabolomic analysis was performed on the hydrophilic fraction of the liver, we were able to observe various mostly hydrophobic metabolites, such as steroidal compounds, that testify of the intense lipid metabolism, especially in males. Principal component analysis (PCA) clearly discriminates specific liver metabolome between males and females according to its component 1 that comprises 76% of the total sample variation (Fig. 2A). Significantly over-represented metabolites in both sexes (|FC| > 2 and Pvalue < 0.05) were highlighted in a volcano plot (Fig. 2B). Whereas 59 molecules are over-represented in females, 103 appear to be over-represented in male metabolome (supplementary Table S2). According to their putative annotation provided by the Batman algorithm, a pathway enrichment analysis indicates that female-enriched metabolomes would rather be implicated in some saccharide and amino acid metabolic processes, whereas male-enriched metabolomes seem to exhibit signatures of steroid hormone biosynthesis and energy (TCA, nitrogen) processes, when the common metabolite set is more relevant to other amino acid and saccharide metabolism (Fig. 2C,D). Another interesting noticeable specificity of the medaka fish metabolome concerns the relative quantity of taurine and hypotaurine that are strongly over-represented in females and males, with 327 and 50 |FC|, respectively. Proteomics The cytosolic fraction of the liver proteome of medaka bred under undisturbed lab condition was investigated by high-throughput method of bottom-up proteomics. The trypsin-cleaved peptides were separated by nano-LC and analysed by high-resolution mass spectrometry, then the proteins were identified thanks to a large genomic dataset available for the Japanese medaka12. Among the 820 proteins identified in the three male pools, 64 appear to be potentially male-biased, because they were detected in both 3 male pools, and not in any of the three female pools. For females, 178 seem to be similarly biased of the 934 total proteins identified in the three female pools (Fig. 2A), and not in any male pool. This large protein sex-biased composition of the liver is also illustrated with semi-quantification of the proteins according to both MS and MS/MS data, represented in a volcano plot (Fig. 2B). Among 1241 identified proteins, 25 and 125 are significantly over-represented in males and females, respectively (|FC| > 3 and p < 0.01) (Supplementary Table S3). A global analysis of these sex-biased proteins according to the described functions of their human orthologs suggests that female-biased proteins are significantly related to tRNA and nucleotide sugar metabolism, that might be related with intense gene expression and synthesis processes, whereas male-biased proteins are related to bile and amino acid metabolisms, and common proteins to other TCA, sugar, lipid and amino acid metabolisms, characteristic of classical liver metabolic pathways (Supplementary Table S3). Among the top highly-abundant proteins in female livers dominate various isoforms of vitellogenins and choriogenins, together with fatty acid-biding, cytochrome P450 and various isoforms of ribosomal and translation-related proteins, whereas male-enriched liver proteins interestingly present other cytochrome P450 (CYP) isoforms, complement proteins, glutathione S-transferases (GSTs) and various TCA metabolism-related proteins, together with wap65 - a protein of unknown function whose transcript over-expression appears characteristic of the male Gulf pipe Syngnathus scovelli13 - highlighting important singularities of the liver proteome in both sexes (Fig. 3C,D). Transcriptomics PCA clearly discriminates between all male and female liver transcriptomes of medaka bred under undisturbed lab condition investigated by RNA-Seq analysis (Fig. 4A), and the volcano plot representation indicates a colossal over-expression of some genes in females compared with males (Fig. 4B). Indeed, some genes such as vitellogenin 1, 3 or 6 reach above 15 to 19 |log2 FC| variations (namely up to 500,000 |FC|) (Supplementary Table S4), representing a large portion of the total female liver transcriptome. In contrast, in male livers the most over-expressed gene, hydroxysteroid dehydrogenase 3, exhibits only up to 8 |log2 FC| (≈250 |FC|) variation in comparison with females. Additional pathway analyses, performed with the 375 and 147 significantly enriched transcripts in females and males, considering their human orthologs (Supplementary Table S4), indicate global enrichments (p < 0.1) of steroid and amino acid processes in both sexes when males also exhibit terpenoid-quinone and fatty acid biosynthesis over-representations. The list of the over-expressed genes in females comprises, along with some genes whose expressions are well known to be female-specific, such as vitellogenins, choriogenins and chorionic protease inhibitors, various forms of FAM20C genes, belonging to the serine threonine kinase 20c-like family. Various isoforms of the FAM20C protein family were similarly over-expressed in the female liver of the Gulf pipefish13. Although the biological function of the proteins of this family remains poorly documented, they are interestingly annotated in GO library as “cellular response to estrogen stimulus” genes, and may constitute female-specific markers in fish livers. With some agreement with our proteomic observation, other genes of interest belong to the cytochrome-P450 (CYP) family. Some member such as, for example, cytochrome-P450 27b1 and cytochrome-P450 2w1 appear clearly over-expressed in male and female medaka livers, respectively (Fig. 4C,D). Integrated pathway analysis To investigate and visualize the biological connectivity of the sex-enriched metabolites and transcripts, the network-generating algorithm of ingenuity pathway analysis (IPA) was used to maximize the interconnectedness of molecules based on all known connectivity in the database developed from Human molecular knowledge in the liver. The results of the IPA biological function analysis (Supplementary Table S6), represented as a bar chart and a heatmap, are shown in Fig. 5A,B, respectively. Lipid metabolism, molecular transport, small molecule biochemistry, inflammatory response, organismal development, vitamin and mineral metabolism, and free radical scavenging appear to be the most significantly represented functional categories according to the combined liver transcriptome and metabolome dataset. The sex-specificity of the molecules involved in these processes is indicated in the heatmap representation (Fig. 5B) that clearly shows a global up-regulation of molecules involved in small molecule biochemistry, lipid metabolism, tissue development, vitamin and mineral metabolism, energy production and carbohydrate metabolism in males, whereas in females most of the molecules involved in cellular movement, haematological system, inflammatory response, or immune cell trafficking appear largely up-regulated. The IPA network search shows that 2 of the top networks consisted predominantly of only female- and male-enriched molecules are related respectively to RNA post-transcriptional modification and lipid metabolism processes, as shown in Fig. 5C,D. These molecular network representations clearly illustrate the selected massive induction of some genes and metabolites related to RNA post-translational modification in females (Fig. 5C), and specific lipid metabolism processes connected with cholesterol metabolism and steroidogenesis, in males (Fig. 5D). Specificities of the toxicological molecular responses The quantitative proteomic analysis of liver of the medaka exposed to various hepatotoxic treatments was performed on nanoLC-ESI-MS/MS, leading to the identification of 1114 proteins and to the differential quantification of 177 and 185 proteins in males and females, respectively (Fig. 6A). Only above a quarter of them (49 proteins) appears to be common between the two sexes. Protein quantification in each group of adult medaka fish, chronically exposed to the various cyanobacterial hepatotoxic treatments (CHT1-3) was reported, according to the relative intensity normalized with the controls set at 0, and the reliable quantifications in male and female livers were represented in a heatmap with hierarchical cluster analysis. This analysis clearly revealed a distinguishable sex-dependent response of medaka fish to the various cyanobacterial hepatotoxic treatments according to the group distribution given by the clustering analysis, which is based on the pattern of relative abundance of up- and down-regulated proteins (Fig. 6B). These cyanobacterial hepatotoxic treatments dysregulate various proteins of various molecular function categories comprising lipid, amino acid, carbohydrate and TCA metabolisms and detoxification processes (Supplementary Table S5). The sex-specificity of the liver molecular response suggests that identical hepatic stress could impact these various molecular processes differently, potentially inducing dissimilar biological repercussions in the organisms of two sexes. In our example, male-enriched protein dysregulations concern rather TCA, steroid, fatty acid, amino acid and vitamin B6-7 metabolism pathway categories, whereas female-enriched protein dysregulations are rather related to tRNA biosynthesis, amino acid, glutathione, xenobiotic and drug metabolism pathways. Discussions The liver is a key organ in vertebrates performing a large diversity of vital functions, including processing and storage of nutrients, maintenance of serum composition, bile production, and xenobiotic detoxification. It is primarily an exocrine gland, secreting bile into the intestine, but it is also an endocrine organ and a blood filter. The liver is a metabolic factory, which synthesizes and breaks down a variety of substances, comprising the production of bile salt anions, the synthesis of urea and many plasma proteins, the metabolism of glycogen, cholesterol and fatty acids, the detoxification of many drugs, and the processing of steroid hormones and vitamin D. Studies carried out in rodents have established that sex-based differences in liver function also characterize many drug-metabolizing enzymes (DMEs), including sulfotransferases, glutathione S-transferases, P450s and other steroid metabolizing enzymes4. The sexual dimorphism of liver gene expression is not confined to DMEs, and it concerns more than 1000 genes in these organisms2, including plasma lipoproteins, pheromone binding proteins, regulators of fatty acid homeostasis, nuclear receptors, and other transcription factors3. Similar to the mammalian liver, the teleost liver plays an important role in the metabolic homeostasis of the whole organism, in addition to that, oviparous vertebrate-specific processes related to the synthesis of various oocyte protein precursors (i.e. mainly vitellogenins and choriogenins, together with other minor vitamin-binding proteins) are synthesized in females under the direct control of estrogens, which bind to estrogen receptor complex and activate the translation of messenger RNA via cis-regulation mechanisms814. The massive rate of synthesis of vitellogenin in the egg-laying animal causes considerable ultrastructural changes in liver cells, which are characterized by extensive proliferation of the rough endoplasmic reticulum and the Golgi apparatus15. In mature female medaka, a remarkable part of the liver metabolism might be dedicated to these reproduction-related processes, as each female can spawn above 30 mature oocytes daily16. This massive synthesis is known to induce large cellular and molecular modifications, as it can also increase the lipid synthesis of hepatocytes5. This metabolic adjustment to maintain the reproductive competency of the female constitutes one of the physiological bases for the extended sexual dimorphism in fish livers. Our histological observations of male and female medaka livers clearly showed differences of reserve vesicle within hepatocytes, which is consistent with ultrastructural modifications in the liver cells between males and females. Interestingly, previous observation of mature medaka liver under transmission electron microscope shows that, in the perinuclear region, granular endoplasmic reticulum, mitochondria, and peroxisomes appear largely increased in number in female hepatocytes15. These hepatocyte sexual differences in cellular organization and content might be related to the intense activity of protein synthesis and consequently the high energy requirement of female hepatocytes. Indeed, in sexually mature fish, as in other oviparous vertebrates, livers globally present morphological, molecular and functional sexual-dimorphisms151718. The liver of female performs an important function in the synthesis of a large set of proteins involved in the oocyte envelop and vitellogen reserves, whereas male liver hepatocytes do not exhibit such activity. Although the precise function of taurine and hypotaurine, which were highly abundant in female and male livers, respectively, and the balance between the two remains poorly documented in fish liver19, one previous study has reported the influence of taurine on egg maturation20. To date, only one investigation performed in adult zebrafish has attempted to compare male and female metabolomes, according to various analytical approaches, including GC-MS, LC-MS and NMR, and has observed a significant up-regulation in various fatty acids, together with valine, acetate, glutamate, glutamine, creatinine and betaine in female liver21. The sex-biased pattern depicted in our transcriptomic analyses appears acutely congruent and even more contrasted than our global proteome investigation. These strong sex-biases testify of the intensity of the female liver efforts for the gene expression and the synthesis of the oocyte precursor proteins814, and the involvement of the male liver in steroid hormone and metabolism processes22, such as urea and energy cycles9, respectively. Previous investigation of medaka fish has observed that liver transcriptome globally exhibits significant enrichment in the expression of genes related to macromolecule, RNA, and nitrogen compound metabolic processes with regard to the gene expression in other tissues, but it has not considered the differences between sexes2324. Previous works performed on zebrafish and Gulf pipefish have highlighted large sets of genes whose expression appears to be driven by sex-dependent processes1625. On one side, zebrafish transcriptomic approach reveals that the female-over-expressed gene list included vitellogenins and zona pellucida glycoproteins, many ribosomal proteins, and estrogen receptor 1, in contrast, the list of male-over-expressed genes contains fatty acid-binding protein2, apolipoprotein 4, and also genes that are supposed to be involved in anti-inflammatory processes, such as complement factors 9b and 3b, together with several chitinases25. On the other side, transcriptomic investigation performed on Gulf pipefish shows a quite similarly high over-expression in characteristic genes of females, such as vitellogenin b and c, choriogenin h, and zpc 4-like, in addition to estrogen receptor 1 and various fam20c isoforms, whereas males exhibit less intense over-expression of specific genes, comprising various metabolism related genes such as hint3, ctl, nsun3 and wap6513. Moreover, primary investigation of the medaka fish liver transcriptome by microarray analysis indicates that the female-specific transcript list comprises some previously characterized female-specific genes such as vitellogenins, choriogenins, ZP family genes, cyclins B and 42S nucleoprotein, whereas most of the male-specific transcripts have not yet been assigned or characterized26. Similarly, previous proteinaceous investigations have highlighted that female liver proteome moreover contains massive amounts of oocyte precursor proteins (i.e. vitellogenins, choriogenins and fatty acid-binding proteins) that are being secreted by the hepatocytes17, together with variations between sexes in drug metabolism capabilities4. CYPs constitute a diverse group of enzymes that are potentially involved in key reactions of oxidation of organic substances, such as drug detoxification17 and steroid hormone metabolism4713. The sexual polymorphism in the expression of these enzymes may have fundamental repercussion on liver physiology such as drug-metabolism processes. In addition to our transcriptomic data that gives a congruent view with previously published observations, our systematic investigation constitutes an unprecedented opportunity to globally depict the medaka liver sexual dimorphism at different molecular and cellular levels. By developing an integrative approach combining high output proteomic, RNA-Seq transcriptomic and non-targeted metabolomics outputs, together with histological examinations, we are able to appreciate the wideness and the deepness of the sexual dimorphism, in terms of both number and intensity of the sex-dependent dysregulations. At the mRNA level, twice more sex-over-expressed transcripts appear in females, comprising some genes involved in ovogenesis (i.e. vitellogins and choriogenins), and reach very high fold change values (up to 50,000 and 250 FC in females and males, respectively). Proteomic investigations also show more proteins that appear to be specific of female proteomes, with higher fold change too (up to 750 and 50 FC in females and males, respectively). The quantitative metabolome analysis performed by NMR indicates that the difference of fold changes between female- and male-enriched metabolites appear to be much higher in female (up to 350 and 50 FC in females and males, respectively). The observations of globally more intensive and numerous molecular up-regulations in female livers are in agreement with the conception of an oviparous female liver that is in charge of extra metabolic activity, according to their massive production of oocyte yolk stocks and chorion precursors, which have substantial impacts on both amino acid, saccharide and fatty acid metabolism of the global liver activities8. As the mature female liver is considered to be more physiologically and energetically solicited than male’s, we assume that female liver could be consequently more susceptible to hepatotoxic stressors. In our experiments and according to other investigations on small fish ecotoxicology, various hepatotoxic stressors such as cyanotoxins11, pesticides2728 or aromatic hydrocarbons29 induce a higher toxicological response in female livers, suggesting that females would be more sensitive to the effects of those molecules than males151718. However, some examples also attest to a higher susceptibility of male livers according to certain specific exposure conditions to toxic chemicals30, and a careful investigation of the dimorphic detoxification capabilities should aim at being performed for each specific chemical or hepatic stressor evaluation. Indeed, as the liver, being the principal detoxification organ, presents noticeable sex-differences in its drug metabolism (e.g. CYP P450 isoforms) and homeostasis capabilities4, it is likely that for some hepatic stressors the detoxification performance, which could be clearly distinguishable between the two sexes, should be systematically considered in environmental toxicology evaluations. Overall, in addition to providing first insight into the molecular mechanism underlying the sex-specificity of the livers of oviparous organisms, and concerning important liver processes, such as energetic metabolism, detoxification and reproduction, our molecular integrated research demonstrates also that, freshwater oviparous organisms such as medaka fish present a net sex-dimorphism in the molecular response specifically induced by chronic hepatotoxin exposures. Furthermore, numerous reports show that fish populations are adversely affected by environmental estrogens, which can potentially present ecological adverse effects, via the induction of the synthesis of oogenesis proteins. These xenoestrogens can also impact various liver processes that also exhibit sex-dimorphisms, such as metabolism and biotransformation enzymes that directly influence the stress resistance capabilities of the organism. Additional studies are still needed to validate these findings at higher levels of biological organization, and to fully estimate their consequences for different populations of the global ecosystem. Methods Medaka fish Medaka fish (Oryzias latipes) of the inbred Cab strain were used for this experiment. The animals were handled and experiments were performed in accordance with European Union regulations concerning the protection of experimental animals and the experimental procedures were approved (N°68-040 for 2013-18) by the “Cuvier’s ethical committee” of the Muséum National d’Histoire Naturelle (French national number C2EA - 68). All fish used in this study (during summer 2014) originate from the same broodstock (F0 from genitors provided in November 2013 by the Amagen CNRS/INRA platform - Gif-sur-Yvette, France). All histopathology, metabolomics, proteomic and transcriptomic analyses, except for iTRAQ proteomic with chronically exposed fish, were performed on the untreated fish breed and kept under control conditions, described as follow. Six month-old adult fish (around 0.55 ± 0.08 and 0.59 ± 0.10 g, for males and females, respectively; n = 36), mature and sexually active (with secondary sexual character well developed; sex determination was confirmed by further histology of the gonads for 36 individuals), were maintained at 25 ± 1 °C, with 15 h:9 h light/dark cycle (in reproductive cycle). Fish were raised in 20 L glass aquaria (in triplicate tanks, containing above 10 male and 10 female per tank) filled with a continuously-aerated mixture of tap water and reverse osmosis filtered water (1/3–2/3, respectively), which was renewed once a week. Fish were fed three times a day with commercial food for juvenile salmon, supplemented once a day with fresh artemia, and were inspected three times daily, and no abnormal behaviour, nor mortality was observed. Fish were randomly selected, briefly anesthetized in buffered 0.1% MS-222, sacrificed. The liver samples were collected and prepared for further analysis, as described below. Mature and sexually active fish, bred in the same conditions of light, temperature and nourishing, were used for chronic exposure to hepatotoxin then for iTRAQ quantitative proteomic analysis, as further described in the following paragraph. Five month-old fish were transferred to experimental tanks (15-L glass tanks, containing 5 males and 5 females each) two weeks prior to the beginning of the experimentation. Hepatotoxic mixture exposure Fish were exposed during 21 days to environmental concentrations of 3 different cyanobacterial hepatotoxic treatments31 (CHT), called CHT1-3, as well as to solvent control conditions (Control). The experiment was performed in triplicate tanks for each treatment (comprising 30 fish, 15 males and 15 females per treatment) and the exposure conditions were maintained by the renewal of two-thirds of the total aquaria volume (10 L) containing hepatotoxic mixture every 2 days. Fish were inspected three times daily, and no abnormal behaviour, nor mortality was observed during all the experimentation. At the end of the experiment, fish were anesthetised in 0.1% MS-222, euthanized and then livers were sampled for further analyses using iTRAQ quantitative proteomics. Histopathology Liver samples from at least 9 males and 9 females bred under unstressed control conditions were fixed in cold 10% buffered formalin (4 °C, 48 h), then transferred into 70% ethanol, dehydrated in successive baths of ethanol (from 70 to 95%), and then embedded in paraffin. Blocks were cut in 3–5 μm-thick sections, and slides were stained with hematoxylin-eosin-saffron (HES) or periodic acid-Schiff/alcian blue (PAS), according to the standard histological procedure. Alternatively, liver samples were fixed with a mixture of paraformaldehyde (2%), glutaraldehyde (0.5%), picric acid (0.5%) and sucrose (0.18 M) in 0.1 M pH 7.4 Sørensen buffer prior to post-fixation in osmium tetroxide (1%). Samples were then dehydrated in ethanol, embedded in the epoxy mixture (Spurr’s resin), and cut in semi-thin 0.5 μm-thick sections and stained with toluidine blue (TB). Metabolome 1H-NMR spectra Liver extraction was carried out using the methanol/chloroform/water method (ratio 2/2/1.8). Fresh frozen livers of 18 individuals for each sex (6 individuals from each triplicate tank) bred under unstressed control conditions were weighted and then homogenized in the ice cold methanol (8 mL.g−1 of tissue; AnalaR Normapur, min. 99.8%) and ice cold milliQ water (2.5 mL.g−1) and then vortexed for 1 min. Subsequently, ice cold chloroform (4 mL.g−1; Normapur, 99.3%) and milliQ water (4 mL.g−1) were added. Then, the mixture was vortexed for 1 min and incubated on ice for 10 min to partition. The supernatant was then centrifuged at 4 °C for 10 min at 2,000 g. The upper polar fraction was then transferred to 2 mL Eppendorf tubes, dried under Speed-vac device and then kept at −80 °C until NMR analysis. The extracts were dissolved in 550 μL of 0.1 M sodium phosphate buffer prepared in D2O (10% v/v) containing 0.25 mM sodium-3-tri-methylsilylpropionate (TMSP) as an internal standard, then were transferred to a 5-mm NMR tube (Norell, France) and analyzed immediately by 1H-NMR. All NMR data were recorded at 298 K on a 600 MHz Bruker AVANCE III HD spectrometer equipped with a 5 mm TCI CryoProbe (1H-13C-15N) with Z-gradient. One-dimensional 1H NMR spectra were acquired using a standard Bruker noesygppr1d pulse sequence to suppress water resonance. Each spectrum consisted of 512 scans of 32,768 data points with a spectra width of 7.2 kHz, a relaxation delay of 3 s and an acquisition time of 2.3 s. A Quality Check (QC) sample was injected every 6 samples, in order to verify that no significant drift of the analysis occurs, according to expected reference. Spectra were then processed with Topspin software (Bruker) for alignment and noise reduction, and analyzed for bucketing, annotation and quantification with the Batman R package32. Individual metabolite intensities were compared according to sex groups using Metaboanalyst 3.0 online tool33 for PCA, PLS-DA, volcano plot reconstruction and metabolite pathway enrichment analyses. Proteomic analysis Nine livers from adult male and female medaka fish bred under unstressed control conditions were randomly pooled (one fish liver from each triplicate tank in each triplicate pool) and homogenized on ice with a Dounce homogenizer in 500 μL of a solution of 6 M guanidine hydrochloride, 500 mM triethylammonium bicarbonate buffer (TEAB, pH 8.3), 0.1% Triton X-100 and 10 μg of protease inhibitor mixture (Roche, Switzerland). The homogenates were centrifuged at 4 °C (12,000 g; 10 min), and then the supernatants were collected. Proteins were precipitated with cold acetone (−20 °C; overnight), centrifuged at 4 °C (2,000 g; 4 °C), and then resuspended in 500 mM TEAB with 6 M urea and 0.1% SDS. The protein concentration was measured using a micro-BCA kit (Sigma-Aldrich, USA), with BSA as a protein standard. 100 μg of each liver protein pool was used for digestion with 5 μg of proteomic-grade trypsin (Sigma-Aldrich, USA), reduced with 2 mM tris-(2-carboxyethyl)phosphine (TCEP) with and cysteine-blocked with 10 mM methyl methane-thiosulfonate (MMTS), prior to analyses with a Q Exactive™ Hybrid Quadrupole-Orbitrap™ mass spectrometer (Thermofisher Scientific). Liver protein digests were concentrated on C18 stages tips, recovered in 40 μL 2% aqueous TFA, 2% ACN before injection in triplicates (6 μL injected). NanoLC was performed on an Ultimate 3000 RSLCnano System (Thermofisher Scientific): digests were desalted on a trap column (Pepmap, C18 300 μm x 5 mm, 5 μm 100 Å, Dionex) with water containing 2% ACN with 0.1% formic acid (solvent A) for 6 min, and the peptides were finally eluted from a separation column (Pepmap, C18 75 μm, x 500 mm, 3 μm 100 Å, Dionex). The separation gradient as optimized for the samples and is divided into 3 successive slopes: 2–20% in 120 min, 20–35% in 45 min and 35–80% ACN + 0.1% formic acid (solvent B) at a flow rate of 300 nL.min−1. Each MS spectrum acquisition (m/z 400–2000, 70,000 Res.) was followed by up to ten data dependent HCD MS/MS spectra (first fixed mass m/z 90, 17,500 Res., 30 normalized collision energy) with an isolation window m/z 2 and a dynamic exclusion window of 30 s. All MS/MS-analyzed samples were analyzed using Mascot 2.4.1 (Matrix Science, UK) and X!Tandem with Scaffold software (version 4.5.1; Proteome Software, USA) to search Uniprot databases of Teleostei (downloaded in December 2015). The ion mass tolerance and the parent ion tolerance were set to 20 mDa and 10 ppm, respectively. The methyl methanethiosulfonate of cysteine was specified as fixed modifications. Oxidation of methionine and deamination of N and Q were specified as variable modifications. The Scaffold was used to probabilistically validate the protein identifications derived from MS/MS sequencing results. Normalized semi-quantifications and identification probability of male and female identified proteins were estimated using Scaffold+ default parameters from MS and MS/MS data for proteins presenting at least two peptides. iTRAQ based quantitative proteomic was performed in triplicates for each experimental group on 3-pooled livers from nine livers from adult male and female medaka fish for each condition (CHT1, CHT2, CHT3 and Control) that were randomly pooled (one fish liver from each triplicate tank in each triplicate pool) as previously described34. Nano-LC-MS/MS analysis of the 8-plex tagged peptide digests was performed using the same top-10 strategy. A Quality Check (QC) sample was injected every 5 samples, in order to confirm that no significant drift of the analysis occurs, according to expected minimal identification number. Protein quantification was performed with Scaffold Q+ (4.5.1) using the isobaric tag peptide and protein identifications. Protein identifications were accepted if they could be established with more than 99.0% probability, and contained at least 2 identified peptides that were quantified using the centroid reporter ion peak intensity. Protein quantitative ratios were calculated as the median of all peptide ratios of the three consecutive runs. Quantitative ratios were log2 normalized for final quantitative testing, using control value set up as a reference sample in both sexes. Heatmap protein quantification was represented using Gene-E freeware (http://www.broadinstitute.org/cancer/software/GENE-E/) using Spearman correlation’s value for sample and protein clustering analyses. RNA-Seq Two or three pooled livers from untreated adult males and females randomly selected from the three replicate tanks were homogenized using a bead beater. Total RNA of pooled female and male livers (6 biological replicates, for each sex) was isolated and purified using RNeasy Plus Mini Kit with gDNA eliminator spin (Qiagen). RNA quantity and quality were evaluated using Qubit RNA Assay Kit in Qubit®2.0 Fluorometer (Life Technologies, USA) and an Agilent Bioanalyzer 2100 eukaryote total RNA Pico series II chip (Agilent Technologies Inc., USA), respectively. The RIN values of all samples (3 pool samples for males and 4 pool samples for females) further selected for RNA-seq analysis were over 7.7. The transcriptome libraries were prepared from total RNA using Illumina TruSeq Stranded mRNA Sample Preparation kits (Illumina Inc., USA) following the manufacturer’s handbook. Briefly, the mRNA was purified using poly-T oligo-attached magnetic beads and then fragmented into small pieces that were then used for synthesizing the first- and second-strand cDNA. After end repair, single nucleotide A (adenine) addition and adaptor ligation, the fragments were amplified with a 10-cycles PCR program. The libraries were sequenced on the Illumina Hi-Seq1000 instrument using the 51 bp single-end sequencing strategy with the TruSeq SBS kit V3-HS 50-cycles (Illumina Inc., USA). Raw reads were first cleaned by removing adaptors using Cutadapt-1.3 and only 51 bp-long reads were kept. The global quality of the reads was checked using the FastQC 0.10.1 and good global Phred scores (>30) were obtained in all the libraries. However, the analysis of the unicity of reads shows a high level of sequence duplications, so a step of duplicated reads removal was conducted using a python script that analyses the quality of reads and keeps the one with the best global quality score. Tophat2 (v2.0.10)35 was used to map the clean unique reads against the medaka genome (release 81) downloaded from Ensembl (ftp://ftp.ensembl.org/pub/release-81/fasta/oryzias_latipes/dna/). Multiple hits were removed by samtools (v0.1.18) and read counting on gene exons was accomplished by HTSeq-count (v0.6.1p1)36 in union mode against the annotation of medaka genomes downloaded from Ensembl (ftp://ftp.ensembl.org/pub/release-81/gtf/oryzias_latipes/). DESeq2 v1.8.137 was used to do differential expressed gene analysis on the raw count data. Genes were considered differently expressed when the p value below 0.001, using the control group as reference. Furthermore, we only included genes with expression level of at least 4 |FC| in order to capture more physiologically relevant genes. Then, we identified 375 female- and 147 male-enriched transcripts (Supplementary Table S4). Messenger RNA expression levels were secondarily confirmed by RT-qPCR on 8 genes randomly selected within the expression intensity gradient that presented correlation coefficient greater than 0.95 between these two techniques in both sexes. Molecular network analysis Molecular pathway was determined for our merged transcriptome and metabolome data using the Ingenuity Pathway analysis software (V01-04; Qiagen) with the Human orthologous of medaka proteins available from Ensembl online platform (http://www.ensembl.org), according to specific Ingenuity Knowledge Database (Genes and Endogenous Chemicals), which is a repository of biological interactions and functional annotations. The fold change values (females vs males) and p values calculated according to the quantifications of all replicates for 2214 gene expressions (FDR < 0.05) and 245 metabolites (HMDB numbers) were imported into IPA, then “Core Analysis” was performed, with default setting on liver tissue and relaxed filters, including both direct and indirect relationship between our dataset and the reference annotations, in order to interpret data in the context of biological pathways, molecular functions and networks. Additional Information How to cite this article: Qiao, Q. et al. Deep sexual dimorphism in adult medaka fish liver highlighted by multi-omic approach. Sci. Rep. 6, 32459; doi: 10.1038/srep32459 (2016). Supplementary Material Supplementary Information This work was supported by grants from the CNRS Défi ENVIROMICS “Toxcyfish” project, from the Sorbonne Universités “DANCE” and “Procytox” project, and from the ATM “Cycles biologiques: evolution et adaptation” of the MNHN to Dr. Benjamin Marie. We would like to thank the China scholarship council and the French minister for the research for their financial supports to Qin Qiao and Séverine Le Manach, respectively. We thank the Amagen platform for providing medaka fish cab strain, and the Imagif platform for RNA sequencing. The NMR spectra were acquired at the Plateau technique de Résonance Magnétique Nucléaire, UMR 7245 Molécules de Communication et d’Adaptation des Micro-organismes, Muséum National d’Histoire Naturelle, Paris, France. We thank the RBCF (RDDM’s bioinformatics core facility of the MNHN) for the use of the server. We also thank Marie-Claude Mercier for its administrative support. Finally, we would like to thank the anonymous reviewers whose constructive comments significantly contribute to improve the present manuscript. Author Contributions Q.Q., S.L.M., B.S., G.B., A.M., H.H., C.B., M.E. and B.M. conceived the experiments, Q.Q., S.L.M., B.S., C.D. and B.M. conducted the experiments, Q.Q., S.L.M., B.S., B.M., E.D.-B., A.P., L.P., A.M., A.B., L.M., J.V., C.D. and B.M. performed the analyses and analysed the results. All authors reviewed the manuscript. Figure 1 Histological investigations of male and female medaka livers. Representative histological observations under a light microscope of thick or thin sections of adult medaka liver stained with HES (A–D), PAS (E,G), and toluidine blue (F,H), for males (A,B,E,F) and females (C,D,G,H). Scale bares represent 20 μm. g, glycogen reserve; n, nucleus; m, membrane; c, cytosol; v, vesicle. Figure 2 Metabolomics of male and female medaka livers by 1H NMR. Principal component analysis (PCA) performed with the quantification values of the 237 metabolites detected by Batman’ R package algorithm (Supplementary Table S2) from the 18 males and 18 female individual NMR spectra (A). Volcano plot representation of the 237 metabolites according to female/male fold change average and significance of the differences (|log2 FC| > 1 corresponding to |FC| > 2 and log10 Pvalue < 1.3 corresponding to Pvalue < 0.05) (B). Female and male over-represented metabolites are determined with positive and negative significant FC (F/M) values and are shown in red and blue, respectively, and are represented with darker colours when metabolite presenting VIP values are superior to 1, according to PLS-DA analysis. Top-25 lists of the putative annotations of male- and female-representative metabolites (C and D, respectively). Figure 3 Proteomics of male and female medaka livers. Unscaled Venn’s diagram of the protein identified with at least 95% protein identification certainty in all of the 3 different 3-individual pools of male and/or female medaka livers (A). Volcano plot representation of the 1241 proteins according to female/male fold change average and significance of the differences (|log2 FC| > 1.5 corresponding to |FC| > 3 and log10 Pvalue < 2 corresponding to Pvalue < 0.01) determined according to Scaffold 4.5.1 semi-quantitative values based on both MS and MS/MS data (B). Female and male over-represented proteins are determined with positive and negative significant FC (F/M) values and are shown in red and blue, respectively. Top-25 lists of the male- and female-superabundant proteins (C and D, respectively). Figure 4 Transcriptomics of male and female medaka livers investigated by RNA-seq approach. Principal component analysis (PCA) performed according to the transcript count for the 16,523 genes encoding medaka liver proteins (Supplementary Table S4) from the 3 male and 4 female pooled cDNA sequenced by Hiseq 1000 comprising at least 30 million reads per libraries (A). Volcano plot representation of the gene expression according to female/male fold change average and significance of the differences (|log2 FC| > 2 corresponds to |FC| > 4 and log10 Pvalue < 3 corresponds to Pvalue < 0.001) (B). Female and male over-expressed genes are determined with positive and negative significant FC (F/M) values and are shown in red and blue, respectively. Top-25 lists of the male- and female-over-expressed genes (C and D, respectively). Figure 5 Ingenuity pathway analysis performed on male versus female fold change values (M/F FC) of both transcriptomic and metabolomic data. Top dysregulated molecular pathways represented on bar chart (A) and heatmap (B). Molecular functions that are specifically activated in males and females (Supplementary Table S6) are indicated in blue and red, respectively. Examples of significant molecular networks related to RNA post-translational modification (C) and lipid metabolism (D) processes (score = 38 and 28). Relative up-regulation of transcripts and metabolites in males or in females are indicated in blue and red, respectively. Figure 6 Global proteomic effects of the different hepatotoxic cyanobacterial extract exposures (CHT1-3) on medaka fish liver. Venn’s diagram (A) and heatmap representation (B) of proteinaceous dysregulation investigated by iTRAQ quantitative proteomic approach for both males and females. Samples are normalized according to male and female controls and significance of protein dysregulation was settled at |log2FC| > 0.25. Main metabolic categories of dysregulated proteins in males, females or in both sexes were determined according to Metaboanalyst 3.0 molecular pathway searches (Supplementary Table S5). Down-regulated proteins are indicated in yellow, up-regulated proteins in blue, and missing values in grey. Clustering was performed according to Pearson’s correlation coefficient values. ==== Refs Davies W. & Wilkinson L. S. It is not all hormones: alternative explanations for sexual differentiation of the brain . Brain Res 1126 , 36 –45 (2006 ).17101121 Yang X. , Zhang B. , Molony C. , Chudin E. , Hao K. . Systematic genetic and genomic analysis of cytochrome P450 enzyme activities in human liver . Genome Res. 20 , 1020 –1036 (2010 ).20538623 Conforto T. & Waxman D. J. Sex-specific mouse liver gene expression: genome-wide analysis of developmental changes from pre-pubertal period to young adulthood . Biol. Sex Diff. 3 , 9 (2010 ). Waxman D. & Holloway M. Sex differences in the expression of hepatic drug metabolizing enzymes . Mol. Pharmacol. 76 , 215 –228 (2009 ).19483103 Roy A. & Chatterjee B. Sexual dimorphism in the liver . Ann. Rev. Physiol. 45 , 37 –50 (1983 ).6189449 Hübener H. J. & Amelung D. , Enzymatische Umwandlungen von steroiden (I). Vergleich der steroidumwandlung in leber end nebenniere . Hoppe-Seylers Z. Physiol. Chem 293 , 137 –141 (1953 ). Mode A. & Gustafsson J.-A. Sex and the liver – a journey through five decades . Drug Metabolism Reviews 38 , 197 –207 (2006 ).16684657 Arukwe A. & Goksøyr A. Eggshell and egg yolk proteins in fish: hepatic proteins for the next generation . Comp. Hepatol. 2 , 4 (2003 ).12685931 Fujisawa K. , Takami T. , Kimoto Y. , Matsumoto, Yamamoto N. . Circadian variations in the liver metabolites of medaka (Oryzias latipes) . Sci. Rep. 6 , 20916 (2016 ).26862003 Marie B. , Huet H. , Marie A. , Djediat C. , Puiseux Dao S. . Effects of a toxic cyanobacterial bloom (Planktothrix agardhii) on fish . Aquat. Toxicol. 114 , 39 –48 (2012 ).22414781 Deng D. F. , Zheng K. , Teh F. C. , Lehman P. W. & Teh S. J. Toxic threshold of the dietary microcystin (-LR) for quart medaka . Toxicon 55 , 787 –794 (2010 ).19941879 Ahsan B. , Kobayashi D. , Yamada T. , Kasahara M. , Sasaki S. . UTGB/medaka: genomic resource database for medaka biology . Nucleic Acids Res. 36 , 747 –52 (2008 ). Rose E. , Flanagan S. P. & Jones A. G. The effects of synthetic estrogen exposure on the sexually dimorphic liver transcriptome of the sex-role-reversed Gulf pipefish . PLoS ONE 10 , e0139401 (2015 ).26448558 Van der Ven L. , Holbech H. , Fenske M. , Van den Brandholf E. J. , Gielis-Proper F. K. . Vitellogenin expression in zebrafish Danio rerio: evaluation by histochemistry, immunohistochemistry, and in situ mRNA hybridization . Aquat. Toxicol. 65 , 1 –11 (2003 ).12932697 Braunbeck T. , The S. , Lester S. & Hinton D. Ultrastructural alterations in liver of medaka (Oryzias latipes) exposed to diethylnitrosamide . Toxicol. Pathol. 20 , 179 –196 (1992 ).1475579 Naruse K. , Tanaka M. & Takeda H. In Medaka a model for organogenesis, Human Disease and Evolution , Springer, Berlin, 387 pp (2011 ). Viitaniemi H. M. & Leder E. H. Sex-biased protein expression in the threespine stickleback, Gasterosteus aculeatus . J. Proteome Res. 10 , 4033 –4040 (2011 ).21726075 Zhang X. , Ung C. , Lam S. , Ma J. , Chen Y. Z. . Toxicogenomic analysis suggests chemical-induced sexual dimorphism in the expression of metabolic genes in zebrafish liver . PLoS ONE 7 , e51971 (2012 ).23272195 Salze G. & Davis D. A. Taurine: a critical nutrient for future fish feeds . Aquaculture 437 , 215 –229 (2015 ). Matsunari H. , Hamada K. , Mushiake K. & Takeuchi T. Effects of taurine levels in broodstock diet on reproductive performance of yellowtail Seriola quinqueradiata . Fish. Sci. 72 , 955 –960 (2006 ). Ong E. S. , Chor C. F. , Zou L. & Ong C. N. A multi-analytical approach for metabolomics profiling of zebrafish (Danio rerio) livers . Mol. Biosys. 5 , 288 –298 (2009 ). Zhang X. , Kecker M. , Park J.-W. , Tompsett A. , Newsted J. . Real-time PCR array to study effects of chemicals on the Hypothalamic-Pituitary-Gonadal axis of the Japanese medaka . Aquat. Toxicol. 88 , 173 –182 (2008 ).18534694 Lai K. P. , Li J.-W. , Wang S. Y. , Chiu J. M. , Tse A. . Tissue-specific transcriptome assemblies of marine medaka Oryzias melastigma and comparative analysis with the freshwater medaka Oryzias latipes . BMC Genomics 16 , 135 (2015 ).25765076 Murata Y. , Yasuda T. , Watanabe-Asaka T. , Oda S. , Mantoku A. . Histological and transcriptomic analysis of adult Japanese medaka sampled onboard the International Space Station . PLoS ONE 10 , e0138799 (2015 ).26427061 Zheng W. , Xu H. , Lam S. , Luo H. , Karuturi R. & Gong Z. Transcriptomic analyses of sexual dimorphism of the zebrafish liver and the effect of sex hormones . PLoS One 8 , e53562 (2013 ).23349717 Kishi K. , Kitagawa E. , Onikura N. , Nakamura A. & Iwahashi H. Expression analysis of sex-specific and 17β-estradiol-responsive genes in the Japanese medaka, Oryzias latipes, using oligonucleotide microarrays . Genomics 88 , 241 –251 (2006 ).16757147 Werner I. , Geist J. Okihiro M. , Rosenkranz P. & Hinton D. E. Effects of dietary exposure to the pyrethroid pesticide esfenvalerate on medaka (Oryzias latipes) . Mar. Environ. Res. 54 , 609 –614 (2002 ).12408625 El-Alfi A. , Bernache E. & Schlenk D. Gender differences in the effect of salinity on aldicarb uptake, elimination, and in vitro metabolism in Japanese medaka, Oryzias latipes. Aquat. Toxicol. 61 , 225 –232 (2002 ).12359392 Oh J. , Moon H.-B. & Choe E. Alterations in differentially expressed genes by exposure to a mixture of carcinogenic polycyclic aromatic hydrocarbons in the liver of Oryzias latipes . Environm. Toxicol. Pharmacol. 33 , 403 –407 (2012 ). Yu W. , Shi Y. , Fong C. , Chen Y. , van de Merwe J. P. . Gender-specific transcriptional profiling of marine medaka (Oryzias melastigma) liver upon BDE-47 exposure . Comp. Biochem. Physiol. part D 8 , 255 –262 (2013 ). Zurawell R. , Chen H. , Burke J. M. & Prepas E. Hepatotoxic cyanobacteria: a review of the biological importance of microcystins in freshwater environments . J. Toxicol. Environ. Health B Crit. Rev. 8 , 1 –37 (2005 ).15762553 Hao J. , Liebeke M. , Astle W. , De Iorio M. , Bundy J. G. & Ebbels T. M. Bayesian deconvolution and quantification of metabolites in complex 1D NMR spectra using BATMAN . Nature Protocols 9 , 1416 –1427 (2014 ).24853927 Xia J. , Sinelnikov I. , Han B. & Wishart D. S. MetaboAnalyst 3.0 - making metabolomics more meaningful . Nucl. Acids Res. , 10.1093/nar/gkv380 (2016 ). Le Manach S. , Khenfesh N. , Huet H. , Qiao Q. , Duval C. . Genedr-specific toxicological effects of chronic exposure to pure microcystin-LR or complex Microcystis aeruginosa extract on adult medaka fish . Environm. Sci. Technol. , 10.1021/acs.est.6b01903 (2016 ). Kim D. , Pertea G. , Trapnell C. , Pemetel H. , Kelley R. & Salzberg S. L. TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions . Genome Biol. 14 , R36 (2013 ).23618408 Anders S. , Pyl P. T. & Wolfgang H. , HTSeq-a python framework with high-throughput sequencing data . Bioinformatics 31 , 166 –169 (2015 ).25260700 Love M. I. , Huber W. & Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2 . Genome Biol. 15 , 550 (2014 ).25516281

PMC005xxxxxx/PMC5000394.txt

==== Front Front Public HealthFront Public HealthFront. Public HealthFrontiers in Public Health2296-2565Frontiers Media S.A. 10.3389/fpubh.2016.00178Public HealthOpinionAn Old Solution for a New Problem: Antiserum against Emerging Infectious Diseases Morais Victor 1*1Department of Biotechnology, Faculty of Medicine, Institute of Hygiene, University of the Republic, Montevideo, UruguayEdited by: Joav Merrick, Ministry of Social Affairs, Israel Reviewed by: Mohammed Morad, Clalit Health Services, Israel *Correspondence: Victor Morais, vmorais@higiene.edu.uySpecialty section: This article was submitted to Immunotherapies and Vaccines, a section of the journal Frontiers in Public Health 26 8 2016 2016 4 17810 6 2016 11 8 2016 Copyright © 2016 Morais.2016MoraisThis is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.antiserumsheterologous antibodiesinfectious diseasesEbolasnake venom ==== Body Introduction Heterologous neutralizing serums or antiserums consist of neutralizing antibodies produced mainly in horses or sheep and have been effectively used for more than a century. Antiserums were born in the golden age of microbiology when, in 1890, von Behering and Kitasato showed that the serum of a diphtheria-infected animal confers immunity against the same disease on naive animals (1, 2). Four years later, antiserum was used in humans. From that point, this method has always been demonstrated to be highly effective in the treatment of both infection and envenoming. However, antiserums did not have good outcomes with respect to safety in their initial applications, causing many life-threatening side reactions (3). Currently, in many applications, heterologous serums have been replaced by other drugs, such as antibiotics or homologous serums. However, in the case of envenoming from snakebites, scorpions and arachnids, antiserums remain the only effective treatment (4). In recent applications, antiserums have demonstrated a good safety profile, with <15% of patients having mild adverse reactions and <1% having severe reactions (4–6). The only weakness antiserums have is that, like most biological products, the induced reactivity in patients generates antibodies against the antiserum (7). This weakness causes the effectiveness and safety to be compromised in successive treatments, or, in other words, heterologous serums can only be used once. Heterologous Serums as Antimicrobials In the first decades of the twentieth century, before the advent of antibiotics, heterologous serums were the best treatment choice against infectious diseases (8, 9). Many diseases were treated with heterologous serum with high effectiveness but with variable safety. For example, in 1904, a Neisseria meningitidis epidemic in New York City was controlled with a heterologous specific serum, decreasing the mortality by one-third (10). Later in the twentieth century, antiserums began to be displaced by drugs with better safety profiles, antibiotics, and vaccination. However, for the treatment of envenomation, tetanus, diphtheria, and rabies, antiserums have seen continued successful use. Currently, the treatment for tetanus and diphtheria has been changed from antiserums to homologous serums obtained from healthy human donors, but in many countries, antiserums remain the only option for such treatment. In the case of snakebite and other envenomations, the antiserum is the only effective treatment. From Now Onward For many emerging diseases, such as the Ebola virus, the risk–benefit equation for the use of antiserum appears to be highly tilted toward benefit. Additionally, it is necessary to use the antiserum only once because surviving patients demonstrate immunity after first contact. Another benefit is the low production cost, which makes this type of drug affordable for most countries (6, 11, 12). Unfortunately, while vaccination, monoclonal and homologous antibodies became the most popular solutions, antiserums had less success. Against many emerging diseases and future threats, however, antiserums could have a chance. Dixit and coworkers (6) are proposed antiserum as a possible solution to avian influenza, MERS-CoV, and viral hemorrhagic fevers. Heterologous serum also has the advantage that it can be made using recombinant proteins, which avoids the risk of manipulating the infective pathogen during the production stage. Conclusion Antiserums are an old drug with more than a century of use. Perhaps for this, most pharmaceutical companies and scientist consider this type of drug as obsolete and opt for the new generation of antibodies (monoclonal, humanized). However, antiserums still have some advantages to monoclonal antibodies, such as shorter product development time, and above all reduced costs in development and production. In summary, antiserums can be a good option for the treatment of emerging infectious diseases when other drugs are unavailable. Author Contributions The author confirms being the sole contributor of this work and approved it for publication. Conflict of Interest Statement The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. ==== Refs References 1 Chippaux JP Goyffon M . Venoms, antivenoms and immunotherapy . Toxicon (1998 ) 36 :823 –46 .10.1016/S0041-0101(97)00160-8 9663690 2 Theakston RDG Warrell DA Griffiths E Report of a WHO workshop on the standardization and control of antivenoms . Toxicon (2003 ) 41 :541 –57 .10.1016/S0041-0101(02)00393-8 12676433 3 Silverstein AM Clemens Freiherr von Pirquet: explaining immune complex disease in 1906 . Nat Immunol (2000 ) 1 :453 –5 .10.1038/77874 11101860 4 WHO . WHO Guidelines for the Production, Control and Regulation of Snake Antivenom Immunoglobulins . (2010 ). Available from: http://www.who.int/bloodproducts/snake_antivenoms/snakeantivenomguideline.pdf?ua=1 5 Otero-Patiño R Cardozo J Higashi H Núñez V Diaz A Toro M A randomized, blinded, comparative trial of one pepsin-digested and two whole IgG antivenoms for Bothrops snake bites in Uraba, Colombia . Am J Trop Med Hyg (1998 ) 58 :183 –9 .9580075 6 Dixit R Herz J Dalton R Booy R . Benefits of using heterologous polyclonal antibodies and potential applications to new and undertreated infectious pathogens . Vaccine (2016 ) 34 :1152 –61 .10.1016/j.vaccine.2016.01.016 26802604 7 Morais V Berasain P Ifrán S Carreira S Tortorella MN Negrín A Humoral immune responses to venom and antivenom of patients bitten by Bothrops snakes . Toxicon (2012 ) 59 :315 –9 .10.1016/j.toxicon.2011.12.006 22206812 8 Vital B Alguns casos de diphtheria tratados pelo serum anti-diphtherico . Rev Med São Paulo (1898 ) 1 :51 –6 . 9 Bull CG . The mechanism of the curative action of antipneumococcus serum . J Exp Med (1915 ) 22 :457 –64 .10.1084/jem.22.4.457 19867929 10 Berghman LR Abi-Ghanem D Waghela SD Ricke SC . Antibodies: an alternative for antibiotics? Poult Sci (2005 ) 84 :660 –6 .10.1093/ps/84.4.660 15844826 11 Kudoyarova-Zubavichene NM Sergeyev NN Chepurnov AA Netesov SV . Preparation and use of hyperimmune serum for prophylaxis and therapy of Ebola virus infections . J Infect Dis (1999 ) 179 (Suppl 1 ):S218 –23 .10.1086/514294 9988187 12 Chippaux JP Boyer L Alagón A . Post-exposure treatment of Ebola virus using passive immunotherapy: proposal for a new strategy . J Venom Anim Toxins Incl Trop Dis (2015 ) 21 :3 .10.1186/s40409-015-0003-1 25705218

PMC005xxxxxx/PMC5000399.txt

==== Front BMC Med EducBMC Med EducBMC Medical Education1472-6920BioMed Central London 74610.1186/s12909-016-0746-6Research ArticleEvaluation of a new community-based curriculum in disaster medicine for undergraduates Bajow Nidaa 00966506202223dr.nidaaa@gmail.com 12Djalali Ahmadreza 1Ingrassia Pier Luigi 1Ragazzoni Luca 1Ageely Hussein 3Bani Ibrahim 3Corte Francesco Della 11 CRIMEDIM - Research Centre in Emergency and Disaster Medicine and Computer Science Applied to Medical Practice, University of Eastern Piedmont, Novara, Italy 2 Disaster Medicine Unit, Mohammad Bin Naif Medical Center, King Fahd Security College, P O Box 89489, Riyadh, 11682 Saudi Arabia 3 Medical School of Jazan University, Jazan, Saudi Arabia 26 8 2016 26 8 2016 2016 16 1 2259 8 2015 18 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Nowadays, many medical schools include training in disaster medicine in undergraduate studies. This study evaluated the efficacy of a disaster medicine curriculum recently designed for Saudi Arabian medical students. Methods Participants were 15 male and 14 female students in their fourth, fifth or sixth year at Jazan University Medical School, Saudi Arabia. The course was held at the Research Center in Emergency and Disaster Medicine and Computer Sciences Applied to the Medical Practice in Novara, Italy. Results The overall mean score on a test given before the course was 41.0 % and it increased to 67.7 % on the post-test (Wilcoxon test for paired samples: z = 4.71, p < 0.0001). There were no significant differences between the mean scores of males and females, or between students in their fourth, fifth or sixth year of medical school. Conclusions These results show that this curriculum is effective for teaching disaster medicine to undergraduate medical students. Adoption of this course would help to increase the human resources available for dealing with disaster situations. Keywords CurriculumDisaster medicineSaudi Arabiaissue-copyright-statement© The Author(s) 2016 ==== Body Background The last two decades have seen a growing interest in disaster medicine [1–5]. This is understandable given the number and magnitude of natural disasters (such as the 2004 Indian Ocean earthquake and consequent tsunami, and hurricane Katrina) as well as man-made and technological disasters (such as the 9/11 bombing, the Fukushima Daiichi nuclear disaster and the ongoing series of worldwide terrorist attacks). Disaster medicine, which is pivotal in such circumstances, has been traditionally a postgraduate form of training. However, in the face of large-scale disasters, the personnel trained in disaster medicine can be overwhelmed by the demands of the disaster and other medical personnel have to pitch in. In view of the gaps in undergraduate and postgraduate disaster medicine education, [6–10] some educational authorities have called for improved disaster medicine education [11, 12] and several medical schools and professional organizations have developed curricula in disaster medicine for education of all physicians [1–4, 13–17]. A survey of all Saudi Arabian medical schools [18] showed that teaching of disaster medicine was scarce, but there was willingness to institute such training at the undergraduate level, with a preference for both didactic and interactive learning activities coupled with a combination of on-site education and distance e-learning. Consequently, a training curriculum contextualized to the local environmental and socioeconomic milieu was developed [19]. At the same time, because crises can span geographic and political regions, the program is also consistent with international educational and disaster medicine standards [20, 21]. The aim of the current study was to evaluate the efficacy of this curriculum in improving the knowledge of Saudi Arabian medical students. Methods Setting and participants Saudi medical students are routinely sent abroad during the summer for further education or training, and the community-based disaster medicine course was piloted at the Research Center in Emergency and Disaster Medicine and Computer Sciences Applied to the Medical Practice (CRIMEDIM) in Novara, Italy. The 2-week course started on 16 June 2014. The participants were 15 male and 14 female students in their fourth, fifth or sixth year at Jazan University Medical School, Saudi Arabia, which agreed to participate in this study. The students were selected on the basis of having a minimum overall grade of 3.5/5.0 and possession of good knowledge of English. The course was held at the Research Center in Emergency and Disaster Medicine (CRIMEDIM), which was established in 2007 as part of the University of Eastern Piedmont, Italy. The course was delivered by 19 instructors, of whom 12 were with University of Eastern Piedmont, Italy, specialized in anesthesia and disaster medicine (n = 7) or in disaster medicine (n = 3), anesthesia and humanitarian medicine (n = 1) or prehospital management (n = 1). The others were from Como Hospital, Italy, Venice University, Italy, Bunbury and Busselton Hospitals, West Australia, and Jazan University, Saudi Arabia, and specialized in disaster medicine, prehospital management, emergency medicine, and emergency and critical care. All the students consented to full participation before the program and had not received any instruction in disaster medicine. The study was approved by the Ethics Committee of Jazan Medical School in Saudi Arabia. Curriculum The development and constitution of the curriculum have been described [19]. The curriculum was developed through a five-stage approach by five international experts in collaboration with stakeholders from the Jazan area, and focus was placed on interactive, student-centered content. The curriculum introduces core principles in emergency medicine, public health, and disaster management using several approaches (lectures, workshops, simulations, group discussions, case studies, and role-playing) to promote higher cognitive engagement (Table 1). Five major domains were presented in this curriculum: (1) general concepts of disaster medicine; (2) disaster risk reduction; (3) disaster & mass casualty incident management; (4) principles of community awareness; (5) training sessions for community education.Table 1 The contents and schedule of the disaster medicine course Time Day 1 Day 2 Day 3 Day 4 Day 5 09:00—10.30 - Course presentation - Student and faculty presentation - Pre-test Lecture Hazard identification and risk assessment Case study Discussion of Jeddah flood 2009 Lecture Medical aspects of different disasters Lecture & GD Prehospital disaster management ISEE Familiarization & hospital disaster preparation 11:00—12:30 Lecture Introduction to disaster management Workshop Using available tools to assess risks to Jazan Lecture & GD International humanitarian law Ethics in disasters XVR Mass casualty triage ISEE Prehospital and hospital disaster response 13:30—15:00 Lecture Differences between emergency and disaster medicine, and humanitarian health Lecture, EL, GD Strategies for preventing and mitigating risks to self and to others VISIT Novara University Hospital Lecture & GD Command and control and the incident command system ISEE -After action review -Debriefing of the first week 15:30—17:00 Video lecture The emergency medical systems used in Jazan Lecture, SDE, GD General concepts of adult teaching Visit Novara EMS dispatch center XVR Disaster scene management Day 6 Day 7 Day 8 Day 9 Day 10 09:00—10.30 Video lectures - Impact & principle of community participation - Tools for community awareness & public education Lectures & GD Top ten priorities & public health assessment during humanitarian emergencies Lecture & RP Techniques to handle psychological reactions caused by disasters Lecture Infectious diseases in emergencies Lecture International coordination for disaster response 11:00—12:30 Lecture & 3 mock exercises Community education for environmental mitigation measures Lectures & CS -Humanitarian standards in context: The Sphere Project -Community education for public health surveillance Brainstorm Preparation of the pilot session for community education Material preparation Group 1 simulation of a pilot session for community education Presentation & feedback Discussion of the group pilot session 13:30—15:00 Lecture Community awareness: Role of volunteers Brainstorm Preparation of the pilot session for community education Video lecture The role of Ministry of Health & Red Crescent during population displacement in Jazan Trip Lake Lago Maggiore Material preparation Group 2 simulation of a pilot session for community education -Post-test -Debriefing & farewell 15:30—17:00 Visit Advanced medical post Material preparation Group 3 as above L lecture, VL video lecture, GD group discussion, WS workshop, RP role-play, ME mock EL experiential learning, SDE self-diagnostic exercise, ISEE Interactive Simulation for Emergencies (www.inovaria.com), EMS Emergency Medical Service, XVR virtual reality training software for education, training and assessment of incident response safety professionals (www.xvrsim.com), MOH Ministry of Health, Sphere Project http://www.sphereproject.org General concepts of disaster medicine included 13 lectures. In the introduction to disaster medicine lectures, the students received an introductory lecture on disaster medicine, including different key terms and classification and identification of the causes of disasters (man-made and natural) by two approaches: trigger events and speed onset events. In trigger events, the disaster cause can be natural or anthropic. The natural cause can be primary such as in earthquakes or secondary as in floods. Anthropic causes are divided into three subcategories: technical (as in industrial incidents and building collapses), social (as in mass gatherings) and war (conventional, or chemical, nuclear or biological bombing). The second approach is based on the speed of onset of the disaster (slow or progressive) both of them expose the students to both types of disaster (man made and natural) in the lectures of medical aspect of disaster the students focused more about direct and indirect impact of natural disasters such as floods and earthquakes. They applied the principles of disaster management cycle to the Jeddah floods. They were also introduced to the definition and level of complex humanitarian emergencies resulting from violent conflict and compared their effect with the effects of different types of natural disasters such as earthquake, floods and high winds. In disaster risk assessment, the students identified the common hazards in their area, such as floods and earthquakes, and population displacements caused by the violent conflict in Yemen, and mass casualties from road traffic accidents. They identified and applied the safety strategies that should be followed in both conventional and nonconventional (chemical hazards) mass casualties. In mass casualties incident management, the students gained an understanding of the different approaches to managing mass casualties, the concept and functions of an advance command post/team (ACP), and the concept of hospital disaster preparedness. They also demonstrated the common types of mass casualty triage. Virtual reality simulation was used to introduce the students to manmade disaster scenarios, such as building collapse and fire on boats at a seaport. In these simulations, they were asked to handle elements of the disaster response, including standard operation procedures for notification and confirmation of mass casualties, incident command system, triage, treatment, evacuation of victims to proper facilities, and activation of the hospital disaster response plan, as well as stabilization of the victims. In principles of community awareness, the students performed an exercise on public health education according to the World Health Organization strategies [22–24] and three mock community education sessions to demonstrate the principles of community awareness. These included health promotion for educators in primary schools in areas affected by earthquakes, preparation of oral rehydration solution for mothers with young children in pediatric clinics in camps, and corpse disposal during a cholera outbreak. In the last domain, training for community education, the students were oriented to the principles of adult learning to enable them to design sessions for educating communities in Saudi Arabia in disaster preparedness. After that, they were divided into groups of 3 − 5 students and asked to prepare community training activities. There were six sessions. In the first two sessions, the students chose the topics on which to prepare workshops for community education and brainstormed with instructors to define their aims, objectives, target audiences, and educational strategy. One topic they agreed on was a workshop for primary school educators about fire disasters, including the medical impact of the fire and the treatment of first and second-degree burns and suffocation. The second was on health promotion and standard precautions in primary schools during a pandemic infection, for example by coronavirus or H1N1. During the next three sessions, each group of students prepared one of the following topics: types of burns and injuries resulting from a fire disaster, the minimum first aid required to save the victims, and video lectures on activation of an evacuation plan and standard precaution procedures during a pandemic influenza. In the last session, on the last day of the program, each group of students presented their community education session and received feedback from the instructors. The post-test was administered after this session. Passing the course required a grade of ≥ 60 % in the post-test and participation in the presentation of the community training session. Evaluation of the efficacy of the course A pre-test was conducted on the first day of the course and a post-test on the last day. The questions were obtained mainly from the question test banks of CRIMEDIM. The instructors were asked to prepare questions relevant to their subjects if they could not find appropriate questions in the CRIMEDIM database. Each 30-min test consisted of 25 multiple choice questions with only one correct answer. One point was given for a correct answer and zero for a wrong answer. The questions in the two tests were different but both sets covered the four major course domains: (1) general concepts of disaster medicine, (2) disaster risk reduction, (3) mass casualty incident management, and (4) community disaster awareness. The number of questions for each domain paralleled the relative weight of the domain in the curriculum. Four experts from CRIMEDIM reviewed the pre-test and post-test. To obtain feedback from the students about the course, they were asked to fill an evaluation form at the end of the course. In addition to the nine questions (mostly Likert scale), they were asked for suggestions on how the course might be improved. We also sought to measure the third level of Kirkpatrick’s evaluation, [25] which specifies for behavior changes after training and education, by contacting the students by email after one and half years. Those who did not respond to the email were contacted by phone whenever possible. Statistical analysis Statistical analysis was done using SPSS, version 11.0 (Statistical Package for Social Science (SPSS, 22®, IBM, NY, USA). The reliability of test items was tested by calculating Cronbach’s alpha coefficients for the knowledge and practice items. The data are presented as mean percent score and standard deviation. The non-parametric Mann-Whitney test was used to compare the means of two groups, and the Kruskal Wallis test was used to compare the means for three groups. Wilcoxon Signed Rank test was used to compare pre-test with post-test mean scores. Level of significance was set at p-value ≤ 0.05. Results Cronbach’s alpha coefficients for the knowledge and practice items, calculated on the pretest, were 0.723 and 0.897, respectively, demonstrating the reliability of the test items. Of the 29 students, 34.4 % were in their fourth year, 24.1 % in their fifth year, and 41.4 % in their sixth year. Demographics of the participants (age, sex and academic year) are presented in Table 2. To assess the gain in knowledge from the course, the pre-test and post-test results were compared (Fig. 1). The overall mean score was 41.0 % ± 6.29 SD on the pre-test and 67.7 % ± 7.70 SD on the post-test (p < 0.0001). There was no significant difference between the mean scores of males and females on the pre-test or on the post-test (Fig. 1).Table 2 Distribution of participants by sex, age and academic year Academic year Male Female TOTAL n Mean age (SD) n Mean age (SD) n Mean age (SD) 4 5 24.0 (4.0) 5 22.3 (0.7) 10 23.1 (3.0) 5 3 23.5 (0.8) 4 23.3 (0.6) 7 23.4 (0.7) 6 5 24.6 (0.7) 5 23.7 (0.2) 10 24.2 (0.7) TOTAL 15 24.1 (2.4) 14 23.1 (0.8) 29 23.6 (1.9) Fig. 1 Mean percent scores of males and females in the pre-test and post-test. *Difference from relative pre-test score is significant at p < 0.001 by Wilcoxon test. No significant difference was found between males and females on the pre-test (p = 0.594) or post-test (p = 0.124). Error bars: standard deviation Given that the students were from three different years of medical school, we wondered whether the one or two years’ difference in medical school education might affect the scores. But comparison of the pre-test mean scores of students in the fourth, fifth and sixth years showed that there was no significant difference between them (39.2 % ± 6.20 SD, 43.4 % ± 7.80 SD and 41.0 % ± 5.44 SD, respectively; p = 0.317). Likewise, there was no significant difference between them on the post-test (data not shown). Finally, we looked at how the students evaluated the course (Table 3). Most of them (76 %) found it interesting and stated that their personal goals were met by the program (72.4 %). On a Likert scale measuring satisfaction with various aspects of the course (Table 2), the statement “Overall, the instructors were effective and responded to questions in an informative, appropriate and satisfactory manner” received the most positive answers (93.1 % agree or strongly agree). The statement receiving the least positive answers (58.6 %) was “The workshop was scheduled at a suitable time of year.” Many students commented that the course was appropriate and relevant to their medical education. Some students complained about some technical difficulties during videoconferencing or suggested that more time should be given for the community education session. In response to open questions concerning improvement of the curriculum, 4 of 29 students (14 %) expressed interest in an additional week of simulation sessions. In conversation after the simulation sessions, all of them expressed strong interest in XVR and ISS.Table 3 Scores of student satisfaction with course Statement % Strongly agree % Agree % Positive answers % Neutral % Disagree Overall, the pre-workshop course was appropriate and informative. 37.9 37.9 75.9 17.2 6.9 The workshop was scheduled at a suitable time of year. 44.8 13.8 58.6 17.2 20.7 Overall, the workshop facilities and location were appropriate and satisfactory. 41.4 48.3 89.7 10.3 0.0 Overall, the workshop material was presented in a clear and organized manner. 37.9 41.4 79.3 17.2 3.4 Overall, the instructors were effective and responded to questions in an informative, appropriate and satisfactory manner. 51.7 41.4 93.1 3.4 3.4 Overall, the handouts for discussion groups and case studies were clear and useful. 41.4 34.5 75.9 20.7 3.4 Overall, the workshop was informative and valuable. 34.5 48.3 82.8 17.2 0.0 We also followed up the students by email and telephone one and a half years after the course to get some understanding of their attitudes and experience with disasters after the course. Only 18 students of the 29 (62 %) could be contacted. Half of them were working as interns or residents in the Jazan area, while the rest had transferred to Riyadh, Saudi Arabia. In December 2015, a fire broke out in the maternity ward at Jazan General Hospital, killing 25 and injuring 123, and though none of these 18 students were working at that hospital, they helped prepare their hospitals and participated in receiving the victims and discharging stable cases. They said that the course had changed their attitude and that they have become less stressed and more confident when faced by emergencies at their hospitals. Overall, they said they feel that they had benefited from the course and, for example, could distinguish what was being done improperly during drill evacuations at their hospital. Discussion This study shows that the curriculum developed for Saudi Arabian medical schools is significantly effective in increasing the students’ average knowledge of disaster medicine. The statistically significant increase in overall mean score from 41.0 % on the pre-test to 67.7 % on the post-test is somewhat better than that described in a similar study, which reported that the scores increased from 39 % on the pre-test to 58 % on the post-test [26]. However, our students were selected for having above average grades and good knowledge of English. Nevertheless, this selection represents the real situation of students who would be sent for overseas training if the course is adopted by Saudi Arabian medical schools. The course was conducted at CRIMEDIM, which started international summer course programs for disaster medicine education two years ago. The performance and competences the students gained from this course motivated Jazan University to sign a memo of understanding with CRIMEDIM to include the course in one of the four international programs for medical students at Jazan University. Though male medical students are separated from female students in Saudi Arabia, they follow the same curricula and the same standards are applied. So it was not surprising that the scores of males and females were not different, either on the pre-test or the post-test. Moreover, given that disaster medicine is not taught in the undergraduate years, the results confirmed our expectation that students in their fourth, fifth or sixth year would perform similarly in the pre-test. According to the National Educational Framework for disaster health, students can be trained and educated in the basic skills needed for the response to mass casualties [13]. These include preparedness, planning, response and recovery, and understanding the roles of different organizations. The course has a community-based disaster medicine curriculum that encourages learners to apply concepts, skills and attitudes to their unique local context. This course considered the specific needs of the medical students and the community at large and consisted of an appropriate number of learning activities in a balanced variety of educational settings. The community disaster awareness sessions (considered a novel and important topic in Middle Eastern countries) was designed according to health needs assessment in Jazan and aim at providing the students with the basic skills and knowledge required for educating the community at different levels (local population and community members). In their final three years, medical students can work as volunteers. One study demonstrated that up to 96 % of medical students are willing to volunteer during disasters by helping triage and by providing first aid and community education [10]. In the 2005 earthquake in Kashmir, medical students participated in search and rescue operations and provided emergency care under supervision [27]. Importantly, students who are not properly prepared and protected can cause dangers to themselves as well, and the first issue that should be considered when involving students in emergency response is their safety. In the course, we included sufficient practical training in the use of personal protective equipment. They were also instructed in how to prepare their personal /family disaster plans and oriented in how to approach a disaster scene, particularly when the hazards are not known. Emergency first responders, whether military or civilian personnel, are given priority access to protective equipment or vaccines, and medical students who serve as volunteers should also be given such priority [28]. Because they are still in training and are newly involved in the disaster response, their occupational risk has to be reduced by giving them priority in receiving all types of protection ahead of the qualified individuals in the medical groups [28]. During disasters, healthcare workers are faced by decision-making difficulties in three main areas: triage of the patients, resource allocation, and clinical care. One study showed that triage can be learned by first year medical students with a degree of accuracy that is comparable to that of more experienced peers [29]. We expect that medical students in their fourth to sixth year of medical school can perform equally well if not better. Providing additional human resources for triage can free the specialized professionals for other duties. This curriculum was developed in the context of Jazan University’s annual international summer school program, which provides for overseas training of 340 students from different colleges. Thus, it was possible to develop a rather extensive curriculum that extends beyond implementation of knowledge for management to aspects of educating peers and the community. The program encompasses more than 53 h of education over 2 weeks of different learning activities structured around five major domains. The complex humanitarian emergency topics, the Sphere project, and techniques to handle psychological reactions caused by disasters are considered new topics in undergraduate programs aiming to help medical students to participate as volunteers in disaster humanitarian relief. The course was designed to teach knowledge and skills sequentially, in that basic principles are first presented in didactic sessions, and then competencies are strengthened by practicing skills in hands-on exercises and simulation sessions. Then, the acquired knowledge and skills are implemented by designing sessions geared towards community education and disaster management education for their colleagues, such as medical and health care students at universities. We also foresee the possibility of recruiting students from other health disciplines. Inter-professional learning with their future colleagues eventually improves the quality of care and assistance provided during a disaster [30]. For example, involvement of medical students with other health care students (nurses, laboratory technicians) helps the students to know the roles and responsibilities of other health care workers and at the same time provides them with greater understanding of teamwork and communication skills in complex patients care situations [31–33]. All the learning activities are inter-related, and this creates a framework for effective disaster medicine training [34]. It has been demonstrated that participants in courses using different teaching methods are more confident in their knowledge for at least 6 months following the training [35]. The gold standard for demonstrating the effectiveness of a curriculum in medical education is the demonstration of a change in behavior or performance in the real-world clinical setting. However, it is often more feasible to measure changes in knowledge and educational satisfaction, [17, 36, 37] which was done in this study. The students’ experience encompasses important factors such as satisfaction with teaching and perception of quality of the learning material, physical environment, and learning culture [38]. The course evaluation process helps guide future modifications of the curriculum in order to better meet the program’s objectives [39]. In addition to evaluating the course by comparing the pre-test and post-test scores, we sought feedback from students. This feedback was quite positive, and so was the less formal follow-up one and a half years after the course. This is the first interventional study that focuses on the education of medical students in Saudi Arabia by using an internationally developed, approved and delivered curriculum. On the level of the Gulf region and other Arab countries, this is the first competency-based disaster medicine curriculum focused on blended learning and simulation-based exercise, and nothing was found in the literature about similar programs in Arab countries. On the other hand, though information can be found about community based programs in the literature, [40] there is not much on providing the students with the skills, knowledge and attitude required for community awareness. One major challenge in this study was the budgetary constraint, which limited the number of participants. Further studies should enroll larger numbers of students, and expenses should be seen as an investment rather than a cost. The extent and depth of the course enables the course graduates to educate others, and in that, the course trains trainers. Establishment of appropriate courses in disaster medicine in Saudi Arabia in which course graduates train students in medicine and other health professionals will increase the capacity and competence for dealing with disasters and will generate a return on the investment. Conclusion This study demonstrates the efficacy of the proposed course in disaster medicine for Saudi Arabian medical students. This course would help to increase the size of the human resources available for dealing with disaster situations. Moreover, graduates of this course can be integrated in teaching disaster medicine to other students in medicine and other health disciplines. This course could be adapted to other countries in the region by replacing the components that have been adapted to the Saudi Arabian context with others that are relevant to the target society. Acknowledgements The authors thank Dr. Amin Bredan for editing the manuscript. Funding This work was done in the framework of a collaboration between University of Eastern Piedmont, Novara, Italy and Jazan University, Jazan, Saudi Arabia. Jazan University covered the travel and accommodation costs for the course and overhead costs. No other funding for the research was received. Availability of data and material The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Authors’ contributions NB, AD, FDC, PLI, LR, HA and IB participated in conception and design of the research. NB acquired and analyzed the data. NB, AD and FDC interpreted the results. NB wrote the manuscript. All authors read the manuscript and approved its submission to BMC Medical Education. Authors’ information Nidaa Bajow is Disaster Medicine Unit Manager at Mohammad Bin Naif Medical Center, King Fahd Security College, Saudi Arabia, and PhD candidate at CRIMEDIM - Research Centre in Emergency and Disaster Medicine and Computer Science Applied to Medical Practice, University of Eastern Piedmont, Novara, Italy Ahmadreza Djalali is Researcher and Lecturer at CRIMEDIM - Research Centre in Emergency and Disaster Medicine and Computer Science Applied to Medical Practice, University of Eastern Piedmont, Novara, Italy Pier Luigi Ingrassia is Vice Director, CRIMEDIM - Research Centre in Emergency and Disaster Medicine and Computer Science Applied to Medical Practice, University of Eastern Piedmont, Novara, Italy Luca Ragazzoni is Disaster Medicine Educationalist and Researcher and PhD Candidate at CRIMEDIM - Research Centre in Emergency and Disaster Medicine and Computer Science Applied to Medical Practice, University of Eastern Piedmont, Novara, Italy Hussein Ageely is Associate Professor of Medicine and Dean of the College of Medicine, Jazan, Saudi Arabia Ibrahim Bani is Associate Dean for Graduate Studies and Research, College of Medicine, Jazan, Saudi Arabia Francesco Della Corte is Director, CRIMEDIM - Research Centre in Emergency and Disaster Medicine and Computer Science Applied to Medical Practice, University of Eastern Piedmont, Novara, Italy Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate The study was approved by the Ethics Committee of Jazan Medical School in Saudi Arabia and all the students included in the study consented to full participation before the start of the program. ==== Refs References 1. Bradt DA Abraham K Franks RA A strategic plan for disaster medicine in Australasia Emerg Med 2003 15 3 271 82 10.1046/j.1442-2026.2003.00445.x 2. Kaji AH Coates W Fung CC A Disaster Medicine Curriculum for Medical Students Teach Learn Med 2010 22 2 116 22 10.1080/10401331003656561 20614377 3. Pfenninger EG DDB Stahl W Bauer A Houser CM Himmelseher S Medical student disaster medicine education: the development of an educational resource Int J Emerg Med 2010 3 9 20 10.1007/s12245-009-0140-9 20414376 4. Schultz CH Koenig KL Whiteside M Murray R All-Hazard NS Development of National Standardized All-Hazard Disaster Core Competencies for Acute Care Physicians, Nurses, and EMS Professionals Ann Emerg Med 2012 59 3 196 208 10.1016/j.annemergmed.2011.09.003 21982151 5. Subbarao I Lyznicki JM Hsu EB Gebbie KM Markenson D Barzansky B A Consensus-based Educational Framework and Competency Set for the Discipline of Disaster Medicine and Public Health Preparedness Disaster Med Public Health Prep 2008 2 1 57 68 10.1097/DMP.0b013e31816564af 18388659 6. Cordi H Cascardo D Disaster medicine: every physician’s second specialty. Twenty-first century fears AJCM 2011 8 141 3 7. Djalali A Ingrassia P Corte F Foletti M Gallardo A Ragazzoni L Identifying deficiencies in national and foreign medical team responses through expert opinion surveys: implications for education and training Prehosp Disaster Med 2014 29 4 364 8 10.1017/S1049023X14000600 24945852 8. Ejaz K Disaster management never part of the medical curriculum in Pakistan JPMA 2011 61 5 425 9. Ingrassia PL Foletti M Djalali A Scarone P Ragazzoni L Corte FD Education and training initiatives for crisis management in the European Union: a web-based analysis of available programs Prehosp Disaster Med 2014 29 02 115 26 10.1017/S1049023X14000235 24642198 10. Kaiser HE Barnett DJ Hsu EB Kirsch TD James JJ Subbarao I Perspectives of future physicians on disaster medicine and public health preparedness: challenges of building a capable and sustainable auxiliary medical workforce Disaster Med Public Health Prep 2009 3 4 210 6 10.1097/DMP.0b013e3181aa242a 20081417 11. Smith J Levy MJ Hsu EB Lee Levy J Disaster curricula in medical education: pilot survey Prehosp Disaster Med 2012 27 05 492 4 10.1017/S1049023X12001215 22985793 12. Swee DE. Medical Student Involvement in Disaster Medicine and Public Health Preparedness Planning and Response. New Orleans, LA: American Medical Association Proceedings of the 2011 Interim Meeting of the House of Delegates. p. 113–23. http://www.ama-assn.org/resources/doc/hod/x-pub/i11-cme-reports.pdf. Accessed 22 Aug 2016. 13. FitzGerald GJ Aitken P Arbon P Archer F Cooper D Leggat P A national framework for disaster health education in Australia Prehosp Disaster Med 2010 25 01 4 11 10.1017/S1049023X00007585 20405454 14. Franc-Law JM Ingrassia PL Ragazzoni L Della Corte F The effectiveness of training with an emergency department simulator on medical student performance in a simulated disaster CJEM 2010 12 1 27 32 10.1017/S1481803500011982 20078915 15. Ingrassia PL Ragazzoni L Tengattini M Carenzo L Della Corte F Nationwide program of education for undergraduates in the field of disaster medicine: development of a core curriculum centered on blended learning and simulation tools Prehosp Disaster Med 2014 29 05 508 15 10.1017/S1049023X14000831 25155942 16. Khanduja PK Bould MD Naik VN Hladkowicz E Boet S The Role of Simulation in Continuing Medical Education for Acute Care Physicians: A Systematic Review Crit Care Med 2015 43 1 186 93 10.1097/CCM.0000000000000672 25343571 17. Parrish AR Oliver S Jenkins D Ruscio B Green JB Colenda C A short medical school course on responding to bioterrorism and other disasters Acad Med 2005 80 9 820 3 10.1097/00001888-200509000-00007 16123460 18. Bajow N Djalali A Ingrassia PL Ageely H Bani I Della Corte F Disaster medicine curricula in Saudi Arabian medical schools JEMTAC 2015 1 8 19. Bajow N Djalali A Ingrassia PL Ageely H Bani I Della Corte F Proposal for a community-based disaster management curriculum for medical school undergraduates in Saudi Arabia Am J Disaster Med 2015 10 2 145 52 10.5055/ajdm.2015.0197 26312495 20. Burkle FM The Development of Multidisciplinary Core Competencies: The First Step in the Professionalization of Disaster Medicine and Public Health Preparedness on a Global Scale Disaster Med Public 2012 6 1 10 2 10.1001/dmp.2012.3 21. Murray V Clifford J Seynaeve G Fisher JM Disaster health education and training: a pilot questionnaire to understand current status Prehosp Disaster Med 2006 21 3 156 67 10.1017/S1049023X00003629 16892880 22. World Health Organization - Western Pacific Region 2008. A guide to establishing event-based surveillance. Available at http://www.wpro.who.int/emerging_diseases/documents/docs/eventbasedsurv.pdf Accessed 15 Apr 2016. 23. World Health Organization - Regional Office for Africa 2015. Integrated Diseases Surveillance and Response in the African Region: Community-based Surveillance (CBS) Training Manual. Available at http://www.afro.who.int/en/clusters-a-programmes/dpc/integrated-disease-surveillance/ids-publications/4837-integrated-diseases-surveillance-and-responsenin-the-african-region-community-based-surveillance-cbs-training-manual.html Accessed 15 Apr 2016. 24. World Health Organization 2014. Disease surveillance and response programme area - Disease prevention and control cluster. Available at http://www.afro.who.int/en/clusters-a-programmes/dpc/integrated-disease-surveillance/ids-publications/4326-integrated-disease-surveillance-and-response-in-the-african-region-a-guide-for-establishing-community-based-surveillance.html Accessed 13 Apr 2016. 25. Kirkpatrick DL, Kirkpatrick JD, Kirkpatrick, WK 2009. The Kirkpatrick model. Available at http://www.kirkpatrickpartners.com/OurPhilosophy/TheKirkpatrickModel Accessed Apr 15 2016 26. Scott LA Carson DS Greenwell IB Disaster 101: a novel approach to disaster medicine training for health professionals J Emerg Med 2010 39 220 6 10.1016/j.jemermed.2009.08.064 20079997 27. Sabri AA Qayyum MA Why medical students should be trained in disaster management: Our experience of the Kashmir earthquake Plos Med 2006 3 9 1452 3 10.1371/journal.pmed.0030382 28. Lim EC Oh VM Koh DR Seet RC The challenges of “continuing medical education” in a pandemic era Ann Acad Med Singapore 2009 38 8 724 6 19736579 29. Sapp RF Brice JH Myers JB Hinchey P Triage performance of first-year medical students using a multiple-casualty scenario, paper exercise Prehosp Disaster Med 2010 25 239 45 10.1017/S1049023X00008104 20586018 30. Atack L Parker K Rocchi M Maher J Dryden T The impact of an online interprofessional course in disaster management competency and attitude towards interprofessional learning J Interprof Care 2009 23 6 586 98 10.3109/13561820902886238 19842952 31. Kaplan BG Connor A Ferranti EP Holmes L Spencer L Use of an emergency preparedness disaster simulation with undergraduate nursing students Public Health Nurs 2012 29 1 44 51 10.1111/j.1525-1446.2011.00960.x 22211751 32. Rega P Interdisciplinary simulation education in disaster medicine Public Health Nurs 2013 30 1 3 4 10.1111/phn.12004 23294381 33. Zapko KA Ferranto ML Brady C Corbisello A Hill D Mullen R Interdisciplinary disaster drill simulation: Laying the groundwork for further research Nurs Educ Perspect 2015 36 6 379 382 10.5480/14-1544 26753297 34. Takemoto K Motoya Y Kimura R A proposal for effective emergency training and exercise program to improve competence for disaster response of disaster responders JDR 2010 5 2 197 207 10.20965/jdr.2010.p0197 35. Cicero MX Blake E Gallant N Chen L Esposito L Guerrero M Impact of an educational intervention on residents’ knowledge of pediatric disaster medicine Pediatr Emerg Care 2009 25 7 447 51 10.1097/PEC.0b013e3181ab78af 19564807 36. Coico R Kachur E Lima V Guidelines for Pre-Clerkship Bioterrorism Curricula Acad Med 2004 79 366 75 10.1097/00001888-200404000-00019 15044173 37. Wakasugi M Nilsson H Hornwall J Vikstrom T Ruter A Can performance indicators be used for pedagogic purposes in disaster medicine training Scand J Trauma Resusc Emerg Med 2009 17 15 10.1186/1757-7241-17-15 19292895 38. Gibson KA Boyle P Black DA Cunningham M Grimm MC McNeil HP Enhancing evaluation in an undergraduate medical education program Acad Med 2008 83 787 93 10.1097/ACM.0b013e31817eb8ab 18667897 39. Kern E Thomas PA Hughes MT Curriculum development for medical education: a six-step approach 2009 Second Baltimore & London Johns Hopkins University Press 40. Adams L, Canclini. S: Disaster Readiness: A Community - University Partnership. Online Journal of Issues in Nursing 2008, 13 (3): Available at http://nursingworld.org/MainMenuCategories/ANAMarketplace/ANAPeriodicals/OJIN/TableofContents/vol132008/No3Sept08/ArticlePreviousTopic/DisasterReadiness.html.

PMC005xxxxxx/PMC5000400.txt

==== Front J NeuroinflammationJ NeuroinflammationJournal of Neuroinflammation1742-2094BioMed Central London 67410.1186/s12974-016-0674-8ResearchFractalkine suppression during hepatic encephalopathy promotes neuroinflammation in mice McMillin Matthew mcmillin@medicine.tamhsc.edu 12Grant Stephanie sgrant@medicine.tamhsc.edu 12Frampton Gabriel gframpton@medicine.tamhsc.edu 12Andry Sarah sandry@sw.org 3Brown Adam adabrown@sw.org 3DeMorrow Sharon demorrow@medicine.tamhsc.edu 121 Department of Internal Medicine, Texas A&M Health Science Center, College of Medicine, Temple, TX USA 2 Central Texas Veterans Healthcare System, 1901 S. 1st Street, Building 205, Temple, TX 76504 USA 3 Department of Internal Medicine, Baylor Scott & White Health, 2401 S. 31st Street, Temple, TX 76508 USA 26 8 2016 26 8 2016 2016 13 1 19824 12 2015 17 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Acute liver failure is associated with numerous systemic consequences including neurological dysfunction, termed hepatic encephalopathy, which contributes to mortality and is a challenge to manage in the clinic. During hepatic encephalopathy, microglia activation and neuroinflammation occur due to dysregulated cell signaling and an increase of toxic metabolites in the brain. Fractalkine is a chemokine that is expressed primarily in neurons and through signaling with its receptor CX3CR1 on microglia, leads to microglia remaining in a quiescent state. Fractalkine is often suppressed during neuropathies that are characterized by neuroinflammation. However, the expression and subsequent role of fractalkine on microglia activation and the pathogenesis of hepatic encephalopathy due to acute liver failure is unknown. Methods Hepatic encephalopathy was induced in mice via injection of azoxymethane (AOM) or saline for controls. Subsets of these mice were implanted with osmotic minipumps that infused soluble fractalkine or saline into the lateral ventricle of the brain. Neurological decline and the latency to coma were recorded in these mice, and brain, serum, and liver samples were collected. Neurons or microglia were isolated from whole brain samples using immunoprecipitation. Liver damage was assessed using hematoxylin and eosin staining and by measuring serum liver enzyme concentrations. Fractalkine and CX3CR1 expression were assessed by real-time PCR, and proinflammatory cytokine expression was assessed using ELISA assays. Results Following AOM administration, fractalkine expression is suppressed in the cortex and in isolated neurons compared to vehicle-treated mice. CX3CR1 is suppressed in isolated microglia from AOM-treated mice. Soluble fractalkine infusion into the brain significantly reduced neurological decline in AOM-treated mice compared to saline-infused AOM-treated mice. Infusion of soluble fractalkine into AOM-treated mice reduced liver damage, lessened microglia activation, and suppressed expression of chemokine ligand 2, interleukin-6, and tumor necrosis factor alpha compared to saline-infused mice. Conclusions These findings suggest that fractalkine-mediated signaling is suppressed in the brain following the development of hepatic encephalopathy. Supplementation of AOM-treated mice with soluble fractalkine led to improved outcomes, which identifies this pathway as a possible therapeutic target for the management of hepatic encephalopathy following acute liver injury. Keywords Acute liver failureAzoxymethaneCX3CL1CX3CR1CCL2http://dx.doi.org/10.13039/100000062National Institute of Diabetes and Digestive and Kidney DiseasesDK082435DeMorrow Sharon http://dx.doi.org/10.13039/100000738U.S. Department of Veterans AffairsBX002638-01DeMorrow Sharon Baylor Scott & White Health Research Mentorship award150156150408Andry Sarah Brown Adam issue-copyright-statement© The Author(s) 2016 ==== Body Background Acute liver failure in Europe and the USA is caused primarily by acetaminophen-induced drug toxicity but can also be the result of viral hepatitis, toxin ingestion, non-acetaminophen-induced drug toxicity, and numerous other rare causative factors [1]. Following the development of acute liver failure, a variety of extrahepatic complications can develop including renal, respiratory, immunological, and neurological dysfunction. Of these complications, neurological dysfunction accounts for around 25 % of the mortality associated with this disease state and remains a significant challenge for clinical management of this disease [2]. These neurological complications are termed hepatic encephalopathy and have a variety of neuropsychiatric presentations from asymptomatic to severe cognitive decline and coma with implications on liver transplant priority, patient quality of life, and survival [3, 4]. A growing body of evidence implicates neuroinflammation and microglia activation involvement with the development of hepatic encephalopathy. Elevations of the proinflammatory cytokines interleukin (IL)-1β, IL-6, and tumor necrosis factor alpha (TNFα) are observed in both patients and in rodent models of acute liver failure [5, 6]. During hepatic encephalopathy, the use of anti-inflammatory therapeutics, such as minocycline, has been shown to reduce neuroinflammation and neurological dysfunction [7]. The production of pathogenic proinflammatory cytokines in the brain is a consequence of microglia activation, which can occur via a variety of signals including signal transduction by chemokines. We have previously shown that chemokine ligand 2 (CCL2) is one pathogenic chemotactic cytokine that is elevated in neurons during hepatic encephalopathy due to acute liver failure and contributes to both microglia activation and the elevation of IL-1β and IL-6 [8]. The receptors for CCL2 and other chemokines are primarily localized to microglia in the brain and therefore chemokine signaling is important in initiating the activation of microglia during states of neuroinflammation. Together, this implicates neuroinflammation and chemokine-signaling pathways as causative factors in the progression of hepatic encephalopathy. Fractalkine, also known as CX3CL1, is the only member of the CX3C class of chemokines. In the brain, fractalkine is highly expressed in neurons but can be induced in astrocytes if they are treated with TNFα or interferon-γ [9, 10]. Fractalkine in its native form is a transmembrane protein that can be cleaved by cathepsin S or a disintegrin and metalloproteinase proteins (ADAMs) [11, 12]. Upon cleavage, fractalkine is released in its soluble form that contains the extracellular N-terminal chemokine domain. Soluble fractalkine can then bind its receptor, CX3CR1, which is a G-protein coupled receptor that is localized to microglia and is thought to keep microglia in a quiescent state [9, 13]. Fractalkine/CX3CR1 signaling has been demonstrated in numerous studies to dampen neuroinflammation, with fractalkine signaling being shown to reduce IL-1β secretion in lipopolysaccharide-treated microglia [14]. This has also been shown in vivo as rats with cerebral ischemia administered CX3CR1 siRNA were found to have increased microglia activation and proinflammatory cytokine expression compared to ischemic controls [15]. In addition to this, in a variety of inflammatory conditions there is a reciprocal relationship between fractalkine and CCL2, though whether these chemokines have a direct effect on each other or are inversely affected by an extraneous signal is unknown [16, 17]. At this time, there have been no studies investigating fractalkine in the context of neuroinflammation associated with hepatic encephalopathy. Therefore, the aims of this study were to assess the expression of fractalkine and its interaction with inflammatory signaling in a murine model of hepatic encephalopathy due to acute liver failure and to determine how fractalkine contributes to the neurological complications associated with this disease state. Methods Materials Fractalkine antibodies were purchased from Abcam (Cambridge, MA). Antibodies against IBA1 were purchased from Wako Chemicals USA (Richmond, VA). NeuN antibodies were ordered from Millipore (Billerica, MA). Antibodies against CD11b and glutamate aspartate transporter (GLAST) were purchased from Miltenyi Biotec (San Diego, CA). Phosphorylated and total extracellular signal-regulated kinase 1/2 (pERK1/2 and tERK1/2) antibodies were purchased from Cell Signaling (Danvers, MA). Mouse soluble fractalkine was purchased from R&D Systems (Minneapolis, MN). Mouse fracktalkine, CCL2, IL-6, and TNFα ELISAs were purchased from R&D Systems. All real-time PCR (RT-PCR) primers were purchased from SABiosciences (Frederick, MD). Total bilirubin was assayed using an ELISA kit from Cusabio (Wuha, China). All other chemicals and reagents were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise noted and were of the highest grade available. Mouse model of hepatic encephalopathy Male C57Bl/6 mice (20–25 g; Charles River Laboratories, Wilmington, MA) were used in all in vivo experiments using methodology described previously [8, 18, 19]. All animal experiments were performed following approval from Baylor Scott & White Health Institutional Animal Care and Use Committee (Temple, TX). Mice were provided free access to water and rodent chow and were housed in constant temperature, humidity, and 12-h light-dark cycling. Acute liver failure was induced via a single intraperitoneal injection of 100 mg/kg of azoxymethane (AOM) as previously described [18, 19]. In parallel, soluble fractalkine was administered to the brain via intracerebroventricular (ICV) infusion (200 ng/day) for 3 days prior to injection of AOM as previously described [18]. This dose was selected to maintain concentrations of fractalkine in the cerebrospinal fluid at concentrations that are at or above those reported by other researchers [20]. Care was taken to minimize the effects of anesthesia on hepatic metabolism in ICV mice by the following: (1) all mice were anesthetized with isoflurane to reach surgical plane with the subsequent surgeries taking approximately 30 min to complete per mouse and (2) AOM injection was delayed for 3 days after the ICV surgery to allow for residual anesthesia to be eliminated from the body. After AOM injection, mice were placed on heating pads adjusted to 37 °C and monitored for signs of neurological decline. To reduce the impact of hypoglycemia and dehydration, cage floors were supplied with hydrogel and rodent chow and after 12 h and every subsequent 4 h, mice were injected subcutaneously with 5 % dextrose in 250 μl of saline. If mice underwent a 20 % weight loss or greater, they were removed from the study. At 12 h following injection (and every 2 h thereafter), body temperature, weight, and neurological assessments were measured. Neurological functioning was assessed by measuring the pinna reflex, corneal reflex, tail flexion reflex, escape response reflex, righting reflex, and ataxia which were scored on a scale of 0 (no reflex) to 2 (intact reflex). The neurological score at each time point was defined as the summation of these reflex scores. In addition, time to coma (defined as a loss of all reflexes) was recorded. Tissue was flash frozen and collected at coma for further analyses. Mice used for histochemical studies were transcardially perfused with PBS followed by 4 % paraformaldehyde. Whole brains were removed and placed into 4 % paraformaldehyde for 24 h, after which they were moved to a 30 % sucrose solution for cryoprotection. Brains were frozen and sectioned using a cryostat for immunofluorescence imaging. Cell isolation and flow cytometry Fresh whole brains from adult C57Bl/6 mice were homogenized using an automated homogenizer from Miltenyi Biotec using solutions provided in the Neural Tissue Dissociation Kit (Miltenyi Biotec). Following dissociation into a single-cell suspension, cells were passed through LS columns (Miltenyi Biotec) containing beads coated with CD11b antibodies (to isolate microglia) or GLAST antibodies (to isolate astrocytes) localized to the columns. The remaining cells not bound to the columns were kept as the neuron-enriched fraction. LS columns were washed to remove the CD11b-bound cells or the GLAST-bound cells. Cells were subsequently spun down and homogenized for subsequent experiments. Liver histology and biochemistry Paraffin-embedded livers were cut into 3-μm sections and mounted onto positively charged slides (VWR, Radnor, PA). Slides were deparaffinized and stained with Hematoxylin QS (Vector Laboratories, Burlingame, CA) for 1 min, followed by staining for 1 min with eosin Y (Amresco, Solon, OH), and rinsed in 95 % ethanol. The slides were then dipped into 100 % ethanol and subsequently through two xylene washes. Coverslips were mounted onto the slides using Vectamount mounting media (Vector Laboratories). The slides were viewed and imaged using an Olympus BX40 microscope with an Olympus DP25 imaging system (Olympus, Center Valley, PA). Serum alanine aminotransferasase (ALT) and bilirubin were assessed using commercially available kits. Alanine aminotransferase measurement was performed using a fluorimetric activity assay. Total bilirubin was assayed using a total bilirubin ELISA. All assays and subsequent analyses were performed according to the instructions provided by the manufacturers. Real-time PCR RNA was extracted from flash frozen tissue or isolated neural cells, and RT-PCR was performed as previously described using commercially available primers designed against mouse fractalkine, CX3CR1, and glyceraldehyde 3-phosphate dehydrogenase [21]. A ΔΔCT analysis was performed using vehicle-treated tissue or isolated cortical neurons/microglia as controls for subsequent experiments [22, 23]. Data for all experiments are expressed as mean relative mRNA levels ± SEM. Immunofluorescence Free-floating immunostaining was performed on brain sections using anti-IBA1 immunoreactivity to detect morphology and relative staining of microglia. In addition, fractalkine and NeuN immunostaining were performed. Immunoreactivity was visualized using fluorescent secondary antibodies labeled with Dylight 488 or Cy3 and counterstained with ProLong© Gold Antifade Reagent containing 4′,6-diamidino-2-phenylindole (DAPI). Slides were viewed and imaged using a Leica Microsystems TCS SP5-X inverted confocal microscope (Buffalo Grove, IL). Quantification of IBA-1 field staining was performed by converting images to grayscale, inverting their color and quantifying the number of stained pixels with ImageJ software. The values for the vehicle group were normalized to 1 with this reference being used for all other quantifications of IBA1 field staining. Z-stack images were created by performing a 30-μm merge of images through the entire brain section using ImageJ software. Cytokine and chemokine ELISAs Cortex tissue from all treatment groups was homogenized using a Miltenyi Biotec gentleMACS™ Dissociator, and total protein was quantified using a Pierce™ BCA Protein Assay kit (ThermoFisher Scientific, Waltham, MA). Capture antibodies against CCL2, IL-6, TNFα, or fractalkine were incubated overnight in 96-well plates. Each ELISA was performed according to the instructions provided from R&D Systems, and the total input protein for each sample was 50 μg. Absorbance was read using a SpectraMax® M5 plate reader from Molecular Devices (Sunnyvale, CA). Data are reported as CCL2, IL-6, TNFα, or fractalkine concentration per milligram of total lysate protein. Immunoblotting Immunoblots were performed as previously described [8] with minor modifications. For western blots, 10 % sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels were loaded with 20 μg of protein diluted in Laemmli buffer. Specific primary antibodies against pERK1/2 and tERK1/2 were used along with appropriate fluorescent secondary antibodies (LI-COR, Lincoln, NE). All imaging was performed using an Odyssey 9120 Infrared Imaging System (LI-COR). Data are expressed as fold change in fluorescent band intensity of pERK1/2 divided by tERK1/2. The values of the vehicle ICV saline group was used as a baseline and set to a relative protein expression value of 1. All band intensity quantifications were performed using ImageJ software (National Institutes of Health, Bethesda, MD). Data for all experiments were expressed as mean relative protein ± SEM (n = 4). Statistical analysis All statistical analyses were performed using Graphpad Prism software (Graphpad Software, La Jolla, CA). Results were expressed as mean ± SEM. For data passing normality tests, significance was established using the Student’s t test when differences between two groups were analyzed, and analysis of variance when differences between three or more groups were compared, followed by the appropriate post hoc test. For neurological score analyses, two-way analysis of variance analysis and a subsequent Bonferroni post hoc test were performed. If tests for normality failed, two groups were compared with a Mann-Whitney U test or a Kruskal-Wallis ranked analysis when more than two groups were analyzed. Differences were considered significant when the p value was less than 0.05. Results Neuronal fractalkine was downregulated following AOM administration C57Bl/6 mice were administered the hepatotoxin AOM to induce acute liver failure and the development of hepatic encephalopathy. The mouse AOM model of hepatic encephalopathy is characterized by a prodromal phase that transitions into mild cognitive deficits and quickly progresses to coma [24]. Fractalkine mRNA expression was downregulated in AOM-treated mice compared to time-matched vehicle controls prior to the development of neurological decline (pre), when minor neurological decline was evident (minor) and when major neurological decline was present (major) but had expression similar to controls when the mice reach coma (Fig. 1a). The expression of fractalkine was not altered in the vehicle groups over the time period studied (data not shown); therefore, all vehicle groups used in the remainder of this study were time-matched to coma AOM-treated mice. The suppression of fractalkine mRNA expression observed in AOM-treated mice correlated with a parallel reduction in fractalkine protein in cortex lysates throughout AOM-induced neurological decline (Fig. 1b). The expression of fractalkine has been reported in multiple cell types in the brain, and therefore, immunofluorescence for fractalkine was performed. This staining demonstrated that the highest expression of fractalkine was in neurons as a high degree of co-staining was found with the neuron marker NeuN (Fig. 1c). In order to determine if AOM treatment led to reduced production of neuronal fractalkine, neurons were isolated from the cortex of vehicle and AOM-treated C57Bl/6 mice using immunoprecipitation. We have previously utilized this technique and found that the neuron-enriched cell fraction had 99.7 % of the cells stain positive for CD90, which indicates that this methodology provides a relatively pure neuron cell fraction from whole brain homogenates [19]. Fractalkine mRNA expression was significantly downregulated in isolated neurons from AOM-treated mice compared to neurons isolated from vehicle-treated mice (Fig. 1d).Fig. 1 Neuronal fractalkine was downregulated following AOM administration. a Fractalkine mRNA expression in the cortex during the timecourse of AOM-induced HE with appropriate time-matched tissue analyses from vehicle-treated controls. b Fractalkine concentrations in cortex lysate normalized to protein concentrations of cortex lysates from vehicle and AOM-treated timecourse mice. c Immunofluorescence for fractalkine (green) and NeuN (red) with DAPI (blue) used as a nuclear counterstain in vehicle cortex brain tissue. d Fractalkine mRNA expression in neurons isolated from vehicle-treated and AOM-treated mice cortices. For mRNA expression and fractalkine ELISA analyses, *p < 0.05 compared to vehicle-treated mice or neurons from vehicle-treated mice, **p < 0.01 compared to vehicle-treated mice. n = 3 for fractalkine mRNA and ELISA analyses CX3CR1 was downregulated in microglia following AOM-induced hepatic encephalopathy Fractalkine transduces its signal through binding CX3CR1, which has been shown in previous studies to be expressed primarily by microglia in the brain [9]. To validate these studies, the mRNA expression of CX3CR1 was measured in neurons, astrocytes, and microglia isolated from vehicle-treated mice. CX3CR1 mRNA expression was significantly higher in microglia compared to either neurons or astrocytes (Fig. 2a). As microglia are the principal cell for CX3CR1 mRNA expression in the brain, microglia were isolated from vehicle and AOM-treated mice cortices at coma using immunoaffinity isolation. This methodology had previously led to the isolation of a microglia cell fraction where 99.3 % of the cells stained positive for CD11b [19]. CX3CR1 mRNA expression was found to be significantly suppressed in isolated microglia from AOM-treated mice when compared to isolated microglia from vehicle-treated mice (Fig. 2b).Fig. 2 The fractalkine receptor CX3CR1 was downregulated in microglia following AOM-induced hepatic encephalopathy. a CX3CR1 mRNA expression in isolated neurons, astrocytes, and microglia from vehicle-treated mice with values normalized to CX3CR1 mRNA expression in vehicle neurons. b CX3CR1 mRNA expression in microglia isolated from the cortex of vehicle and AOM-treated mice. For mRNA analyses, *p < 0.05 compared to vehicle neurons (a), or microglia from vehicle-treated mice (b), n = 3 for CX3CR1 mRNA analyses Central infusion of soluble fractalkine improved outcomes of AOM-treated mice Due to the downregulation of both fractalkine and CX3CR1 during hepatic encephalopathy, supplementation of fractalkine could be used to induce this signaling pathway to help alleviate the pathology associated with this disease state. ICV infusion of soluble fractalkine was found to slow the neurological decline that is present during AOM-induced hepatic encephalopathy with significant differences between soluble fractalkine-infused and saline-infused AOM-treated mice particularly at 20 and 22 h following AOM injection (Fig. 3a). Furthermore, soluble fractalkine-infused mice treated with AOM took significantly longer to reach coma compared to AOM-treated mice infused with saline (Fig. 3b). To see if the neuroprotective protective effect of soluble fractalkine was due to protective effects on the liver, assessments of overall liver damage and function were performed. Hematoxylin and eosin staining in liver sections from vehicle and AOM-treated mice infused with fractalkine or saline indicates that the hepatic damage in AOM-treated mice infused with soluble fractalkine or saline is similar with significant hepatic necrosis being present in both groups (Fig. 3c). In order to better assess overall liver function, bilirubin (Fig. 3d) and ALT (Fig. 3e) concentrations were measured in serum from vehicle and AOM-treated mice infused with fractalkine or saline. The concentrations of both bilirubin and ALT in AOM-treated mice infused with saline were significantly higher than vehicle. Interestingly, bilirubin and ALT levels were significantly reduced in AOM-treated mice infused with soluble fractalkine when compared to AOM-treated mice infused with saline indicating that that ICV infusion of soluble fractalkine may improve liver function. However, both AOM-treated groups were still significantly higher than the vehicle-treated groups indicating that liver damage was still present.Fig. 3 Soluble fractalkine ICV infusion improved outcomes of AOM-treated mice. a Neurological decline in AOM-treated mice infused ICV with either saline or soluble fractalkine (sFKN). The neurological score is a summation of assessments for five reflexes and ataxia that indicate a worse neurological state with a lower score, which is displayed in hours post AOM injection. b Time to coma in hours for AOM-treated mice infused with either saline or sFKN. c Hematoxylin and eosin staining in liver sections from vehicle or AOM-treated mice infused ICV with saline or sFKN. d Total serum bilirubin concentrations in vehicle or AOM-treated mice infused with saline or sFKN. e Serum ALT concentrations in vehicle or AOM-treated mice infused ICV with saline or sFKN. For neurological score and time to coma analyses, *p < 0.05 compared to AOM-treated mice infused with saline. For ALT and bilirubin assays, *p < 0.05 compared to vehicle-treated ICV saline mice, # p < 0.05 compared to AOM-treated ICV saline mice. n = 4 for all analyses Microglia activation in AOM-treated mice was reduced by ICV infusion of soluble fractalkine Determination that soluble fractalkine infusion was generating an increase in CX3CR1-mediated signaling in the brain was accomplished by performing immunoblots for pERK1/2 and tERK1/2 as these proteins are downstream effectors of CX3CR1 signaling [25]. In the vehicle-treated mice, there was no observable difference in cortical pERK1/2 expression but pERK1/2 was reduced in AOM-treated ICV saline-infused mice compared to the vehicle-treated groups and this reduction of pERK1/2 expression was reversed by ICV soluble fractalkine infusion (Fig. 4a). As fractalkine has been reported to play a role in neuroinflammation, microglia activation was investigated. Microglia activation and proliferation has been reported in a variety of neuroinflammatory conditions to include Alzheimer’s disease and excitotoxic neuronal injury [26, 27]. One method to assess microglia activation is to determine if these cells undergo a phenotypic shift from a ramified resting state to an amoeboid active proinflammatory state. In order to examine if soluble fractalkine infusion reduces microglia activation, microglia morphology was assessed in the cortex of vehicle and AOM-treated mice that were infused with either soluble fractalkine or saline. Morphological analysis of microglia using the marker IBA-1 demonstrated that microglia in AOM-treated mice infused with saline have a larger cell body and retracted processes when compared to both vehicle-treated groups or AOM-treated mice infused with soluble fractalkine (Fig. 4b). This suggests that soluble fractalkine infusion reduced the activation state of microglia and allowed for microglia to remain in a quiescent state. In addition to phenotypic changes in microglia, neuroinflammation is also characterized by proliferation of microglia. IBA1 field staining was increased in the cortex of AOM-treated mice infused with saline, which was reduced in AOM-treated mice infused with soluble fractalkine (Fig. 4c). These data support that infusion of soluble fractalkine significantly reduced IBA1 staining in AOM-treated cortex when compared to saline-infused mice (Fig. 4d). Together, these data support that soluble fractalkine infusion reduced microglia activation and proliferation that occurs during hepatic encephalopathy.Fig. 4 Microglia activation in AOM-treated mice was reduced by ICV infusion of soluble fractalkine. a Representative immunoblot images and quantification of pERK1/2 and tERK1/2 in cortex lysates from vehicle or AOM-treated mice infused ICV with saline or soluble fractalkine (sFKN). b Z-stack confocal images of individual microglia stained with IBA1 (red). Microglia when activated become more amoeboid shaped with retracted cellular processes. c Field staining for IBA1 (red) in the cortex of vehicle or AOM-treated mice infused ICV with saline or sFKN. d Quantification of IBA1 field staining in the cortex of vehicle or AOM-treated mice infused with saline or sFKN. For quantitative IBA1 immunofluorescence analyses, *p < 0.05 compared to vehicle-treated mice infused with saline, # p < 0.05 compared to AOM-treated mice infused with saline. n = 4 for immunoblot analyses and n = 10 for IBA1 field quantification analysis Elevated cytokine expression observed during hepatic encephalopathy was reduced via central fractalkine infusion Activation of microglia can lead to a variety of responses with the best characterized being the release of proinflammatory cytokines to promote neuroinflammation. CCL2 is a proinflammatory chemokine involved with the activation of microglia and has been demonstrated to play a significant role in AOM-induced hepatic encephalopathy [8]. Similar to previous studies [8], CCL2 protein expression was increased in the cortex after AOM treatment (Fig. 5a). However, pretreatment with soluble fractalkine attenuated this cortical upregulation of CCL2 expression (Fig. 5a). Downstream of chemokines are proinflammatory cytokines with both IL-6 and TNFα playing a significant role in neuroinflammation. Both IL-6 (Fig. 5b) and TNFα (Fig. 5c) were found to be elevated in the cortex of AOM-treated mice infused with saline, but expression of both cytokines was reduced following soluble fractalkine ICV infusion in AOM-treated mice. These data support the concept that soluble fractalkine ICV infusion reduces the expression of proinflammatory chemokines and cytokines during AOM-induced HE.Fig. 5 Elevated cytokine expression observed during HE can be reduced via soluble fractalkine infusion. a CCL2 concentrations in the cortex of vehicle or AOM-treated mice infused with saline or soluble fractalkine (sFKN) normalized for total lysate protein concentration. b Cortical IL-6 levels in vehicle or AOM-treated mice infused with saline or sFKN normalized for total lysate protein concentration. c TNFα concentrations in the cortex of vehicle or AOM-treated mice infused with saline or sFKN normalized for total lysate protein concentration. For ELISA analyses *p < 0.05 compared to vehicle-treated mice infused with saline. n = 3 for all ELISA analyses Discussion Hepatic encephalopathy is a devastating neurological complication of liver disease with few effective treatments; therefore, the need to identify novel therapeutic targets is important [28]. Taking this approach, the current study indicates the role that fractalkine signaling plays in the progression of hepatic encephalopathy through modulating microglia activation and subsequent neuroinflammation. The results from this study support that (1) fractalkine and its receptor are downregulated during AOM-induced neurological decline and (2) supplementing fractalkine levels directly in the brain reduces neurological decline, microglia activation, and the expression of proinflammatory cytokines during AOM-induced hepatic encephalopathy. These findings suggest that increasing fractalkine concentrations, which are suppressed during hepatic encephalopathy, significantly reduces inflammation and indicates that fractalkine supplementation may be a potential treatment strategy for the management of hepatic encephalopathy. Fractalkine is a chemokine that is expressed in neurons in normal physiological conditions, and its suppression during neuroinflammatory states contributes to the elevation of proinflammatory cytokines. Indeed, fractalkine has been demonstrated to reduce both neurotoxicity and the activation of microglia in the 6-hydroxydopamine rat model of Parkinson’s disease [29]. The results from the current study indicate that during AOM-induced hepatic encephalopathy, fractalkine is downregulated in the cortex as an early event, and its expression decreases throughout the duration of this model until coma is reached. This positions fractalkine suppression as an early event in the pathogenesis of hepatic encephalopathy. One of the only other studies investigating fractalkine expression in the brain during liver pathology found that rats with partial portal vein ligation had unchanged fractalkine expression compared to control rats, though the expression of other chemokines and their receptors was significantly altered [30]. That being said, this current report is the first to investigate fractalkine signaling in overt hepatic encephalopathy due to acute liver failure and therefore, gaining understanding into fractalkine expression and signaling during other models, and patients with hepatic encephalopathy will be required to fully understand the role of this chemokine during this disease state. In order for fractalkine to generate downstream signaling, this ligand must bind CX3CR1. As microglia are the primary source of CX3CR1, the gene expression of this receptor was assessed in isolated microglia from vehicle and AOM-treated mice. CX3CR1 mRNA was downregulated in microglia isolated from the cortices of AOM-treated mice when compared to control-treated mice. Because of the technical difficulty and low yield in isolating microglia from adult mouse brains, analysis of CX3CR1 protein was not possible in this study. Furthermore, fractalkine, via the activation of CX3CR1 leads to the downstream phosphorylation of ERK1/2. Our data support a suppression of fractalkine-induced ERK1/2 activation after AOM. Taken together, our data indicate that the fractalkine/CX3CR1 signaling axis is suppressed as an early event during AOM-induced hepatic encephalopathy that persists until coma. A similar suppression of fractalkine signaling has been reported during various other neuropathies. For example, during neuroinflammation associated with scrapie infection in hamsters, microglia activation occurs and downregulation of fractalkine and its receptor is observed [31]. Similarly, in mice with genetic knockout of CX3CR1, there is reduced expression of IL-1β and TNFα as well as reduced infarct size following focal cerebral ischemia compared to wild-type mice [32]. These effects have also been observed in the clinic, as lower plasma fractalkine concentrations have been observed in patients who have worse 6-month outcomes following stroke [33]. Similar findings regarding fractalkine were observed in the current study as ICV infusion of soluble fractalkine slowed the neurological decline and reduced the elevated concentrations CCL2, IL-1β, and TNFα that are associated with AOM-induced hepatic encephalopathy. That being said, fractalkine appears to play an immunomodulatory role that can reduce, but not completely reverse, the pathology of AOM-treated mice. AOM-induced hepatic encephalopathy is characterized by elevations of ammonia and bile acids as a result of liver injury and both of these contribute to the progression of this disorder [34]. Therefore, fractalkine supplementation may provide an even greater therapeutic benefit if used in conjunction with other therapies for the management of acute liver failure or hepatic encephalopathy. Contrary to the data reported here, some reports in the brain, and a few investigating other inflammatory diseases, suggest that fractalkine can promote the infiltration of immune cells and generate inflammation. One of these studies demonstrates that increased fractalkine expression during experimental autoimmune encephalomyelitis, a mouse model of multiple sclerosis, promotes lymphocyte entry into the brain further exacerbating disease severity [35]. Further support of this is demonstrated in the α-synuclein-induced inflammation model of Alzheimer’s disease, where the use of CX3CR1 knockout mice has been demonstrated to reduce inflammation during this disease state compared to wild-type mice [13]. These differential effects during various neuropathies indicate that fractalkine signaling may generate different effects depending on the disease state and the pathology that contributes to the progression of the disease. Of interest is that CX3CR1 was initially thought to respond to fractalkine alone; however, more recently, CX3XR1 was found to also bind chemokine ligand 26, making the findings using CX3CR1 knockout mice difficult to interpret with respect to specific fractalkine signaling [36]. Neuroinflammation has been associated with a variety of neuropathies including multiple sclerosis, stroke, and Alzheimer’s disease [37]. In regard to hepatic encephalopathy, microglia activation has been observed during hepatic devascularization, toxic liver injury, and biliary cirrhosis [38–40]. Elevations of ammonia occur during hepatic encephalopathy and have been demonstrated to promote microglia activation in rats fed an ammonium-enriched diet indicating that ammonia may generate some of its effects on hepatic encephalopathy pathogenesis through exacerbating neuroinflammation [41]. During hepatic encephalopathy, microglia activation and subsequent neuroinflammation is thought to be primarily driven by increased concentrations of ammonia, lactate, manganese, or glutamate in the brain [42]. Currently, many of these effects are not well understood, but recent studies have begun to expand on how the initiation of neuroinflammation may occur. Neuroinflammation can be induced by chemokine signaling, such as through CCL2, which is upregulated during AOM-induced hepatic encephalopathy and chemokine receptor 2 or chemokine receptor 4 antagonist treatment improved outcomes in AOM-treated mice [8]. Similar findings were observed during bile duct ligation in mice where cerebral concentrations of CCL2 were elevated and were associated with monocyte infiltration in the brain, which was eliminated in CCL2 or chemokine receptor 2 knockout mice [40]. Of interest is that there appears to be an inverse relationship between CCL2 and fractalkine during hepatic encephalopathy. Gaining a greater understanding of the mechanism that generates these signaling events may provide better understanding of the initiators of inflammation during this disease state. This study was the first to report that supplementing fractalkine during hepatic encephalopathy suppresses CCL2 concentrations in the cortex, which indicates that the increase of CCL2 observed during AOM-induced hepatic encephalopathy is dependent upon the suppression of fractalkine. Therefore, there is support that chemokine signaling plays a significant role in the initiation of inflammation during hepatic encephalopathy, though more studies are needed to fully understand these signaling pathways and their roles in the pathogenesis of this disorder. As AOM-induced hepatic encephalopathy is the result of acute liver injury, the finding that soluble fractalkine infusion directly into the brain conferred some protection to the liver is interesting and unexpected. Due to fractalkine being administered directly into the lateral ventricle of the brain and at a low concentration, one would logically infer that fractalkine infusion should have little effect on the liver. Recent studies have identified a variety of findings that support that this treatment could influence the liver. The first is that the blood-brain barrier becomes permeable at the later stages AOM-induced hepatic encephalopathy, which could allow small amounts of soluble fractalkine to enter the circulation at the end stages of this model [43]. That being said, fractalkine signaling during liver disease is generally thought to contribute to hepatic inflammation. In liver sections from patients with fibrosis and cirrhosis, it has been reported that hepatic stellate cells upregulate ADAMs and subsequently lead to the release of soluble fractalkine that recruits inflammatory cells to the liver and contributes to increased liver pathology [44]. This contribution to liver injury has also been shown in primary biliary cirrhosis where fractalkine triggers the infiltration of CX3CR1 expressing immune cells to the liver that promote inflammation [45]. Therefore, the protective effects of soluble fractalkine ICV infusion on the liver may be a result of improved neurological function. The connection between brain pathology and liver dysfunction is beginning to become recognized as patients with stroke and traumatic brain injury have elevations of liver enzymes following injury [46, 47]. That being said, this change in serum transaminases and liver enzymes observed during neuropathies is not well classified and gaining an understanding of brain-liver axis signaling will require more studies. Conclusions The results presented suggest that following liver failure and the development of hepatic encephalopathy, fractalkine expression is reduced which contributes to the neurological decline associated with this disease. Direct cranial infusion of soluble fractalkine was found to slow neurological decline, reduce microglia activation, and decrease the concentrations of proinflammatory cytokines in the cortex. Therefore, treatments aimed at increasing fractalkine concentrations may be a potential therapeutic strategy for patients with hepatic encephalopathy. Abbreviations ADAMsA disintegrin and metalloproteinase proteins ALTAlanine aminotransferasase AOMAzoxymethane CCL2Chemokine ligand 2 DAPI4′,6-Diamidino-2-phenylindole ERK1/2Extracellular signal-regulated kinase 1/2 GLASTGlutamate aspartate transporter ICVIntracerebroventricular ILInterleukin RT-PCRReal-time PCR sFKNSoluble fractalkine TNFαTumor necrosis factor-α Acknowledgements This material is the result of work supported with resources and the use of facilities at the Central Texas Veterans Health Care System, Temple, Texas. The content is the responsibility of the author(s) alone and does not necessarily reflect the views or policies of the Department of Veterans Affairs or the United States Government. Funding This study was funded by an NIH R01 award (DK082435) awarded to Dr. DeMorrow, a VA Merit award (BX002638-01) from the United States Department of Veterans Affairs Biomedical Laboratory Research and Development Service (BLR&D) to Dr. DeMorrow, a Baylor Scott & White Health Research Mentorship Award (#150156) to Dr. Andry, and a Baylor Scott & White Health Research Mentorship Award (#150408) to Dr. Brown. Availability of data and materials No project specific resources (cell lines, plasmids, or mouse strains) were used in this study. All reagents used are commercially available. Raw data for these studies will be made to interested parties available upon request. Authors’ contributions MM, SG, GF, SA, AB, and SD performed the technical work and data analysis. MM, GF, and SD performed the statistical analyses. MM and SD conceived the study, designed and coordinated the experiments, and drafted the manuscript. All authors have read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate All animal experiments were performed following approval from Baylor Scott & White Health Institutional Animal Care and Use Committee (Temple, TX). ==== Refs References 1. Ostapowicz G Fontana RJ Schiodt FV Larson A Davern TJ Han SH McCashland TM Shakil AO Hay JE Hynan L Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States Ann Intern Med 2002 137 947 954 10.7326/0003-4819-137-12-200212170-00007 12484709 2. Bernal W Auzinger G Dhawan A Wendon J Acute liver failure Lancet 2010 376 190 201 10.1016/S0140-6736(10)60274-7 20638564 3. Butterworth RF Hepatic encephalopathy: a central neuroinflammatory disorder? Hepatology 2011 53 1372 1376 10.1002/hep.24228 21480337 4. Ferenci P Lockwood A Mullen K Tarter R Weissenborn K Blei AT Hepatic encephalopathy—definition, nomenclature, diagnosis, and quantification: final report of the working party at the 11th World Congresses of Gastroenterology, Vienna, 1998 Hepatology 2002 35 716 721 10.1053/jhep.2002.31250 11870389 5. Wright G Shawcross D Olde Damink SW Jalan R Brain cytokine flux in acute liver failure and its relationship with intracranial hypertension Metab Brain Dis 2007 22 375 388 10.1007/s11011-007-9071-4 17899343 6. Bemeur C Qu H Desjardins P Butterworth RF IL-1 or TNF receptor gene deletion delays onset of encephalopathy and attenuates brain edema in experimental acute liver failure Neurochem Int 2010 56 213 215 10.1016/j.neuint.2009.11.010 19931338 7. Jiang W Desjardins P Butterworth RF Cerebral inflammation contributes to encephalopathy and brain edema in acute liver failure: protective effect of minocycline J Neurochem 2009 109 485 493 10.1111/j.1471-4159.2009.05981.x 19220703 8. McMillin M Frampton G Thompson M Galindo C Standeford H Whittington E Alpini G DeMorrow S Neuronal CCL2 is upregulated during hepatic encephalopathy and contributes to microglia activation and neurological decline J Neuroinflammation 2014 11 121 10.1186/1742-2094-11-121 25012628 9. Harrison JK Jiang Y Chen S Xia Y Maciejewski D McNamara RK Streit WJ Salafranca MN Adhikari S Thompson DA Role for neuronally derived fractalkine in mediating interactions between neurons and CX3CR1-expressing microglia Proc Natl Acad Sci U S A 1998 95 10896 10901 10.1073/pnas.95.18.10896 9724801 10. Yoshida H Imaizumi T Fujimoto K Matsuo N Kimura K Cui X Matsumiya T Tanji K Shibata T Tamo W Synergistic stimulation, by tumor necrosis factor-alpha and interferon-gamma, of fractalkine expression in human astrocytes Neurosci Lett 2001 303 132 136 10.1016/S0304-3940(01)01699-8 11311510 11. Fonovic UP Jevnikar Z Kos J Cathepsin S generates soluble CX3CL1 (fractalkine) in vascular smooth muscle cells Biol Chem 2013 394 1349 1352 10.1515/hsz-2013-0189 23893684 12. Hurst LA Bunning RA Sharrack B Woodroofe MN siRNA knockdown of ADAM-10, but not ADAM-17, significantly reduces fractalkine shedding following pro-inflammatory cytokine treatment in a human adult brain endothelial cell line Neurosci Lett 2012 521 52 56 10.1016/j.neulet.2012.05.057 22641052 13. Thome AD Standaert DG Harms AS Fractalkine signaling regulates the inflammatory response in an alpha-synuclein model of Parkinson disease PLoS One 2015 10 e0140566 10.1371/journal.pone.0140566 26469270 14. Lyons A Lynch AM Downer EJ Hanley R O'Sullivan JB Smith A Lynch MA Fractalkine-induced activation of the phosphatidylinositol-3 kinase pathway attentuates microglial activation in vivo and in vitro J Neurochem 2009 110 1547 1556 10.1111/j.1471-4159.2009.06253.x 19627440 15. Briones TL Woods J Wadowska M Chronic neuroinflammation and cognitive impairment following transient global cerebral ischemia: role of fractalkine/CX3CR1 signaling J Neuroinflammation 2014 11 13 10.1186/1742-2094-11-13 24447880 16. Koyama Y Kotani M Sawamura T Kuribayashi M Konishi R Michinaga S Different actions of endothelin-1 on chemokine production in rat cultured astrocytes: reduction of CX3CL1/fractalkine and an increase in CCL2/MCP-1 and CXCL1/CINC-1 J Neuroinflammation 2013 10 51 10.1186/1742-2094-10-51 23627909 17. Zimmermann HW Trautwein C Tacke F Functional role of monocytes and macrophages for the inflammatory response in acute liver injury Front Physiol 2012 3 56 10.3389/fphys.2012.00056 23091461 18. McMillin M Galindo C Pae HY Frampton G Di Patre PL Quinn M Whittington E DeMorrow S Gli1 activation and protection against hepatic encephalopathy is suppressed by circulating transforming growth factor beta1 in mice J Hepatol 2014 61 1260 1266 10.1016/j.jhep.2014.07.015 25046848 19. McMillin M Frampton G Tobin R Dusio G Smith J Shin H Newell-Rogers K Grant S DeMorrow S TGR5 signaling reduces neuroinflammation during hepatic encephalopathy J Neurochem 2015 135 565 576 10.1111/jnc.13243 26179031 20. Rancan M Bye N Otto VI Trentz O Kossmann T Frentzel S Morganti-Kossmann MC The chemokine fractalkine in patients with severe traumatic brain injury and a mouse model of closed head injury J Cereb Blood Flow Metab 2004 24 1110 1118 10.1097/01.WCB.0000133470.91843.72 15529011 21. Frampton G Invernizzi P Bernuzzi F Pae HY Quinn M Horvat D Galindo C Huang L McMillin M Cooper B Interleukin-6-driven progranulin expression increases cholangiocarcinoma growth by an Akt-dependent mechanism Gut 2012 61 268 277 10.1136/gutjnl-2011-300643 22068162 22. DeMorrow S Francis H Gaudio E Venter J Franchitto A Kopriva S Onori P Mancinelli R Frampton G Coufal M The endocannabinoid anandamide inhibits cholangiocarcinoma growth via activation of the noncanonical Wnt signaling pathway Am J Physiol Gastrointest Liver Physiol 2008 295 G1150 1158 10.1152/ajpgi.90455.2008 18832445 23. Livak KJ Schmittgen TD Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method Methods 2001 25 402 408 10.1006/meth.2001.1262 11846609 24. Matkowskyj KA Marrero JA Carroll RE Danilkovich AV Green RM Benya RV Azoxymethane-induced fulminant hepatic failure in C57BL/6 J mice: characterization of a new animal model Am J Physiol 1999 277 G455 462 10444460 25. Cambien B Pomeranz M Schmid-Antomarchi H Millet MA Breittmayer V Rossi B Schmid-Alliana A Signal transduction pathways involved in soluble fractalkine-induced monocytic cell adhesion Blood 2001 97 2031 2037 10.1182/blood.V97.7.2031 11264168 26. Puga DA Tovar CA Guan Z Gensel JC Lyman MS McTigue DM Popovich PG Stress exacerbates neuron loss and microglia proliferation in a rat model of excitotoxic lower motor neuron injury Brain Behav Immun 2015 49 246 254 10.1016/j.bbi.2015.06.006 26100488 27. Marlatt MW Bauer J Aronica E van Haastert ES Hoozemans JJ Joels M Lucassen PJ Proliferation in the Alzheimer hippocampus is due to microglia, not astroglia, and occurs at sites of amyloid deposition Neural Plast 2014 2014 693851 25215243 28. Cash WJ McConville P McDermott E McCormick PA Callender ME McDougall NI Current concepts in the assessment and treatment of hepatic encephalopathy QJM 2010 103 9 16 10.1093/qjmed/hcp152 19903725 29. Pabon MM Bachstetter AD Hudson CE Gemma C Bickford PC CX3CL1 reduces neurotoxicity and microglial activation in a rat model of Parkinson's disease J Neuroinflammation 2011 8 9 10.1186/1742-2094-8-9 21266082 30. Merino J Aller MA Rubio S Arias N Nava MP Loscertales M Arias J Arias JL Gut-brain chemokine changes in portal hypertensive rats Dig Dis Sci 2011 56 2309 2317 10.1007/s10620-011-1625-y 21347560 31. Xie WL Shi Q Zhang J Zhang BY Gong HS Guo Y Wang SB Xu Y Wang K Chen C Abnormal activation of microglia accompanied with disrupted CX3CR1/CX3CL1 pathway in the brains of the hamsters infected with scrapie agent 263K J Mol Neurosci 2013 51 919 932 10.1007/s12031-013-0002-z 23526370 32. Denes A Ferenczi S Halasz J Kornyei Z Kovacs KJ Role of CX3CR1 (fractalkine receptor) in brain damage and inflammation induced by focal cerebral ischemia in mouse J Cereb Blood Flow Metab 2008 28 1707 1721 10.1038/jcbfm.2008.64 18575457 33. Donohue MM Cain K Zierath D Shibata D Tanzi PM Becker KJ Higher plasma fractalkine is associated with better 6-month outcome from ischemic stroke Stroke 2012 43 2300 2306 10.1161/STROKEAHA.112.657411 22798324 34. McMillin M Frampton G Quinn M Ashfaq S de los Santos M 3rd Grant S DeMorrow S Bile acid signaling is involved in the neurological decline in a murine model of acute liver failure Am J Pathol 2016 186 312 323 10.1016/j.ajpath.2015.10.005 26683664 35. Mills JH Alabanza LM Mahamed DA Bynoe MS Extracellular adenosine signaling induces CX3CL1 expression in the brain to promote experimental autoimmune encephalomyelitis J Neuroinflammation 2012 9 193 10.1186/1742-2094-9-193 22883932 36. Nakayama T Watanabe Y Oiso N Higuchi T Shigeta A Mizuguchi N Katou F Hashimoto K Kawada A Yoshie O Eotaxin-3/CC chemokine ligand 26 is a functional ligand for CX3CR1 J Immunol 2010 185 6472 6479 10.4049/jimmunol.0904126 20974991 37. Janssen B, Vugts DJ, Funke U, Molenaar GT, Kruijer PS, van Berckel BN, Lammertsma AA, Windhorst AD. Imaging of neuroinflammation in Alzheimer’s disease, multiple sclerosis and stroke: recent developments in positron emission tomography. Biochim Biophys Acta. 2016;1862(3):425–41. 38. Jiang W Desjardins P Butterworth RF Direct evidence for central proinflammatory mechanisms in rats with experimental acute liver failure: protective effect of hypothermia J Cereb Blood Flow Metab 2009 29 944 952 10.1038/jcbfm.2009.18 19259110 39. Rangroo Thrane V Thrane AS Chang J Alleluia V Nagelhus EA Nedergaard M Real-time analysis of microglial activation and motility in hepatic and hyperammonemic encephalopathy Neuroscience 2012 220 247 255 10.1016/j.neuroscience.2012.06.022 22728097 40. D'Mello C Le T Swain MG Cerebral microglia recruit monocytes into the brain in response to tumor necrosis factoralpha signaling during peripheral organ inflammation J Neurosci 2009 29 2089 2102 10.1523/JNEUROSCI.3567-08.2009 19228962 41. Rodrigo R Cauli O Gomez-Pinedo U Agusti A Hernandez-Rabaza V Garcia-Verdugo JM Felipo V Hyperammonemia induces neuroinflammation that contributes to cognitive impairment in rats with hepatic encephalopathy Gastroenterology 2010 139 675 684 10.1053/j.gastro.2010.03.040 20303348 42. van Rossum D Hanisch UK Microglia Metab Brain Dis 2004 19 393 411 10.1023/B:MEBR.0000043984.73063.d8 15554430 43. McMillin MA Frampton GA Seiwell AP Patel NS Jacobs AN DeMorrow S TGFbeta1 exacerbates blood-brain barrier permeability in a mouse model of hepatic encephalopathy via upregulation of MMP9 and downregulation of claudin-5 Lab Invest 2015 95 903 913 10.1038/labinvest.2015.70 26006017 44. Bourd-Boittin K Basset L Bonnier D L'Helgoualc'h A Samson M Theret N CX3CL1/fractalkine shedding by human hepatic stellate cells: contribution to chronic inflammation in the liver J Cell Mol Med 2009 13 1526 1535 10.1111/j.1582-4934.2009.00787.x 19432809 45. Sasaki M Miyakoshi M Sato Y Nakanuma Y Chemokine-chemokine receptor CCL2-CCR2 and CX3CL1-CX3CR1 axis may play a role in the aggravated inflammation in primary biliary cirrhosis Dig Dis Sci 2014 59 358 364 10.1007/s10620-013-2920-6 24185682 46. Fox A Sanderlin JB McNamee S Bajaj JS Carne W Cifu DX Elevated liver enzymes following polytraumatic injury J Rehabil Res Dev 2014 51 869 874 10.1682/JRRD.2013.10.0233 25479083 47. Muscari A Collini A Fabbri E Giovagnoli M Napoli C Rossi V Vizioli L Bonfiglioli A Magalotti D Puddu GM Zoli M Changes of liver enzymes and bilirubin during ischemic stroke: mechanisms and possible significance BMC Neurol 2014 14 122 10.1186/1471-2377-14-122 24903748

PMC005xxxxxx/PMC5000401.txt

==== Front BMC Public HealthBMC Public HealthBMC Public Health1471-2458BioMed Central London 357010.1186/s12889-016-3570-3Research ArticleSedentary bout durations and metabolic syndrome among working adults: a prospective cohort study Honda Takanori honda-t@eph.med.kyushu-u.ac.jp 123Chen Sanmei sanmei.chen@kyudai.jp 1Yonemoto Koji yonemoto_koji@med.kurume-u.ac.jp 4Kishimoto Hiro hiro4@eph.med.kyushu-u.ac.jp 2Chen Tao chentwhy@gmail.com 1Narazaki Kenji narazaki@fit.ac.jp 5Haeuchi Yuka yuka_haeuchi@yahoo.co.jp 1Kumagai Shuzo +81-92-583-7853shuzo@ihs.kyushu-u.ac.jp 161 Department of Behavior and Health Sciences, Graduate School of Human-Environment Studies, Kyushu University, 6-1 Kasuga kouen, Kasuga City, Fukuoka Prefecture 816-8580 Japan 2 Department of Epidemiology and Public Health, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka City, Fukuoka Prefecture 812-8582 Japan 3 Research Fellow of the Japan Society for the Promotion of Science, 5-3-1 Kojimachi, Chiyoda-ku, Tokyo 102-0083 Japan 4 Biostatistics Center, Kurume University, 67 Asahi-machi, Kurume, Fukuoka Prefecture 830-0011 Japan 5 Department of Socio-Environmental Studies, Fukuoka Institute of Technology, 3-30-1 Wajiro-higashi, Higashi-ku, Fukuoka City, Fukuoka Prefecture 811-0295 Japan 6 Faculty of Arts and Science, Kyushu University, 6-1 Kasuga kouen, Kasuga City, Fukuoka Prefecture 816-8580 Japan 26 8 2016 26 8 2016 2016 16 1 88810 2 2016 20 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background This study aimed to examine the associations between time spent in prolonged and non-prolonged sedentary bouts and the development of metabolic syndrome. Methods We used data from a prospective study of Japanese workers. Baseline examination was conducted between 2010 and 2011. A total of 430 office workers (58 women) aged 40-64 years without metabolic syndrome were followed up by annual health checkups until 2014. Metabolic syndrome was defined as having ≥ 3 out of 5 diagnostic criteria from the Joint Interim Statement 2009 definition. Sedentary time was assessed using a tri-axial accelerometer. Time spent in total, prolonged (accumulated ≥ 30 min) and non-prolonged sedentary bouts (accumulated < 30 min) was calculated. Cox proportional hazards models were used to estimate the risk of developing metabolic syndrome. Results During a median follow-up of 3 years, 83 participants developed metabolic syndrome. After adjustment for age, sex, education, smoking, and family income, positive associations were observed between time spent in prolonged sedentary bouts and the development of metabolic syndrome. After additional adjustment for moderate-to-vigorous physical activity, those in the three highest quartiles of time spent in prolonged sedentary bouts showed higher risk of metabolic syndrome compared to the lowest quartile group, with adjusted hazard ratios (95 % confidence intervals) of 2.72 (1.30 – 5.73), 2.42 (1.11 – 5.50), and 2.85 (1.31 – 6.18), respectively. No associations were seen for time spent in total and non-prolonged sedentary bouts. Conclusions Sedentary behavior accumulated in a prolonged manner was associated with an increased risk of metabolic syndrome. In devising public health recommendations for the prevention of metabolic disease, the avoidance of prolonged uninterrupted periods of sedentary behavior should be considered. Electronic supplementary material The online version of this article (doi:10.1186/s12889-016-3570-3) contains supplementary material, which is available to authorized users. Keywords EpidemiologyAccelerometrySedentary lifestylePhysical activityCentral obesityMetabolic syndromeThe Japanese Society for the Promotion of Science Fellows15J03431Honda Takanori Grant-in-Aid for Scientific Research (A) from the Ministry of Education, Culture, Sports, Science and Technology of Japan22240073Kumagai Shuzo The Practical Research Project for Life-Style related Diseases including Cardiovascular Diseases and Diabetes Mellitus from the Japan Agency for Medical Research and Development, AMED15ek0210001h0003Kumagai Shuzo issue-copyright-statement© The Author(s) 2016 ==== Body Background Metabolic syndrome represents a cluster of metabolic disorders that include central obesity, elevated blood pressure, dyslipidemia, and hyperglycemia [1]. Metabolic syndrome is prevalent worldwide [2–4]; in the US, a recent report estimated that approximately one-fourth of adults have metabolic syndrome [2]. Since metabolic syndrome confers an elevated risk of cardiovascular diseases and type 2 diabetes [1], studies exploring modifiable risk factors for metabolic syndrome are essential to develop public health strategies for chronic disease prevention. A large body of epidemiological literature has shown that physical inactivity (i.e., lack of physical activity) is a driving factor for the global epidemic of non-communicable diseases [5]. Alongside physical inactivity, sedentary behavior, defined as prolonged periods of inactivity involving sitting or reclining, has recently been revealed to be associated with adverse metabolic and vascular health outcomes [6, 7]. Importantly, the detrimental effects of sedentary behavior on health are independent of lack of moderate-to-vigorous physical activity (MVPA) [8, 9]. Although a meta-analysis of cross-sectional studies revealed associations between sedentary behavior and metabolic syndrome [10], the role of sedentary behavior on the development of metabolic syndrome over time has been poorly understood. Only one study has reported a longitudinal association of sedentary behavior with metabolic syndrome; in that report, longer sedentary time was shown to increase the subsequent risk of developing metabolic syndrome independent of leisure-time physical activity [11]. The lack of evidence from prospective studies has precluded conclusions on the causality of the relationship between sedentary behavior and metabolic syndrome. Recent studies have indicated that, in addition to total sedentary time, the manner in which sedentary time was accumulated has important health implications [12]. Several experimental studies have suggested that uninterrupted periods of sedentary behavior, compared to interrupted ones, exerted a detrimental impact on postprandial glucose and lipid responses [13, 14]. A cross-sectional observational study using an accelerometer reported that sedentary time in non-prolonged bouts was not associated with any cardiometabolic biomarkers, while that in prolonged bouts was associated with higher waist circumference and body mass index [15]. Other cross-sectional studies have even reported favorable associations between non-prolonged sedentary time and cardiometabolic and anthropometric measures [16, 17]. Accelerometers can objectively record minute-by-minute activities at different levels of intensity, and this objective measurement of physical activity allows us to capture periods (bouts) of consecutive minutes of activities. An analysis of accelerometry data showed that office workers spent 75 % of their workday being sedentary, with much of the time accumulated in prolonged (>30 min) bouts [18], suggesting that the health of office workers may be at risk of prolonged sedentary time and its consequences. Thus far, to our knowledge, there have been no studies examining the associations between different durations of sedentary bouts and metabolic syndrome. We therefore addressed these issues by examining the prospective associations between objectively-measured time spent in sedentary behavior of different bout lengths and the development of metabolic syndrome. Here, we tested our hypothesis that a greater amount of sedentary time in prolonged bouts is associated with an increased risk of developing metabolic syndrome, independent of the levels of moderate-to-vigorous physical activity. Methods Participants This study was conducted by using data from the Ryobi Health Survey, which is a prospective study carried out among Japanese workers to investigate social and behavioral risk factors for metabolic syndrome among working adults. The subjects were Japanese workers aged 30 years and over working in a Japanese enterprise group in Okayama prefecture, located on the western half of the main island of Japan. The enterprise, Ryobi Holdings, consists of seven companies involved primarily in information technology and transportation. Potential Subjects consisted of employees of the enterprises aged 40 to 64 years (n = 691) who were contacted by mail to participate in the Ryobi Health Survey. We did not include those aged between 30 and 39 in the present analyses since the Specific Health Checkups and Guidance in Japan (Tokutei Hoken Shido), a national screening and interventional program for metabolic syndrome, is geared toward those aged 40 and over. Baseline examination was conducted between January 2010 and March 2011, and the participants without metabolic syndrome at baseline were followed up in annual health checkups until March 2014. Among the 691 subjects contacted by mail, 682 agreed to participate in the present study, representing a response rate of 98.7 %. Of this initial sample, data on the components of metabolic syndrome at baseline were available for 660 participants, and 502 participants without metabolic syndrome at baseline were eligible for the present study. Thirty-six participants without valid accelerometer data and 25 participants with missing data on covariates were further excluded. In addition, 11 individuals were lost to follow-up. Finally, 430 participants were included in the analyses. A comparison between subjects excluded from and those included in the present analysis is shown (see Additional file 1: Table S1). There was no evidence of selection bias due to the exclusion. All participants provided written informed consent. This study was conducted in accordance with the Declaration of Helsinki. The entire study protocol was approved by the ethics committee of the Institute of Health Science, Kyushu University, Fukuoka, Japan. Assessment of metabolic syndrome All data on anthropometry, blood pressure, lipid and glucose profile, and medication use were obtained from the annual health examinations, which were conducted in accordance with the Industrial Safety and Health Act [19]. Height and body weight were measured in light clothing without shoes. Waist circumference was measured to the nearest 0.1 cm at the umbilical level while standing. Systolic and diastolic blood pressures were measured at rest by an automated sphygmomanometer. Serum triglycerides, high-density lipoprotein (HDL) cholesterol, and blood glucose were measured using enzymatic methods. All participants were asked to fast overnight before the blood test. Metabolic syndrome was defined based on the Joint Interim Statement 2009 definition [1]. Specifically, participants having ≥ 3 of the following five clinical measures were considered as having metabolic syndrome: (1) central obesity (waist circumference (≥90 cm in men, or ≥ 85 cm in women); (2) elevated blood pressure, defined as systolic blood pressure ≥ 130 mmHg or diastolic blood pressure ≥ 85 mmHg, or taking an antihypertensive medication; (3) elevated fasting blood glucose level ≥ 5.6 mmol/L or taking a hypoglycemic medication; (4) decreased HDL-cholesterol level (<1.0 mmol/L in men or < 1.3 mmol/L in women); and (5) hypertriglyceridemia (≥1.7 mmol/L) or taking a lipid-lowering medication. Objective measurement of sedentary behavior and physical activity Sedentary behavior and MVPA were assessed using a tri-axial accelerometer device (Active style Pro HJA 350-IT; Omron Healthcare Co., Ltd., Kyoto, Japan). Details of the accelerometer measurement procedure have been reported elsewhere [20]. Briefly, participants wore the device during waking hours for 10 consecutive days, except while bathing or sleeping. Data were recorded in 60-sec epochs. The accuracy of the intensity estimation has been validated with the Douglas bag method [21, 22]. Non-wear time was defined as a period of at least 60 consecutive minutes of no activity (i.e., estimated activity intensity < 1.0 metabolic equivalent, or MET) with allowance for up to two consecutive minutes of activities with intensity equal to 1.0 MET. We adapted the SAS macro program for the ActiGraph monitor provided by the National Cancer Institute to compute daily non-wear time [23], with modifications for our accelerometer [20]. Days with at least 600 min of wear time were considered valid [24]. Participants with at least four valid days were included in the analysis. In this study, sedentary behavior was defined as any activity with an accelerometer-estimated intensity of ≤ 1.5 METs. We considered each minute that the activity intensity was ≤ 1.5 METs as sedentary time. A sedentary bout was defined as a period of time in continuous sedentary time where the activity intensity fell into the sedentary range with no interruption. For example, a bout of 30 min of sedentary time was not an accumulation of three 10-min bouts but rather a consecutive 30-min period of sedentary time. The amounts of time spent in prolonged (accumulated ≥ 30 min) and non-prolonged sedentary bouts (accumulated < 30 min) were calculated separately. Total sedentary time was calculated as the sum of the prolonged and non-prolonged sedentary periods. MVPA was defined as activities of ≥ 3 METs. An MVPA bout was defined as a period of time in continuous activities where the activity intensity was ≥ 3 METs. A bout of MVPA lasting for at least 10 min, with allowance for up to 2 min of non-MVPA activity, was considered an MVPA period, which is consistent with the consensus recommendation that physical activity accumulated in periods lasing for ≥ 10 min benefits health [25]. Other variables Information on education (higher or lower than university education) and current smoking habits (yes or no) was obtained from a self-report questionnaire. Household income data were obtained by questionnaire and categorized as <4 million, 4-8 million, or ≥ 8 million Japanese yen per year. Statistical analysis All statistical analyses were performed with the SAS software version 9.3 (SAS Institute, Cary, NC, USA) with a significance level of α = 0.05. Person-time for each participant was calculated from the date of baseline examination to the date of the first occurrence of metabolic syndrome or the last examination, whichever came first. Only two participants missed a follow-up examination and were confirmed as having metabolic syndrome in the subsequent year. Since the findings were basically unchanged after excluding these participants, results were presented including data from these participants. Given that the wear-time potentially influences the sedentary time, sedentary time variables were adjusted for wear-time using the residual method [26, 27]. Wear time-adjusted sedentary time variables were divided into sex-specific quartiles for analyses. The quartile boundaries for each exposure are shown (see Additional file 1: Table S2). The variables were expressed as the median (interquartile range, IQR) for continuous data, and the frequency for categorical data. To model the effects of sedentary variables on first-ever metabolic syndrome, Cox proportional hazards models were used to calculate hazard ratios (HRs) with 95 % confidence intervals (CIs) for the development of metabolic syndrome by quartiles of total, prolonged and non-prolonged sedentary time. The proportional hazards assumption was assessed by visual inspection of log-log plots. The first model was adjusted for sex and age. The second model was adjusted for sex, age, education, and current smoking. To examine whether the associations of sedentary time with metabolic syndrome were independent of physical activity, the third model was adjusted for variables in the second model and MVPA. Waist circumference was further adjusted in the full model to examine whether the associations were independent of central obesity. We tested interactions of sedentary variables with age, sex, and MVPA (<150 or ≥ 150 min/week) in each survival model, to examine potential effect-modifications by age, sex, and physical activity. None of the interaction terms were statistically significant, showing that these factors did not modify the effect of sedentary behavior on the development of metabolic syndrome. Here, we should note that metabolic syndrome is a “reversible” state. To elucidate the potential impact of this reversibility on the Cox proportional hazard model, we repeated the analysis while excluding those who took medications at baseline (we assumed that these subjects would be more likely to receive treatment or lifestyle intervention by physicians during the follow-up period). The result was not substantially changed from the main analysis, suggesting that such cases would not have affected the present findings, and thus we presented the results drawn from the whole sample. Sensitivity analyses were performed using different cut-off points (≥10 and ≥ 20 min) for differentiating prolonged sedentary time from non-prolonged sedentary time. Additionally, the associations between sedentary variables and metabolic syndrome were analyzed in those at a higher risk of developing metabolic syndrome (with two affected components at baseline) and those with a lower risk (with one or no affected components) separately. Results Baseline characteristics of the study participants Among the 430 participants, 87.5 % were men and the median age at baseline was 48 years. The participants wore the accelerometer for 8.5 valid days on average, and the median (IQR) device wear-time was 846 (786 – 920) minutes/day. Table 1 summarizes participants’ baseline characteristics by quartiles of prolonged sedentary time. Participants who had higher amounts of prolonged sedentary time were younger and more highly educated, and were less likely to be current smokers. There was a significant difference in household income among groups.Table 1 Sample characteristics by quartiles of time spent in prolonged sedentary bouts (≥30 min), the Ryobi Health Survey 2009 – 2010 (n = 430) Characteristics Time spent in prolonged sedentary boutsa p for trend Q1 (n = 107) Q2 (n = 108) Q3 (n = 107) Q4 (n = 108) Age group, % (n) <0.001  40–49 48.6 (52) 45.4 (49) 62.6 (67) 76.9 (83)  50–59 40.2 (43) 41.7 (45) 28.0 (30) 20.4 (22)  60–64 11.2 (12) 13.0 (14) 9.3 (10) 2.8 (3) Women, % (n) 13.1 (14) 13.9 (15) 13.1 (14) 13.9 (15) 0.9130 Education, college or university level, % (n) 36.4 (39) 57.4 (62) 72.0 (77) 73.1 (79) <0.001 Current smoker, % (n) 36.4 (39) 38.9 (42) 29.0 (31) 25.0 (27) 0.028 Family income (JPY), % (n) 0.028   < 4 million 19.6 (21) 20.4 (22) 5.6 (6) 2.8 (3)  4–8 million 57.9 (62) 49.1 (53) 65.4 (70) 74.1 (80)  8+ million 22.4 (24) 30.6 (33) 29.0 (31) 23.1 (25) Moderate-to-vigorous physical activity, min/wk, median (interquartile range) 50 (11, 141) 66 (23, 168) 57 (11, 151) 41 (11, 174) 0.564 Device wear-time, min/d, median (interquartile range) 850 (785, 916) 843 (790, 922) 834 (787, 912) 865 (776, 929) 0.7503 Central obesity, % (n) 14.0 (15) 11.1 (12) 13.1 (14) 11.1 (12) 0.635 Elevated blood pressure, % (n) 29.0 (31) 33.3 (36) 28.0 (30) 32.4 (35) 0.801 Hypertriglyceridemia, % (n) 10.3 (11) 18.5 (20) 16.8 (18) 24.1 (26) 0.015 Low HDL-cholesterol level, % (n) 1.9 (2) 2.8 (3) 4.7 (5) 1.9 (2) 0.799 Hyperglycemia, % (n) 40.2 (43) 35.2 (38) 32.7 (35) 25.9 (28) 0.026 Number of affected components, % (n)b 0.878  Zero 30.8 (33) 29.6 (32) 33.6 (36) 27.8 (30)  One 43.0 (46) 39.8 (43) 37.4 (40) 49.1 (53)  Two 26.2 (28) 30.6 (33) 29.0 (31) 23.1 (25) Data are presented as a median (interquartile range) or % (n). HDL-cholesterol, high-density lipoprotein cholesterol aTime spent in prolonged sedentary time was adjusted for time spent wearing the device using the residual method prior to classifying into sex-specific quartiles (Q1 – Q4). Cut-offs for quartiles were 106.7, 165.5, and 269.2 min/day for men, and 65.1, 122.7, and 195.4 for women for those who wore the accelerometer device for the average amount of wear-time bNumber of components of metabolic syndrome at baseline survey Prospective effects of baseline sedentary behavior on the risk of metabolic syndrome During a median follow-up period of 3 years (IQR 3-4 years), 76 men and 7 women developed metabolic syndrome. The HRs (95 % CIs) of total, non-prolonged, and prolonged sedentary time for metabolic syndrome are shown in Table 2. No associations between total sedentary time and metabolic syndrome were found in any models. Similarly, the association of non-prolonged sedentary time (<30-min bouts) with the development of metabolic syndrome was not significant in any models. On the other hand, significant associations were observed between prolonged sedentary time (≥30-min bouts) and increased risk of metabolic syndrome. Those in the second and the higher quartiles showed significantly higher risk of metabolic syndrome compared with the lowest quartile group in the sex and age-adjusted and multivariate-adjusted models. After adjustment for MVPA, the association remained significant with adjusted HRs (95 % CI) of 2.72 (1.30 – 5.73), 2.42 (1.11 – 5.5), and 2.85 (1.31 – 6.18). This association did not change even after adjustment for waist circumference. When prolonged sedentary time was defined as ≥ 10-min or ≥ 20-min bouts, neither prolonged nor non-prolonged sedentary time was associated with increased risk of metabolic syndrome (Additional file 1: Table S3).Table 2 Multivariable-adjusted hazard ratios (95 % confidence intervals) for the development of metabolic syndrome, the Ryobi Health Survey 2009 – 2010 (n = 430) Cases (n) Incident rate (per 1000 person-years) Model 1 Model 2 Model 3 Model 4 HR 95 % CI p value HR 95 % CI p value HR 95 % CI p value HR 95 % CI p value Total sedentary time (≥1-min bout)  Q1 22 62.5 1.00 1.00 1.00 1.00  Q2 23 68.0 1.23 (0.63 – 2.39) 0.542 1.39 (0.67 – 2.86) 0.376 1.37 (0.66 – 2.82) 0.398 1.50 (0.73 – 3.09) 0.272  Q3 20 63.3 1.66 (0.88 – 3.13) 0.116 1.87 (0.94 – 3.72) 0.075 1.84 (0.92 – 3.68) 0.083 1.76 (0.87 – 3.55) 0.118  Q4 18 56.6 1.12 (0.56 – 2.21) 0.752 1.30 (0.61 – 2.76) 0.500 1.26 (0.59 – 2.69) 0.559 1.55 (0.70 – 3.43) 0.278 Non-prolonged sedentary time (<30-min bout)  Q1 16 52.6 1.00 1.00 1.00 1.00  Q2 18 55.9 0.71 (0.38 – 1.31) 0.268 0.72 (0.39 – 1.34) 0.298 0.71 (0.38 – 1.33) 0.287 0.79 (0.42 – 1.48) 0.465  Q3 27 81.3 0.82 (0.45 – 1.48) 0.500 0.82 (0.45 – 1.51) 0.520 0.81 (0.44 – 1.49) 0.491 1.09 (0.59 – 2.03) 0.785  Q4 22 60.1 0.85 (0.47 – 1.52) 0.573 0.85 (0.47 – 1.56) 0.606 0.83 (0.45 – 1.52) 0.546 1.08 (0.57 – 2.02) 0.817 Prolonged sedentary time (≥30-min bout)  Q1 20 58.3 1.00 1.00 1.00 1.00  Q2 23 70.6 2.58 (1.24 – 5.37) 0.011 2.71 (1.29 – 5.68) 0.009 2.72 (1.30 – 5.73) 0.008 3.03 (1.42 – 6.49) 0.004  Q3 21 64.4 2.16 (1.02 – 4.59) 0.045 2.41 (1.11 – 5.25) 0.026 2.42 (1.11 – 5.25) 0.026 2.25 (1.03 – 4.92) 0.040  Q4 19 57.8 2.49 (1.18 – 5.24) 0.017 2.86 (1.31 – 6.21) 0.008 2.85 (1.31 – 6.18) 0.008 2.90 (1.30 – 6.44) 0.009 HR hazard ratio, CI confidence interval. Sedentary variables were adjusted for time spent wearing the device using the residual method prior to classifying into sex-specific quartiles. Model 1 was adjusted for sex and age. Model 2 was adjusted for sex, age, education, smoking, and family income. Model 3 was additionally adjusted for moderate-to-vigorous physical activity. Model 4 was additionally adjusted for waist circumference. Cut-offs for quartiles were 106.7, 165.5, and 269.2 min/day for men, and 65.1, 122.7, and 195.4 for women among those who wore the accelerometer device for the average amount of wear-time Figure 1 illustrates the associations of prolonged sedentary time (≥30-min bouts) with metabolic syndrome by the number of components of metabolic syndrome at baseline. In the group with 0-1 metabolic syndrome components, although the multivariable-adjusted hazard rates in the three highest quartiles were 1.9 to 2.7-fold greater than that in the lowest group, these differences were not statistically significant (panel A). In contrast, among participants with two affected components, the three highest quartile groups had about 3.5-fold significantly greater risk of developing metabolic syndrome compared with the lowest quartile group (panel B).Fig. 1 Associations of time spent in prolonged sedentary bouts with metabolic syndrome according to the number of metabolic syndrome components at baseline. Panel a, Participants who at baseline had zero or one affected component. The numbers of participants in each group were 79, 75, 76, and 83 for Q1 to Q4 in panel A, respectively. Panel b, Participants with two affected components at baseline. The numbers of participants in each group were 28, 33, 31, and 25 for Q1 to Q4 in panel B, respectively. *p < 0.05 Discussion In this longitudinal study of adult office workers, we found that longer time spent in prolonged sedentary bouts (≥30-min bouts), but not in shorter ones, was significantly associated with higher risk of metabolic syndrome. The significance did not change after additional adjustment for MVPA, suggesting that the associations between time spent in prolonged sedentary bouts and metabolic syndrome was independent of MVPA. There were several strengths in the present study. To our knowledge, this is the first study to examine prospective associations between objectively-measured sedentary time and metabolic syndrome. The prospective design allowed to establish a temporal sequence. Secondly, the use of a tri-axial accelerometer with a validated algorithm for estimating low-intensity physical activities enhanced the accuracy of our assessment. In addition, the minute-by-minute monitoring approach enabled us to quantify sedentary periods of prolonged and non-prolonged bout lengths. Our study also has several limitations. First, the results may not be generalizable to other working adult populations as the participants were mainly engaged in sedentary occupations. Second, sedentary time was assessed only at baseline, which could have led to misclassification. Such misclassification, if present, would have weakened the relationship between sedentary time and the development of metabolic syndrome, and biased the results toward the null hypothesis. Third, the accelerometer is unable to differentiate standing and sitting. Indeed, we observed that quiet standing could be incorrectly classified as a sedentary period by the accelerometer in a laboratory setting (data not shown). Future research should use an objective measure which is able to distinguish sitting from standing postures. However, we would expect that most office workers would spend a majority of their work time sitting rather than standing, and few occupations require standing quietly for half of an hour or more. Finally, we did not measure some important covariates, such as diet and family history of type 2 diabetes. The findings from the present study suggested that objectively-measured sedentary time is an independent risk factor for metabolic syndrome, which was consistent with a previous longitudinal observation using a self-reported measure of sedentary behaviors [11]. Also, our findings were in line with cross-sectional studies using an accelerometer device to assess sedentary time [28–30], although some exceptions exist [31]. Our findings have extended these works by examining prolonged sedentary time and metabolic syndrome over time. Similarly, in a cross-sectional study of overweight/obese adults it was reported that prolonged sedentary time, but not non-prolonged sedentary time (<30-min bouts), was associated with cardio-metabolic risks [15]. Indeed, another study even found favorable associations between sedentary time in very short bouts (representing frequent transitions of postures) and cardio-metabolic risk profiles [16]. There has also been a cross-sectional study which reported no associations between sedentary time and the presence of metabolic syndrome [31]. The discrepancy may be partly attributable to methodological issues. Typically, sedentary time has been quantified by counting every single minute (i.e., ≥ 1-min bouts) in which the activity counts were below a threshold for sedentary behavior [29, 30]. By this method, sedentary time would be compounded not only by prolonged but also non-prolonged sedentary periods; thus the effects on metabolic syndrome would partly cancel each other out, which would presumably be inappropriate to reflect the prolonged nature of sedentary behavior. To our knowledge, only one study examining metabolic syndrome employed sedentary bouts (≥5-min) to calculate time spent in sedentary behavior [28]; however, a period of 5 min now seems to be insufficient for a definition of sustained sedentary behavior, based on our present results suggesting that prolonged sedentary time of 30 min contributes to the development of metabolic syndrome. In the present analysis, the three highest quartiles of prolonged sedentary time had similar HRs, suggesting that there was a threshold of duration of prolonged sedentary time. Given that the residual method was used to adjust for the device wear-time, the cut-off value between the lowest and second quartiles (106.7 min for men and 65.1 min for women) should be interpreted as applying to individuals wearing the device for the average wear-time of this population. These values could be underestimated and thus conservative compared to real-world settings, considering that, in practice, the device wear-time is likely to be shorter than actual waking hours. We also observed that the associations were stronger in participants with two affected components at baseline compared with those with no or one affected component, suggesting that those at higher metabolic risk could benefit more from a reduction in prolonged sedentary behavior. However, the absence of a significant association among those with no or one component does not necessarily indicate that the recommendation to reduce prolonged sedentary periods should be withheld from this population. Due to the short follow-up periods, only 32 cases of metabolic syndrome occurred in this group, which could potentially have resulted in insufficient statistical power to detect to a significant difference. Further investigation, including studies with longer follow-up, is needed to address this issue. The mechanisms by which sedentary behavior independently increases the risk of chronic disease remain to be fully elucidated. Our results showed no substantial change after adjustment for waist circumference, suggesting that mechanisms other than central obesity may contribute to the deleterious impact of sedentary behavior on metabolic syndrome. Hamilton and colleagues suggested that the activities of lipoprotein lipase, which locally regulate the uptake of triglycerides into muscle and the HDL-cholesterol concentration, were suppressed by prolonged periods without muscle contraction [32, 33]. Bed rest, a model of extreme sedentary behavior, has been shown to induce insulin resistance in skeletal muscle, reduced fatty acid oxidation, muscle atrophy, and a shift in muscle fiber type and ectopic fat storage [34]. Those physiological adaptations could also be induced by a certain prolonged period of sedentary time (i.e., ≥ 30 min). Our findings suggest the need for public health messages and policies to reduce not total but sustained sedentary periods, which has not yet been considered [35]. Interruptions of sedentary bouts in the present analysis were probably made by standing up or walking or by movement during sitting in which the intensity exceeded 1.5 METs. Therefore, not only the use of brief activity breaks to disrupt prolonged periods of sitting but also increasing movements of ≥ 1.5 METs while sitting, such as stretching, may be beneficial for prevention of metabolic syndrome. Conclusions Sedentary behavior accumulated in a prolonged manner was shown to be associated with an increased risk of developing metabolic syndrome. Reducing time spent in prolonged sedentary bouts may be beneficial for the prevention of metabolic syndrome. These results highlight the importance of sedentary bouts, which should be taken into account in the recommendations for the primary prevention of metabolic syndrome. Public health recommendations regarding the prevention of metabolic diseases may need to include avoiding prolonged uninterrupted periods. Additional file Additional file 1: Table S1. Comparisons between included and excluded subjects. Table S2. Quartile boundaries for time spent in total, non-prolonged and prolonged sedentary bouts. Table S3. Multivariable-adjusted hazard ratios (95 % confidence intervals) for the development of metabolic syndrome by different bout thresholds. (DOCX 23 kb) Abbreviations CIsConfidence intervals HDL-cholesterolHigh-density lipoprotein cholesterol HRsHazard ratios IQRInterquartile range METMetabolic equivalent MVPAModerate-to-vigorous physical activity Acknowledgements The authors would like to thank the employees of Ryobi Holdings, who kindly participated in the study, and the project staff for assisting in the data collection. TH is supported by the Japanese Society for the Promotion of Science. SC is supported by the China Scholarship Council (CSC). Funding The present study was funded by a Grant-in-Aid for the Japanese Society for the Promotion of Science Fellows (15 J03431) to TH, by a Grant-in-Aid for Scientific Research (A) (22240073) from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and by funds from the Practical Research Project for Life-Style-related Diseases including Cardiovascular Diseases and Diabetes Mellitus from the Japan Agency for Medical Research and Development, AMED (15ek0210001h0003), to SK. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Availability of data and materials The datasets during and/or analysed during the current study available from the corresponding author on reasonable request. Authors’ contributions All authors have contributed sufficiently to warrant authorship: TH had full access to all the data, participated in the design of the study and data collection, performed statistical analysis, interpreted data, and drafted the manuscript. SC, TC, and YH contributed to the design of the study, interpretation of the results, and the revision of the manuscript. KY participated in the design of the study, supervised the statistical analyses, interpreted data, and made critical revisions. KN and HK contributed to the interpretation of results and critical revision of the manuscript. HK contributed to data acquisition and survey coordination. SK supervised the study and contributed to survey planning and coordination. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate All participants provided written informed consent. This study was conducted in accordance with the Declaration of Helsinki. The entire study protocol was approved by the ethics committee of the Institute of Health Science, Kyushu University, Fukuoka, Japan. ==== Refs References 1. Alberti KG Eckel RH Grundy SM Zimmet PZ Cleeman JI Donato KA Harmonizing the metabolic syndrome: A joint interim statement of the international diabetes federation task force on epidemiology and prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; Atherosclerosis Society; and International Association for the Study of Obesity Circulation 2009 120 1640 5 10.1161/CIRCULATIONAHA.109.192644 19805654 2. Beltrán-Sánchez H Harhay MO Harhay MM McElligott S Prevalence and trends of metabolic syndrome in the Adult U.S. Populatoion, 1999-2010 J Am Coll Cardiol 2013 62 697 703 10.1016/j.jacc.2013.05.064 23810877 3. Scuteri A Laurent S Cucca F Cockcroft J Cunha PG Mañas LR Metabolic syndrome across Europe: different clusters of risk factors Eur J Prev Cardiol 2015 22 486 91 10.1177/2047487314525529 24647805 4. Li R Li W Lun Z Zhan H Sun Z Kanu JS Prevalence of metabolic syndrome in Mainland China: a meta-analyssi of publichsed studies BMC Public Health 2016 16 296 10.1186/s12889-016-2870-y 27039079 5. Lee IM Shiroma EJ Lobelo F Puska P Blair SN Katzmarzyk PT for the Lancet Physical Activity Secies Working Group Effect of physical inactivity on major non-communicable diseases worldwide: an analysis of burden of disease and life expectancy Lancet 2012 380 219 29 10.1016/S0140-6736(12)61031-9 22818936 6. Wilmot EG Edwardson CL Achana FA Davies MJ Gorely T Gray LJ Sedentary time in adults and the association with diabetes, cardiovascular disease and death : systematic review and meta-analysis Diabetologia 2012 55 2895 905 10.1007/s00125-012-2677-z 22890825 7. Dunstan DW Howard B Healy GN Owen N Too much sitting-a health hazard Diabetes Res Clin Pract 2012 97 368 76 10.1016/j.diabres.2012.05.020 22682948 8. Owen N Healy GN Matthews CE Dunstan DW Too much sitting: the population-health science of sedentary behavior Exerc Sport Sci Rev 2010 38 105 13 10.1097/JES.0b013e3181e373a2 20577058 9. Biswas A Oh PI Faulkner GE Bajaj RR Silver MA Mitchell MS Sedentary time and its association with risk for disease incidence, mortality, and hospitalization in adults: a systematic review and Meta-analysis Ann Intern Med 2015 162 123 32 10.7326/M14-1651 25599350 10. Edwardson CL Gorely T Davies MJ Gray LJ Khunti K Wilmot EG Association of sedentary behaviour with metabolic syndrome: a meta-analysis PLoS One 2012 7 e34916 10.1371/journal.pone.0034916 22514690 11. Greer AE Sui X Maslow AL Greer BK Blair SN The effects of sedentary behavior on metabolic syndrome independent of physical activity and cardiorespiratory fitness J Phys Act Health 2015 12 68 73 10.1123/jpah.12.1.68 25760431 12. Healy GN Dunstan DW Salmon J Cerin E Shaw JE Zimmet PZ Breaks in sedentary time: beneficial associations with metabolic risk Diabetes Care 2008 31 661 6 10.2337/dc07-2046 18252901 13. Dunstan DW Kingwell BA Larsen R Haely GN Cerin E Hamilton MT Breaking up prolonged sitting reduces postprandial glucose and insulin responses Diabetes Care 2012 35 976 83 10.2337/dc11-1931 22374636 14. Thorp AA Kingwell BA Sethi P Hammond L Owen N Dunstan DW Alternating bouts of sitting and standing attenuates postprandial glucose responses Med Sci Sports Exerc 2014 46 2053 61 10.1249/MSS.0000000000000337 24637345 15. Healy GN Winkler EAH Brakenridge CL Reeves MM Eakin EG Accelerometer-derived sedentary and physical activity time in overweight/obese adults with type 2 diabetes: cross-sectional associations with cardiometabolic biomarkers PLoS One 2015 10 e0119140 10.1371/journal.pone.0119140 25775249 16. Kim Y Welk GJ Braun SI Kang M Extracting objective estimates of sedentary behavior from accelerometer data: measurement considerations for surveillance and research applications PLoS One 2015 10 e0118078 10.1371/journal.pone.0118078 25658473 17. Jefferis BJ Parsons TJ Sartini C Ash S Lennon LT Wannamethee SG Does duration of physical activity bouts matter for adiposity and metabolic syndrome? A cross-sectional study of older British men Int J Behav Nutr Phys Act 2016 13 36 10.1186/s12966-016-0361-2 26980183 18. Thorp AA Healy GN Winkler E Clark BK Gardiner PA Owen N Prolonged sedentary time and physical activity in workplace and non-work contexts: a cross-sectional study of office, customer service and call centre employees Int J Behav Nutr Phys Act 2012 9 128 10.1186/1479-5868-9-128 23101767 19. Ministry of Justice, Industrial Safety and Health Act. [Online]. 2009. Available from: URL: http://www.japaneselawtranslation.go.jp/law/detail/?id=1926. Accessed 9 May 2016. 20. Honda T Chen S Kishimoto H Narazaki K Kumagai S Identifying associations between sedentary time and cardio-metabolic risk factors in working adults using objective and subjective measures: a cross- sectional analysis BMC Public Health 2014 14 1307 10.1186/1471-2458-14-1307 25526746 21. Ohkawara K Oshima Y Hikihara Y Ishikawa-Takata K Tabata I Tanaka S Real-time estimation of daily physical activity intensity by a triaxial accelerometer and a gravity-removal classification algorithm Br J Nutr 2011 105 1681 91 10.1017/S0007114510005441 21262061 22. Oshima Y Kawaguchi K Tanaka S Ohkawara K Hikihara Y Ishikawa-Takata K Classifying household and locomotive activities using a triaxial accelerometer Gait Posture 2010 31 370 4 10.1016/j.gaitpost.2010.01.005 20138524 23. National Cancer Institute: SAS Programs for Analyzing NHANES 2003-2004 Accelerometer Data [online material]. Available from http://appliedresearch.cancer.gov/nhanes_pam/. Accessed 17 Nov 2015. 24. Matthews CE Chen KY Freedson PS Buchowski MS Beech BM Pate RR Amount of time spent in sedentary behaviors in the United States, 2003-2004 Am J Epidemiol 2008 167 875 81 10.1093/aje/kwm390 18303006 25. Haskell WL Lee IM Pate RR Powell KE Blair SN Franklin BA Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association Med Sci Sport Exerc 2007 39 1423 34 10.1249/mss.0b013e3180616b27 26. Willett WC Howe GR Kushi LH Adjustmentfor total energy intake in epidemiologic studies Am J Clin Nutr 1997 65 suppl 1220S 8S 9094926 27. Healy GN Matthews CE Dunstan DW Winkler EAH Owen N Sedentary time and cardio-metabolic biomarkers in US adults: NHANES 2003-06 Eur Heart J 2011 32 590 7 10.1093/eurheartj/ehq451 21224291 28. Bankoski A Harris TB McClain JJ Brychta RJ Caserotti P Chen KY Sedentary activity associated with metabolic syndrome independent of physical activity Diabetes Care 2011 34 497 503 10.2337/dc10-0987 21270206 29. Kim J Tanabe K Yokoyama N Zempo H Kuno S Objectively measured light-intensity lifestyle activity and sedentary time are independently associated with metabolic syndrome: a cross-sectional study of Japanese adults Int J Behav Nutr Phys Act 2013 10 30 10.1186/1479-5868-10-30 23452372 30. Gennuso KP Gangnon RE Thraen-Borowski KM Colbert LH Dose–response relationships between sedentary behaviour and the metabolic syndrome and its components Diabetologia 2015 58 485 92 10.1007/s00125-014-3453-z 25476524 31. Scheers T Philippaerts R Lefevre J SenseWear-determined physical activity and sedentary behavior and metabolic syndrome Med Sci Sports Exerc 2013 45 481 9 10.1249/MSS.0b013e31827563ba 23034646 32. Bey L Hamilton MT Suppression of skeletal muscle lipoprotein lipase activity during physical inactivity: a molecular reason to maintain daily low-intensity activity J Physiol 2003 551 673 82 10.1113/jphysiol.2003.045591 12815182 33. Hamilton MT Hamilton DG Zderic TW Role of low energy expenditure and sitting in obesity, metabolic syndrome, type 2 diabetes, and cardiovascular disease Diabetes 2007 56 2655 67 10.2337/db07-0882 17827399 34. Bergouignan A Rudwill F Simon C Blanc S Physical inactivity as the culprit of metabolic inflexibility: evidence from bed-rest studies J Appl Physiol 2011 111 1201 10 10.1152/japplphysiol.00698.2011 21836047 35. Garber CE Blissmer B Deschenes MR Franklin BA Lamonte MJ Lee IM American College of Sports Medicine position stand. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults: guidance for prescribing exercise Med Sci Sports Exerc 2011 43 1334 59 10.1249/MSS.0b013e318213fefb 21694556

PMC005xxxxxx/PMC5000402.txt

==== Front Scand J Trauma Resusc Emerg MedScand J Trauma Resusc Emerg MedScandinavian Journal of Trauma, Resuscitation and Emergency Medicine1757-7241BioMed Central London 29510.1186/s13049-016-0295-3Original ResearchEvaluation of a trauma team activation protocol revision: a prospective cohort study http://orcid.org/0000-0001-7709-9385Dehli Trond +47 47 28 35 79trond.dehli@gmail.com 12Monsen Svein Arne Svein.Arne.Monsen@Helgelandssykehuset.no 3Fredriksen Knut Knut.Fredriksen@unn.no 24Bartnes Kristian Kristian.Bartnes@unn.no 251 Department of Gastrointestinal Surgery, University Hospital North Norway (UNN), 9038 Tromsø, Norway 2 Department of Clinical Medicine, UiT- The Arctic University of Norway, 9037 Tromsø, Norway 3 Department of Anesthesiology, Helgeland Hospital, 8801 Sandnessjøen, Norway 4 Division of Emergency Medical Services, UNN, 9038 Tromsø, Norway 5 Department of Cardiothoracic and Vascular Surgery, UNN, 9038 Tromsø, Norway 25 8 2016 25 8 2016 2016 24 1 10517 4 2016 22 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Correct triage based on prehospital information contributes to a better outcome for potentially seriously injured patients. In 2011 we changed the trauma team activation (TTA) criteria in our center in order to improve the high over- and undertriage properties of the protocol. Five criteria that were unable to predict severe injury were removed. In the present study, we evaluated the protocol revision by comparing over- and undertriage in the former and present set of criteria. Methods All severely injured patients (Injury Severity Score (ISS) > 15) and all patients admitted with TTA in the period of 01.01.2013 – 31.12.2014 were included in the study. We defined overtriage as the fraction of patients with TTA when ISS ≤15 and undertriage as the fraction of patients without TTA when ISS > 15. We also evaluated triage with the occurrence of emergency procedures immediately after admission. Results 324 patients were included, 164 patients had ISS>15, 287 were admitted with TTA. Over- and undertriage were 74 % and 28 % respectively. When we used emergency procedure as reference, the figures were 83 % and 15 % respectively. Undertriaged patients had significantly more neurosurgical injuries and were significantly more often transferred from an acute care hospital. Discussion Over- and undertriage are almost the same as before the criteria were revised, and higher thanrecommended levels. Conclusions Revision of the TTA criteria has not improved triage, and further measures are necessary to achieveacceptable levels. Keywords TraumaTriagePatient transferEmergency treatmentissue-copyright-statement© The Author(s) 2016 ==== Body Background Multidisciplinary trauma teams are important in trauma care by allowing timely diagnosis and treatment of unstable trauma patients [1]. Criteria-based trauma team activation (TTA) protocols are widely implemented throughout Scandinavia [2]. TTA criteria vary among trauma centers, but they usually resemble the recommendations from the American College of Surgery – Committee on Trauma (ACS-COT) [3–7]. The criteria comprise parameters of physiologic compromise, anatomic injury, and mechanism of injury (MOI), and TTA is recommended when prehospital information indicates that at least one criterion is fulfilled. A substantial overtriage (TTA despite only minor or moderate injury) is common and may reach 70 % in some centers, mostly reflecting a low specificity of the MOI criteria [5–7]. Overtriage is mainly a resource problem, as it diverts personnel from other activities in the hospital. However, undertriage (admission of severely injured without TTA) may delay timely intervention and even increase mortality [8]. Some overtriage seems to be necessary to avoid unacceptable undertriage, and the ACS-COT suggests that 50 % may be acceptable [3]. However, no scientific evidence supports this figure, and it may depend on factors related to the individual facility and trauma system. Injury patterns vary considerably between regions, and the predictive properties of triage criteria depend on the incidence of severe injuries. In Scandinavia penetrating injuries are rare [9, 10], and the frequency of severe polytrauma admissions is low, except in some large cities [11]. In order to optimize patient outcomes and resource utilization we evaluated the TTA protocol of the University Hospital of North Norway Tromsø (UNN) in 2011 using data from the period 2005–6 [12]. Five criteria were removed from the TTA-protocol. In the present study we evaluate the impact of the TTA protocol revision on protocol performance of under- and overtriage. Methods Design The study is a prospective cohort study based on the trauma registry at UNN. Study setting and local trauma registry UNN has a primary catchment population of 80.000, and is the regional trauma center (level I-II according to ACS-COT) for 480.000 people. The mainland covers an area of 107 000 km2, and is the same length as the British Isles from south to north, with 9 acute care hospitals. The closest acute care hospital is 250 km from the trauma center by road; therefore all patient transfers between hospitals are done with air ambulance. According to the regional health trust’s destination and transfer protocol, patients with injuries that exceed the resources of the acute care hospitals will be transferred to the UNN either directly from the scene or from the local hospital after initial stabilization. The emergency medical communication center (EMCC) at UNN mobilizes the trauma team before patient arrival based on defined TTA criteria (Table 1). Trauma patient transferred to the UNN are admitted with TTA if there was TTA at the local hospital and if they arrive within 24 h after the injury.Table 1 An analysis of individual criteria applied for trauma team activation Criteria category Criterion Criterio n(/reason) applied to the patient (no. of patients) Criterion applied to a severely injured patient (ISS > 15a), (no. of patients) Criterion applied to a patient receiving an emergency procedure (no. of patients) Vital parameters 1. Airway obstruction, stridor 4 2 3 2. Tachypnoe (adults, respiratory rate > 30) 14 10 3 3. Heart rate > 130 (adults) 3 0 2 4. Systolic BP <90 mmHg 9 5 4 5. Lowered level of consciousness (GCSb <13) 87 33 28 Extent of injuries 6. Flail chest 2 1 0 7. Unstable fracture of the pelvis. Fracture in two or more long bones 5 2 0 8. Traumatic amputation or crush injury above wrist/ankle 1 0 1 9. Injury in two or more body regions (head/neck/chest/abdomen/pelvis/femur/back) 61 9 8 10. Paralysis 10 8 1 11. Penetrating injury of the head/neck/chest/abdomen/pelvis/groin/back 5 0 3 12. 2. or 3. degree burn injury > 15 % body surface (children > 10 %) 5 2 3 13. Burn injury with inhalation injury 5 2 2 14. Hypothermia (core temperature <32 °C) 11 3 2 Mechanism of injury 15. Ejected from vehicle 6 4 0 16. Co-passenger dead 5 2 0 17. Trapped in wreck 9 3 1 18. Pedestrian or cyclist hit by motor vehicle 15 2 2 19. Fall from >5 m 20 10 3 20. Avalanche accident 1 0 0 Unknown criteria ETAc < 15 min 8 0 0 Trauma team leader requested TTAd 5 0 0 Anesthesiologist in ambulance helicopter requested TTAd 6 0 0 Unknown/undocumented reason for TTAd 20 0 0 The table shows the number of times an individual criteria is applied for trauma team activation based on prehospital information in potentially severely injured patients primarily admitted at the University Hospital of North Norway Tromsø during 2013–14, n = 223. Transferred patients are not included in this analysis. The two last columns shows the number of times the individual criterion correctly activated the trauma team assessed with ISS and the appearance of an emergency procedure. More than one criterion can be applied to one patient Emergency surgical procedure include endotracheal intubation, damage control thoracotomy, damage control laparotomy, extraperitoneal packing in the pelvis, revascularization of an extremity, intervention radiology, craniotomy, insertion of intracranial pressure bolt, chest tube insertion, external fracture stabilization or other emergency procedures aiming at stabilizing airway, respiration or circulation ISSa: Injury Severity Score, ISS > 15: Severely injured patient, GCSb: Glasgow Coma Score, ETAc: estimated time of arrival, TTAd: trauma team activation The trauma registry at UNN includes patients admitted with TTA, patients with penetrating injury to the head/neck/torso/extremities proximal for elbow or knee, all patients with New Injury Severity Score (NISS) >12 [13], and all patients with a head injury with abbreviated injury score (AIS) ≥3 [14]. The registry excludes patients with chronic subdural hematoma, with injuries from drowning, inhalation and strangulation, and those admitted for rehabilitation after trauma. Patients that did not have TTA but fulfilled the inclusion criteria were identified from the hospital admission charts. Trained and authorized registrars scored the AIS and calculated the ISS [14, 15]. The revision of the UNN trauma team activation protocol The former TTA protocol of the UNN was evaluated in 2011 based on data from the period 2005–6 [12]. The inclusion criteria were the same as in the present study. When we analyzed the data with an injury severity score (ISS) > 15 as indication of severe trauma [14, 15], the overtriage was as high as 71, and the undertriage was 32 %. With the occurrence of an emergency procedure as standard of reference, the overtriage was 71 and the undertriage 21 %. Three criteria in the mechanism of injury-category (“Motorcycle accident”, “Considerable deformation of vehicle compartment” and “Traffic accident with speed >60 km/h”) conferred a substantial overtriage and were removed from the TTA protocol. Another criterion was removed because it had not been used (“Convulsions”), and the criterion “Dilated or not responding pupils” was deemed unnecessary because the same patients were identified by “Lowered level of consciousness (GCS <13)”, both in order to simplify the criteria. The revised TTA criteria are listed in Table 1. The protocol also allows TTA if time to arrival of the injured patient <15 min and prehospital information on fulfillment of criteria is missing (e.g. immediate transport to the hospital is prioritized over examination). The 2011 study also showed that undertriaged patients were mainly patients who were transferred from local hospitals or patients admitted directly to the neurosurgical department. Based on the results from this study, the UNN also attempted to increase awareness of the inter-hospital transfer protocol among employees in the EMCC, in the neurosurgical department, and among trauma team leaders. Inclusion of patients In the present study, we included all patients admitted to the UNN with TTA and all patients with an ISS > 15 (with or without TTA) during the period 1.1.2013 – 31.12.2014. Patients transferred >24 h after injury were excluded. Data collection Most data could be collected from the trauma registry, but TTA criteria were recorded from the EMCC record AMIS (acute medical communication system, Nirvaco, Norway) and supplementary information was collected from the patient records of the UNN. Data from other hospitals were recorded from transfer documents. Triage Triage was considered correct when the trauma team was mobilized for primary admittance of injured patients with ISS > 15, or an injured patient transferred to our trauma center < 24 h after injury and admitted with TTA in the local hospital. Overtriage was calculated as the proportion of TTA for patients with minor or moderate injuries (ISS ≤ 15). Undertriage was calculated as the proportion of severely injured patients (ISS > 15) admitted without TTA. The ability of an individual TTA criterion to predict appropriate TTA, was given as the number of patients that fulfilled the specific criterion that also had an ISS > 15. The same triage- and criterion analyzes are done with the occurrence of emergency procedures as standard of reference. Parameters Baseline characteristics are listed in Table 2. The primary endpoint is TTA or no TTA for severely injured patients (ISS > 15), and the corresponding over- and undertriage.Table 2 Main characteristics of patients received by a trauma team or having ISSa > 15, admitted at the University Hospital of North Norway Tromsø in 2013–2014, n = 324 Male patients (proportion) 226 (69.8 %) Mean age, years (range) 41 (0–101) Median ISS* (interquartile range) 10 (2, 20) Proportion with ISS > 15** (percentage of total) 131 (40.4 %) Predominant mechanism of injury (proportion)  Penetrating 3.4 %  Blunt 96.6 % Mean length of stay (days) 6.7 Mean length of stay in intensive care unit (days) 2.0 Interhospital transfer, patients (n (proportion)) 74 (22.8 %) 30 day mortality (n (proportion)) 18 (5.6 %) 30 day mortality, patients with ISS > 15 (n (proportion)) 17 (14.9 %) ISSa: Injury Severity Score, ISS > 15: seriously injured patient * Injury Severity Score **Severely injured patient Secondary outcome parameters include triage calculated with the occurrence of an emergency procedure as standard of reference, and the type and frequency of emergency procedures. Emergency procedures are listed in Table 3.Table 3 Emergency procedures for a total of 324 trauma patients admitted with activation of the trauma team or ISS > 15* at the University Hospital of North Norway Tromsø during 2013–2014. One patient can receive more than one procedure, both in the local hospital and in the trauma center Emergency procedure Number of patients receiving a procedure among all patients (percentage of patients), n = 324 Number of patients receiving a procedure at the local hospital before transfer, n = 74 Number of undertriaged patients (no TTA** and ISS > 15) receiving a procedure, n = 37 Endotracheal intubation 57 (17 %) 27 - Damage control thoracotomy 2 (0.6 %) - - Damage control laparotomy 12 (3.7 %) 4 - Extraperitoneal pelvic packing 1 (0.3 %) - - Revascularization of an extremity - - - Intervention radiology 8 (2.5 %) - 2 Craniotomy 26 (8.0 %) - 8 insertion of intracranial pressure bolt 19 (5.9 %) - 4 chest tube insertion 26 (8.0 %) 13 2 external fracture stabilization 7 (2.2 %) 2 2 Other procedures to stabilize airways, respiration or circulation 11 (3.4 %) - - Sum of all patients with an emergency procedures 92 46 14 *ISS: Injury Severity Score, ISS > 15: seriously injured patient, **TTA: trauma team activation Statistical methods Results are given as sum, percentage, mean and median with interquartile (IQR) range. Individual criteria which are seldom used (<5 of patients) and/or with a low positive predictive value (<10 %) will be considered omitted. Comparison of triage results with the 2011 study is done with Pearson correlation-test although data from the former study is not shown [12]. Characteristics of undertriaged patients are tested with Pearsson chi-square test. Significance is assumed for p < 0.05. Ethics and publication The study was approved by the hospital’s data protection officer (case number 2013/501). Approval from the Regional Medical Research Ethical Committee was not necessary as the study was assessed as quality control and not medical research by the Ethical Committee itself (case number 2012/1912/REK Nord). Results A total of 324 patients were included. Baseline characteristics are presented in Table 2. There were 131 patients with ISS > 15 and 94 of these (72 %) were admitted with TTA. Results for triage are presented in Tables 4 and 5. There were no significant changes in over- and undertriage of the TTA protocol following the revision after the 2011 study.Table 4 Performance of the trauma team activation protocol during 2013–2014 at the University Hospital of North Norway Tromsø assessed with Injury Severity Score (ISS) Number of patients Number of patients with ISS > 15 Correct triage (TTAa and ISS >15) Undertriage (no TTAa and ISS > 15) Overtriage (TTAa and ISS < 15) Primarily admitted patients 250 81 (32 %) 72 % (58/81) 28 % (23/81) 74 % (169/227) Transferred patients 74 50 (68 %) 72 % (36/50) 28 % (14/50) 40 % (24/60) TTAa: trauma team activation, ISS > 15: severely injured patient Table 5 Performance of the trauma team activation protocol during 2013–2014 at the University Hospital of North Norway Tromsø assessed with the occurrence of emergency procedure Number of patients Number of patients with emergency procedure Correct triage (TTAa and emergency procedure) Undertriage (no TTAa and emergency procedure) Overtriage (TTAa and no emergency procedure) Primarily admitted patients 250 46 (18 %) 85 % (39/46) 15 % (7/46) 83 % (188/227) Transferred patients 74 46 (62 %) 83 % (38/46) 17 % (8/46) 38 % (28/74) Emergency procedures include damage control thoracotomy, damage control laparotomy, packing of the pelvis, revascularization of a limb, intervention radiology, craniotomy, insertion of intracranial pressure monitor, thoracostomy, external fixation of fractures for hemostasis, endotracheal intubation and other surgical procedures that aimed at stabilizing airways, respiration and circulation TTAa: trauma team activation A total of 92 patients received an emergency procedure, and the most common procedure was endotracheal intubation, followed by pleural drainage and neurosurgical interventions (Table 3). Hemostatic emergency surgery was required in <5 % of patients. Undertriaged patients The 37 patients who were undertriaged according to ISS had a median ISS of 19 (IQR 17, 25), and 8 died within 30 days (22 %). Neurosurgical injuries (head and/or cervical spine injury) were present in 32 of the 37 patients, and 84 % of these injuries were severe or critical injuries (AIS 4 or 5). This was significantly different from patients admitted with TTA (p < 0.001). Among the 32 undertriaged patients with neurosurgical injuries, 14 (44 %) patients were operated with a craniotomy or insertion of an intracranial pressure bolt (Table 3). Two patients had severe chest injuries (AIS 4), there were no other injuries classified as AIS 4 or 5 below the neck. Undertriaged patients were more often transferred from another hospital rather than admitted directly from the scene of accident (14 out of 37), compared to other included patients (p < 0.05). Triage criteria The performance of the individual criteria is shown in Table 1. Several criteria are used in less than 5 % of the admissions, but with a relatively high fraction of correct triage (criteria number 1, 3, 4, 6–8, 11–17, 20). Discussion Of the 131 seriously injured patients admitted at UNN and included in the present study, only 94 were admitted with TTA. This gives an undertriage of 28 and an overtriage of 74 % evaluated with extent of injury as the standard of reference (ISS). With occurrence of emergency procedures as reference, the undertriage was 15 and the overtriage 83 %. Based on the data from the 2011 study, we revised the TTA criteria and reinforced the transfer protocol. The present re-evaluation of our TTA protocol shows no significant change in over- or undertriage compared to the 2011 study. Overuse of the TTA is a matter of resource utilization, and the only reason to accept overtriage is that some overtriage might be necessary to avoid undertriage. Our results are in line with several other studies that report an overtriage of ca 70 % [5–9]. Some authors report overtriage as high as 90 % [16], and in this context our results may still be acceptable. Indeed, some overtriage may help to increase clearance of patients from the emergency department, and unnecessary members of the trauma team may be dismissed early after the initial survey. In hospitals with a low TTA frequency overtriage also represents a potential for training the trauma team, which may improve team performance with seriously injured patients for which the TTA may make an important difference. The low number of patients makes it difficult to draw conclusions about individual criteria for TTA, and even though some criteria were seldom used, the proportion of correctly triaged patients was high. However, based on the findings in the present study, we suggest that further changes in the present TTA criteria will not help to reduce overtriage. Another way to limit unnecessary use of limited resources may be to introduce a two-tiered TTA, with a low threshold for mobilizing a smaller team based on MOI information from the prehospital services, and the full team only when alarming vital signs or anatomical injuries are reported. Reports indicate a reduction of resource use and undertriage with a two tiered TTA [17, 18]. However, this has not been addressed by the present study. A low undertriage is important for a favorable patient outcome, and there are reports indicating increased mortality in undertriaged patients [8]. According to the ACS-COT, an undertriage of up to 5 % is acceptable [3]. Our results are far above this figure, and we believe that two factors might explain some of the undertriage. First, at the UNN trauma registrars continuously screen all admissions to the surgical departments and assess for inclusion in the trauma registry. This might contribute to a better detection of undertriaged patients compared to other centers. Second, referring and receiving doctors may have exchanged critical patient information and agreed that a patient has been adequately stabilized before transfer and a new TTA at the trauma center is not needed. This is formally violating the existing transfer protocol in our center, but may represent sound clinical judgment and can in part explain some of the undertriaged transfers. Neurosurgical patients dominate the group of undertriaged patients that have been directly admitted from the scene. These were scored to an ISS > 15, but the mechanism of injury and physiologic status, including level of consciousness, did not fulfill any TTA criteria. It seems possible to have a significant head injury, without being identified as seriously injured based on clinical findings. A recent study has shown that elderly patients with traumatic brain injury might have a higher Glasgow Coma Score (GCS) than younger patients, making our GCS-based triage criteria less reliable [19]. We have experienced an unexpected reduction in severe trauma admissions over the last ten years in our trauma center (unpublished data). In the present study, we report a decrease in TTA from 382 in 2005–2006 to 287 in 2013–2014, and likewise a reduction in number of seriously injured patients (ISS > 15) from 161 to 131 [12]. The results in this study cannot explain this general reduction, but we believe that it is not related to change in inclusions of patients. The recent introduction of a local trauma registry secures an even more complete inclusion of relevant trauma patients than before, and it is unlikely that the registrars now are missing more severe trauma patients than in the former study. Interestingly, the 30-day mortality was relatively unchanged with 6.6 in the 2011 study and 5.6 % in the present publication, supporting the notion that the inclusion process has remained unchanged. Traditionally TTA has been evaluated against the ISS, with an ISS >15 indicating severe injury. AIS and ISS is an anatomical grading of injuries, with good correlation to mortality. However, the scoring of injuries is retrospectively derived, most often after discharge or death of the patients. AIS and ISS are therefore not available as an aid in decision-making in the prehospital setting or for the trauma team at admission. Furthermore, the most precise TTA-criteria for correct triage is the physiologic criteria, and there is a small paradox in assessing physiologic TTA criteria against anatomic injury grading. Anatomic and physiologic parameters often, but not always, correlate well with each other after injury. For this reason, we also evaluated TTA against immediate use of emergency procedures that aim to secure vital functions in physiologically instable patients. The main reason for a TTA is to identify such potentially life-threatening instability, and to restore airway control, and adequate ventilation and circulation. We believe, as other authors [20, 21], that the traditional ISS-based evaluation of TTA appropriateness should be complemented by the clinically recognized need for stabilizing emergency interventions. When we evaluated TTA against immediate emergency procedures, we found a slightly better undertriage, but still a high overtriage. The number of severely injured patients is low, and this explains the low number of emergency procedures we have found. Hence, it has not been justified to have dedicated trauma surgeons on call at the UNN. Instead, the center relies on the general surgeons on call, who do both elective and acute care surgery on a daily basis. This indicates that specific trauma training of the surgeons is necessary, since real life experience with emergency procedures is limited. The recently suggested revised national trauma plan for Norway describes hemostatic emergency courses as mandatory for surgeons and their teams in all hospitals admitting trauma patients, in order to compensate for the limited experience in trauma surgery [22]. The weakness of our study is the low number of patients, which precludes an analysis of individual TTA criteria. The sample size was given by the time period we collected data, and a further extension of the time would both delay the evaluation, and it could also introduce other variables that would make it impossible to evaluate the effect of the criteria revision. Furthermore, we believe that not all criteria are assessed for every admission; if one criterion is fulfilled, it has no consequence to record more criteria, as the trauma team will be activated by the first criterion alone. Another potential limitation is the theoretical possibility that the results are influenced by an over-grading especially in head injuries, but all ISS scoring was done by trained and authorized registrars. We therefore believe that this possibility is unlikely. The major strengths of the study are the prospective design and the complete inclusion of undertriaged patients. Conclusion Both overtriage and undertriage remains high despite changes in TTA criteria aimed to improve TTA protocol precision. Both indicators are still higher than desired. The study indicates a lack in the TTA criteria’s ability to identify patients with severe head injuries. The number of potential lifesaving surgical procedures is low. The revision of TTA criteria has not improved triage, and further studies are needed to find better ways to improve it. Acknowledgements None. Funding The study has not received any funding. Availability of data and materials Data was obtained from electronic patient records. These data are protected by law, and not openly available. Authors’ contributions TD wrote the protocol, registered data, made analysis and drafted the manuscript. SAM contributed on the protocol, registered data and contributed on the manuscript. KF contributed on the protocol, registered data and contributed on the manuscript. KB supervised the study, contributed on the protocol and the manuscript. All authors have read and acknowledged the final version of the manuscript. Authors’ information All authors are involved in the medical management of trauma patients. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate The study was approved by the hospital’s data protection officer (case number 2013/501). Approval from the Regional Medical Research Ethical Committee was not necessary as the study was assessed as quality control and not medical research by the Ethical Committee itself (case number 2012/1912/REK Nord). Informed patient consent was not necessary assessed by the Ethical Committee. ==== Refs References 1. MacKenzie EJ Rivara FP Jurkovich GJ Nathens AB Frey KP Egleston BL A national evaluation of the effect of trauma-center care on mortality N Engl J Med 2006 354 366 78 10.1056/NEJMsa052049 16436768 2. Kristiansen T Søreide K Ringdal K Rehn M Krüger AJ Reite A Trauma systems and early management of severe injuries in Scandinavia: review of the current state current state Injury 2010 41 442 52 10.1016/j.injury.2009.05.027 3. The American College of Surgeons Resources for optimal care of the injured patient 2006 [Internet]. 2006th ed. J. Am. Coll. Surg 2006 Chicago IL American College of Surgeons Committee on Trauma 4. Larsen KT Uleberg O Skogvoll E Differences in trauma team activation criteria among Norwegian hospitals Scand J Trauma Resusc Emerg Med 2010 18 21 10.1186/1757-7241-18-21 20406456 5. Uleberg O Vinjevoll OP Eriksson U Aadahl P Skogvoll E Overtriage in trauma - What are the causes? Acta Anaesthesiol Scand 2007 51 1178 83 17714579 6. Clemmesen ML Rytter S Birch K Lindholt JS Jensen SS Troelsen S [Should high-energy traumas always result in a trauma team call?]. Ugeskr Laeger 2006 168 2916 20 7. Kann SH Hougaard K Christensen EF Evaluation of pre-hospital trauma triage criteria: a prospective study at a Danish level I trauma centre Acta Anaesthesiol Scand 2007 51 1172 7 17711567 8. Rehn M Eken T Krüger AJ Steen PA Skaga NO Lossius HM Precision of field triage in patients brought to a trauma centre after introducing trauma team activation guidelines Scand J Trauma Resusc Emerg Med 2009 17 1 10.1186/1757-7241-17-1 19134177 9. Krüger AJ Hesselberg N Abrahamsen GT Bartnes K [When should the trauma team be activated?] Tidsskr den Nor lægeforening 2006 126 1335 7 10. Meisler R Thomsen AB Abildstrøm H Guldstad N Borge P Rasmussen SW Triage and mortality in 2875 consecutive trauma patients Acta Anaesthesiol Scand 2010 54 218 23 10.1111/j.1399-6576.2009.02075.x 19817720 11. Wisborg T Castrèn M Lippert A Valsson F Wallin CJ Training trauma teams in the Nordic countries: an overview and present status Acta Anaesthesiol Scand 2005 49 1004 9 10.1111/j.1399-6576.2005.00742.x 16045663 12. Dehli T Fredriksen K Osbakk SA Bartnes K Evaluation of a university hospital trauma team activation protocol Scand J Trauma Resusc Emerg Med 2011 19 18 10.1186/1757-7241-19-18 21439095 13. Osler T Baker SP Long W A modification of the injury severity score that both improves accuracy and simplifies scoring J Trauma 1997 43 922 5 10.1097/00005373-199712000-00009 9420106 14. Association for the Advancement of Automotive Medicine Abbreviated Injury Scale 2005 2008 Barrington, IL, USA Update 2008 15. Baker S O’Neill B Haddon W Long W The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care J Trauma-Injury 1974 14 187 196 10.1097/00005373-197403000-00001 16. Stordahl H, Passas E, Hopland A, Nielsen EW. Nine out of ten trauma calls to a Norwegian hospital are avoidable: a retrospective analysis. BMC Emerg Med. 2015;15. 17. Jenkins P Rogers J Kehoe A Smith JE An evaluation of the use of a two-tiered trauma team activation system in a UK major trauma centre Emerg Med J 2015 32 364 7 10.1136/emermed-2013-203402 24668398 18. Rehn M Lossius HM Tjosevik KE Vetrhus M Østebø O Eken T Efficacy of a two-tiered trauma team activation protocol in a Norwegian trauma centre Br J Surg 2012 99 2 199 208 10.1002/bjs.7794 22190166 19. Kehoe A Rennie S Smith JE Glasgow coma scale is unreliable for the prediction of severe head injury in elderly trauma patients Emerg Med J 2014 32 613 5 10.1136/emermed-2013-203488 25280479 20. Bressan S Franklin KL Jowett HE King SK Oakley E Palmer CS Establishing a standard for assessing the appropriateness of trauma team activation : a retrospective evaluation of two outcome measures Emerg Med J 2015 32 716 721 10.1136/emermed-2014-203998 25532103 21. Falcone RA Haas L King E Moody S Crow J Moss A A multicenter prospective analysis of pediatric trauma activation criteria routinely used in addition to the six criteria of the American College of Surgeons J Trauma Acute Care Surg 2012 73 377 84 10.1097/TA.0b013e318259ca84 22846943 22. No Title [Internet]. [cited 2016 Jan 1]. Available from: http://traumeplan.no/wp-content/uploads/2015/04/Traumesystem-i-Norge-2015.pdf. Accessed 24 Aug 2016.

PMC005xxxxxx/PMC5000403.txt

==== Front Genome BiolGenome BiolGenome Biology1474-75961474-760XBioMed Central London 103910.1186/s13059-016-1039-4MethodStreamlined analysis of duplex sequencing data with Du Novo Stoler Nicholas 1Arbeithuber Barbara 2Guiblet Wilfried 1Makova Kateryna D. kdm16@psu.edu 3Nekrutenko Anton aun1@psu.edu 141 Graduate Program in Bioinformatics and Genomics, The Huck Institutes for the Life Sciences, Penn State University, 505 Wartik Lab, University Park, PA 16802 USA 2 Institute of Biophysics, Johannes Kepler University, Linz, Austria 3 Department of Biology, Penn State University, 310 Wartik Lab, University Park, PA 16802 USA 4 Department of Biochemistry and Molecular Biology, Penn State University, University Park, PA USA 26 8 2016 26 8 2016 2016 17 1 18022 1 2016 5 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Duplex sequencing was originally developed to detect rare nucleotide polymorphisms normally obscured by the noise of high-throughput sequencing. Here we describe a new, streamlined, reference-free approach for the analysis of duplex sequencing data. We show the approach performs well on simulated data and precisely reproduces previously published results and apply it to a newly produced dataset, enabling us to type low-frequency variants in human mitochondrial DNA. Finally, we provide all necessary tools as stand-alone components as well as integrate them into the Galaxy platform. All analyses performed in this manuscript can be repeated exactly as described at http://usegalaxy.org/duplex. Electronic supplementary material The online version of this article (doi:10.1186/s13059-016-1039-4) contains supplementary material, which is available to authorized users. Keywords Duplex sequencingLow frequency polymorphism discoveryNext generation sequencingGenomic data analysishttp://dx.doi.org/10.13039/100000051National Human Genome Research InstituteHG005542Nekrutenko Anton http://dx.doi.org/10.13039/100000057National Institute of General Medical SciencesGM116044issue-copyright-statement© The Author(s) 2016 ==== Body Background The term “genetic variation” is often used to imply allelic combinatorics within a diploid organism such as humans or Drosophila. Yet the majority of organisms in the biosphere are not diploid (prokaryotes and viruses), and even those that are include non-diploid genomes such as mitochondria and chloroplasts. Identification of genetic variants—e.g., single nucleotide polymorphisms (SNPs) and small indels—is especially challenging in non-diploid systems due to the lack of a simple “homozygote-or-heterozygote” expectation: a heterozygous site may have not just two but multiple allelic variants, with frequencies ranging anywhere from 0 to 1 [1, 2]. Because high-throughput sequencing technologies exhibit considerable amounts of noise [3], it becomes increasingly difficult to reliably call variants with frequencies below 1 % [4–9]. In these situations increased sequencing depth does not improve the predictive power but instead introduces additional noise. This complicates the identification of low-frequency variants that is becoming critically important in a variety of applications. For example, humans have numerous disease-causing mitochondrial variants where the disorder penetrance is proportional to the allele frequency [10]. Because dramatic shifts in allele frequency can occur during mitochondrial bottleneck during oogenesis, a disease-causing variant present at a very low frequency in the mother may increase in frequency in the child to exhibit a disease phenotype. The lack of cures for diseases caused by mitochondrial DNA mutations and the recent regulatory approval of tri-parental in vitro fertilization by the UK House of Commons makes it critical to identify low-frequency variants in the human mitochondrial genome [11]. Other examples illustrating the importance of discovering low-frequency genome alterations include tracking mutational dynamics in viral genomes, malignant lesions, and somatic tissues [12, 13]. Today the vast majority of strategies for the identification of low-frequency sequence variants rely on next-generation sequencing technologies. Noise reduction in these approaches ranges from simple base-quality filtering to complex statistical strategies incorporating instrument and mapping errors [4, 7, 14]. However, there is still considerable uncertainty about alternative alleles with frequencies below 1 %. For example, Fig. 1 illustrates the number of potential polymorphisms observed within the human mitochondrial genome as a function of the allele frequency cutoff. At 1 % there is an average of three sites [7], while at 0.75 % the number surpasses 10, and, finally, around 0.1 % almost all sites appear polymorphic. Clearly, the majority of these sites are false-positives but how does one know for certain? Potentially, highly sensitive techniques with a high dynamic range, such as droplet digital PCR [15, 16], can be used to validate each site, but it would quickly become prohibitively expensive and laborious to perform this on hundreds or thousands of sites.Fig. 1 The relationship between the minor allele frequency (maf) threshold (x-axis) and the total number of variable sites (y-axis) detected by [7]. Lowering the MAF threshold leads to an exponential increase in the number of variable positions. The image was generated by applying variable MAF thresholds to data from 156 human samples and plotting the average number of variable sites at a given MAF threshold. The line thickness corresponds to the 95 % confidence interval around the mean value An approach that offers a potential solution is duplex sequencing [17]. This recently developed method was designed to increase sequencing accuracy by over four orders of magnitude. Duplex sequencing uses randomly generated barcodes to uniquely tag each molecule in a sample. The tagged fragments are then PCR amplified prior to the preparation of a sequencing library, creating fragment families characterized by unique combinations of barcodes at both 5′ and 3′ ends (a conceptually similar primer ID approach [12] allows tagging of cDNA fragments at the 5′ end only). A family contains multiple reads, each originating from a single input DNA fragment. A legitimate sequence variant will thus be present in all reads within a family. In contrast, sequencing and amplification errors will manifest themselves as “polymorphisms” within a family and so can be identified and removed. A consensus can be called from these read families. The consensus of all the reads originating from the same strand reduces errors originating from sequencing and PCR amplification. Then, comparing consensus sequences from complementary strands can identify early PCR errors. Despite the fact that duplex sequencing promises great advances, the methods for both experimental and computational aspects of this technique are still evolving. In fact, the latter is lagging as it is based on alignment to a reference genome, which is disadvantageous for several reasons. The use of a reference genome biases results toward that reference, affecting studies using de novo assembly or studies examining indels or other alleles that diverge far enough from the reference to cause alignment difficulties. The current analysis method also removes a large (and potentially useful) fraction of the original data due to stringent filters and uses suboptimal tools for variant identification. Here we describe an alternative analysis strategy which removes reliance on a reference sequence, preserves a higher proportion of the input reads, and can be deployed as a standalone application or as a part of the Galaxy system. We demonstrate the application of this approach by validating rare variants in the human mitochondrial genome. Results and discussion A reference-free approach Our approach is outlined in Fig. 2. First, paired reads generated from a duplex sequencing experiment are merged into families. This is performed by sorting according to the barcode. Each fragment is expected to be represented by two single-stranded families corresponding to each strand. These two single-stranded families are expected to have the same unique tags but in the opposite order: the α tag from one single-stranded family will be the β tag in the other (also see Fig. 1 in [17]). In order to group single-stranded families from the same fragment together, we normalize the order of the concatenation to produce a “canonical barcode” (a concatenated string consisting of α and β tags), which will be identical for both strands. The order of the canonical barcode is determined by a simple string comparison. Sorting the output groups the reads so that the two families constituting each duplex will be adjacent, with the read pairs separated by strand.Fig. 2 The Du Novo approach. First, reads tagged with identical barcodes are grouped into strand-specific families. Reads within each family are aligned and single-stranded consensus sequences (SSCSs) are generated. Finally, the SSCSs are reduced into duplex consensus sequences (DCSs) Next, the reads in each single-stranded family are aligned to themselves and these alignments are used to call the single-stranded consensus sequence (SSCS). First, a threshold is applied, requiring a user-specified number of reads to produce a consensus (three by default). The consensus calling is conducted by determining the majority base at each position. If no base is in the majority, “N” is used. Positions with gaps are considered in the same way as bases. Quality filtering is performed at this stage: bases with a Phred quality score [18] lower than a user-specified threshold are not counted (20 is used by default). For positions with gaps, a quality score is calculated by considering the qualities of eight neighboring bases. The calculated score is a weighted average, with the weight decreasing linearly with distance from the gap. Finally, a duplex consensus is called using the two SSCSs. The SSCSs are aligned using the Smith–Waterman algorithm [19] and then each pair of bases is compared. If the bases agree, that base is used in that position to generate duplex consensus. If they disagree, the International Union of Pure and Applied Chemistry (IUPAC) ambiguity code for the two bases is used. Gap and non-gap characters produce an “N”. In the end the above approach reduces the initial set of sequencing reads to a collection of duplex consensus sequences (DCSs; as the duplex sequencing experiments are performed with paired-end reads, the output of the procedure also consists of pairs corresponding to forward and reverse double-stranded consensuses). DCSs are then filtered (i.e., sequences with ambiguous nucleotides can be removed or trimmed), mapped against the reference genome, and realigned to normalize gap-containing regions and the resulting alignments are used to call variants. In this scenario variants are expected to have the full spectrum of allele frequencies between 0 and 1 and do not follow a diploid expectation. For that reason we use variant callers capable of dealing with this limitation such as the Naive Variant Caller (NVC) [20] or Freebayes [21]. Finally, variant calls are post-processed to compute the strand bias (using formulae from Guo et al. [22]). This approach is implemented in a pipeline relying exclusively on open-source software (https://github.com/galaxyproject/dunovo and accessible through the Galaxy system). We termed this approach Du Novo—for duplex sequencing de novo assembly-based calling. Du Novo reliably identifies very low frequency variants First, we evaluated the performance of Du Novo by applying it to a dataset generated from a simulated mixing experiment. The advantage of performing the simulation is that the “truth” is known explicitly. We randomly generated 21 “heteroplasmies” by modifying human mitochondrial sequence. This altered version of the mitochondrial genome was then “mixed” with unmodified reference sequence at a ratio of 1:10,000 (thus, each “heteroplasmy” in this mix has the frequency of 0.0001) and a duplex experiment was simulated on the mixture. This was done by randomly generating 2500 fragments from the altered sequence and 25,000,000 fragments from unmodified reference, adding barcodes, and performing in silico PCR and sequencing (see “Methods”). The polymerase error rate in PCR and sequencing was set at 0.1 % per base. After applying Du Novo to the simulated reads and aligning the DCSs to the mitochondrial reference, the median read depth was 166,574×. Next, we identified all variable sites and filtered them using a series of minor allele frequency (MAF) thresholds and requiring a minimum DCS coverage of 10,000. The relationship between MAF thresholds and the numbers of false positives and false negatives is shown in Additional file 1: Figure S1. Du Novo correctly identifies 20 of the 21 variants with no false positives. The remaining variant was present at a frequency of 0.00004 (likely a result of random fluctuation), along with 46 false positives with an equal or higher MAF. Comparison with original approach: Du Novo replicates published estimates To assess the performance of our method on real-world data and to compare it head-to-head with the original approach of Kennedy et al. [23], we re-analyzed a recently published dataset by Schmitt and colleagues [13] using both methods. In [13] the authors employed duplex sequencing to identify a rare mutation at the ABL1 locus responsible for resistance to the chronic myeloid leukemia therapeutic compound imatinib. The resistance is conferred by the presence of G-to-A substitutions within the ABL1 coding region resulting in an E279K amino acid replacement. This substitution is present in a small sub-clonal subset of cells at an ~1 % frequency. The dataset (Sequence Read Archive accession SRR1799908) contains 6,921,891 read pairs representing 1,468,089 unique tag combinations (potential families; Table 1).Table 1 Characteristics of ABL1 and SC8 duplex sequencing experiments Number of ABL1 SC8 Read pairs 6,921,891 17,385,100 Unique tags 1,467,768 2,100,705 Unique αβ configurations 748,411 1,148,444 Unique αβ configurations with 1 read pair 677,069 884,295 Unique αβ configurations with ≥3 read pairs 60,333 222,823 Unique βα configurations 743,669 1,092,748 Unique βα configurations with 1 read pair 672,946 832,875 Unique βα configurations with ≥3 read pairs 60,032 140,486 Unique αββα 24,313 140,485 Unique αββα with ≥3 read pairs on both strands 20,746 109,999 Reads within αββα families with ≥3 read pairs on both strands 2,156,105 8,636,692 First, we analyzed this dataset with Du Novo. Requiring each family to contain at least three reads reduced this number to 120,365 SSCSs and reconciling these into DCSs further reduced this number to 20,746 DCSs constructed from 2,083,140 read pairs (the remaining 6,921,891 − 2,083,140 = 4,838,751 were represented by families with less than three reads and were omitted; see Additional file 2: Figure S2). Mapping DCSs to the reference human genome showed the G-to-A substitution with frequency varying from 1.28 to 1.31 % depending on the variant caller (NVC [20] and FreeBayes [21], respectively) but irrespective of the mapper used (BWA-MEM [24] or BWA [25]). Next, we repeated this experiment with the published duplex sequencing pipeline [23]. This produced 1.29 and 1.31 % frequencies at the G-to-A substitution site for NVC and FreeBayes, respectively. Thus, the allele frequency estimates were essentially identical between the two approaches. Du Novo produced a higher depth at the variable site: 1099 for our method versus 618 for the published pipeline [23]. However, at such low allele frequencies even formidable coverage results in a relatively small proportion of reads supporting the minor allele. For example, in the case of this analysis the minor allele (“A”) is supported by 14 duplex consensuses from the total of 1099, resulting in a MAF of 1.28 %. Yet each of these 14 families is in turn derived from multiple starting reads ranging from a minimum of 5 to a maximum of 102 (Fig. 3a), providing additional support for the reliability of the minor allele calls.Fig. 3 Distribution of family sizes (number of reads per family) supporting A and G alleles on both strands (plus and minus) for a site 130,872,141 in the ABL1 dataset and b site 13,708 in the SC8 dataset Using Du Novo to call low-frequency heteroplasmies at mitochondrial DNA After ensuring the adequate performance of Du Novo on the ABL1 data, we applied it to the identification of low-frequency variants in human mitochondrial DNA (mtDNA). Previously, we have reported 174 point heteroplasmies identified from the analysis of mtDNA in 39 mother–child pairs (a total of 156 samples = 39 mothers × 2 tissues + 39 children × 2 tissues [7]). We chose family SC8 as it displays significant variability across samples and individuals. This family contains two heteroplasmic sites—at positions 7607 and 13,708. According to our published results [7], the MAFs at site 7607 are 0.7, 1.1, 0.0, and 0.0 % in mother’s buccal tissue, mother’s blood, child’s buccal tissue, and child’s blood, respectively. The corresponding MAFs at site 13,708 are 0.0, 0.0, 2.2, and 1.6. To verify these frequencies, we performed the duplex sequencing experiment using genomic DNA extracted from SC8 child buccal tissue in which mtDNA has been enriched via long-range PCR as previously described in [7]. We started with 17,385,100 read pairs that contained 2,100,704 unique tags and were assembled into 82,230 DCSs. The estimated allele frequency at position 13,708 was 0.53 %, a figure substantially lower than the 2.2 % estimated previously [7]. The coverage at this site was 1138 with six reads representing the minor allele (“A”). To check the reliability of this call we estimated strand bias (SB; using formula 1 from [22]) for all sites with MAF ≥0.5 %. There were 20 sites (excluding 13,708) with MAF ranging from 0.51 to 21.2 % and with SB values ranging from 0.94 to 6.08 (the lower the value, the less SB there is at a site; 0 is an ideal value [22]). SB = 0.01 at site 13,708, which is outside of the SB distribution for all other variable sites in our sample, strongly suggesting that this is the only true heteroplasmy in this sample. In addition, examining individual DCSs at this site indicates that each of them is generated from a large number of original reads (Fig. 3b) confirming this polymorphism, albeit at a significantly lower frequency. The utility of SSCSs In the SC8 experiment described above, we estimated the MAF at site 13,708 to be 0.53 %—a much lower value compared with the original one (2.2 %) obtained from re-sequencing [7]. The likely cause of this deviation lies in the design of the duplex experiment. In this study we performed duplex sequencing not directly on mtDNA but instead on products of a long-range PCR (see “Methods”). In this particular case this is unavoidable as the samples are obtained by a minimally invasive “cheek swab”, resulting in a very low concentration of mtDNA. The core issue is that complementary strands of the resulting PCR products (the starting material for our duplex sequencing experiment) can randomly pair after amplification, forming heteroduplexes and leading to an underestimation of MAFs when using DCSs only (Additional file 3: Figure S3). To test whether this indeed is the cause of MAF underestimation, we performed variant calling using SSCSs instead of DCSs and obtained a MAF of 1.7 % (strand bias = 0.02 and depth = 4548), a value much closer to the 2.2 % reported in the original publication. Thus, although the background error frequency is higher for SSCSs in comparison with DCSs [17] in certain situations, such as experiments using ampliconic DNA, the use of SSCSs for polymorphism detection may be preferable to obtain more accurate allele frequencies. Loss of data as a result of sequencing errors in duplex tags One of the fundamental weaknesses of duplex sequencing is the fact that the majority of families in a duplex experiment contain only a single read pair (Additional file 4: Figure S4). This eliminates a substantial amount of otherwise useful data from the analysis, contributing to the inefficiency of the current protocol. To understand the potential sources of read loss, we examined individual stages of the duplex analysis process. This information is compiled in Table 1 and is based on the re-analysis of both previously published data (ABL1 data) [13] and results generated in our laboratory (the SC8 dataset described above). Both cases feature a large number of initial read pairs and unique tags. However, these numbers are rapidly reduced by requiring at least three reads within each single-stranded family. Combining SSCSs into DCSs also greatly reduces the number of useful sequences since both strands must be present and meet the three-read threshold. One potential explanation for the large number of families with only one read pair is sequencing errors within duplex tags. Each barcode with an error will almost certainly be unique, creating an entirely new apparent family with only one member. The number of reads with an erroneous barcode may be a minority but this can still result in the number of families with erroneous barcodes being very high (a majority). The fraction of erroneous barcodes (r) can be expressed in the following form: 1 r=1–1–El where E is the per-base error rate and l is the barcode length (in this case 24 as it is a combination of α and β tags, each of which is 12 nucleotides). Here, E is a cumulative error rate taking into account the chance of a mutation at every cycle of PCR plus the sequencing reaction. The cumulative error rate can be calculated from the error rate at each stage using the same equation (Eq. 1), this time using E as the error rate per base per stage, l as the number of stages (number of PCR cycles plus 1 for the sequencing reaction), and r as the cumulative error rate. Even assuming a low per-stage error rate of 0.1 %, this gives a cumulative error rate of about 3 %. Using this in Eq. 1 again, we obtain the fraction of barcodes expected to contain an error to be 52.5 %: 2 r=1–1–0.0324≈0.525 Now, suppose in a hypothetical duplex experiment ten initial fragments of DNA were ligated with α and β adapters (a unique α and β for each of the ten fragments) and the subsequent PCR amplification and Illumina sequencing process produced 100 read pairs (10 pairs per original fragment). If there are no errors, these 100 read pairs should be recognized as members of ten duplex families during the analysis stage. If we now factor in the erroneous barcode rate of ~52 % calculated above, one would observe 62 total families: ten real families and 52 artifactual families consisting of a single read pair. This phenomenon increases the total number of families by reducing the read count within legitimate families—a trend apparent in real data (Additional file 2: Figure S2). Furthermore, the relationship between the number of single-read families and the total number of reads can serve as a proxy for the error rate. For example, in the SC8 experiment there were 1,717,170 single read families and 17,385,100 total read pairs. Assuming that all single read families are byproducts of sequencing errors within duplex tags, this gives 1,717,170/17,385,100 = 0.098 as the fraction of erroneous barcodes (r). With l = 24 we can solve Eq. 1 for E obtaining an estimate of ~0.4 % for the cumulative error rate. To test this reasoning we simulated duplex experiments with different error rates. The starting distribution of family sizes was constant in each case, with 1.20 % of fragments producing a family with only one read. With an error rate of zero, the proportion of output families which were composed of a single read was, as expected, precisely 1.20 % (Additional file 4: Figure S4), meaning no excess beyond those with a natural family size of one. When the error rate was raised to 0.1 % per base per cycle of PCR/sequencing reaction, 75.5 % of output families were composed of a single read. This meant that 74.3 % of families were artifacts consisting of a read originating from a fragment that produced multiple reads. Instead of being grouped with its sibling reads, each of these instead was grouped by itself because of an error in the barcode. While this was only a simulation and the above calculations make a number of simplifying assumptions, they nevertheless highlight the significance of sequencing errors within tags as one of the main causes of data loss. We are currently developing a family reconstruction approach that would allow mismatches in tags and is expected to significantly reduce the number of single read families. Interactive analysis of duplex data The underlying components of the Du Novo process are distributed as an open source software and can be used from the command line (https://github.com/galaxyproject/dunovo). However, to increase the number of potential users we also make Du Novo accessible through the Galaxy system (http://usegalaxy.org). Figure 4 illustrates all stages of the duplex analysis workflow. This example begins with fastq datasets generated by an Illumina machine that are used as inputs in the Du Novo pipeline. Initially, reads are processed to identify and count duplex tags (Make families). Reads having identical tags (families) are aligned (Align families) and alignments are reduced to DCSs (Make consensus reads). The DCSs are trimmed to remove ambiguous nucleotides (Sequence Content Trimmer), converted to fastq format (this is because DCSs are reported as fasta datasets; Combine FASTA and QUAL), and mapped to the reference genomes (in this example with both BWA and BWA-MEM). BAM datasets produced by mappers are combined (MergeSamFiles) and realigned (BamLeftAlign) and variable sites are identified with the Naive Variant Caller (NVC). A Variable Call Format (VCF) dataset generated by NVC is processed by Variant Annotator, which tabulates allele frequencies and strand bias values. Finally, the data are filtered on MAF (≥0.5 %) and strand bias (<1). This workflow is available at https://usegalaxy.org/u/aun1/w/duplex-analysis-from-reads. The most computationally demanding portion of the workflow is the alignment of reads within each family (Align Families). For instance, processing of 6,921,891 read pairs comprising the ABL1 dataset [9] took an average of 0.004 s per pair or approximately 9 h of wall time on a 16-CPU cluster node. One of the advantages of using Galaxy at https://usegalaxy.org for the analysis of duplex sequencing data is that its underlying infrastructure relies on high-performance resources provided by the Texas Advanced Computing Center (TACC) and the Extreme Science and Engineering Discovery Environment (XSEDE), making it possible to perform analyses of multiple duplex datasets by multiple users simultaneously.Fig. 4 A complete workflow implementing the Du Novo approach to variant discovery from duplex sequence data. It is accessible from http://usegalaxy.org/duplex Conclusions The continuing drop in the price of massively parallel sequencing will expand the use of the duplex technique and will amplify the need for a scalable analysis solution such as Du Novo reported here. Our approach allows the use of both single- and double-stranded consensuses for variant discovery depending on the experimental design and is parallelized to take advantage of the advanced high performance compute infrastructure. By allowing our tools to be used both from the command line and through the Galaxy interface we hope to reach a wide audience of computational and experimental researchers. Methods Duplex sequencing protocol used for human mitochondrial amplicons Two overlapping mtDNA regions (each ~9 kb, representing the entire mitochondrial genome) were amplified from sample SC8C1-k1169-A*B (DNA extracted from buccal swabs of the child of family SC8 collected under IRB 30432EP), using the primer pairs L*2817 + H*11570 and L10796 + H3370 and mixed at equimolar quantities, as described previously [7, 26]. Amplicons (2 μg) were sheared to ~550 bp and purified using 1.6 volumes of Agencourt AMPure XP beads (Beckman Coulter). Duplex sequencing libraries were prepared as described in Kennedy et al. [23] with several minor modifications. Briefly, T-tailed adapters were prepared by hybridization of MWS51 and MWS55, followed by extension, and a restriction digest with TaaI (HypCH4III) at 60 °C for 16 h. Adapters were purified by precipitation with two volumes of absolute ethanol and 0.5 volumes of 5 M NH4OAc. The hybridized PCR amplicon was end-repaired with the End-Repair Enzyme Mix provided in the Illumina TruSeq Kit according to the manufacturer’s protocol and A-tailed and the adapter was ligated with 1800 units of T4 ligase (NEB) with 20× molar excess at 16 °C for 30 min. Amplified tag families were generated from 15 attomoles of adapter-ligated amplicon by 23 cycles of PCR (the optimal cycle number was evaluated by real-time PCR). The library was quantified with the KAPA Library Quantification Kit (Kapa Biosystems) according to the manufacturer’s instructions. Sequencing was performed on an Illumina MiSeq platform using 301-bp paired-end reads. Construction of read families Read pairs are grouped into families according to the random tags which constitute the first 12 bp of each read using the Du Novo pipeline either in Galaxy or through the command line. For each pair, we first construct a barcode which is the concatenation of the two tags from the two reads. Then the reads are sorted according to this compound barcode. Single-stranded families from the same fragment will have the same 12-bp tags but in the opposite order: the α tag from one family will be the β tag in the other. In order to group single-stranded families from the same fragment together, we normalize the order of the concatenation to produce a “canonical barcode” which will be identical for both strands. The order of the canonical barcode is determined by a simple string comparison. Then the original order of the tags is recorded in a separate field. Sorting the output groups the reads so that the two families constituting each duplex will be adjacent, with the read pairs separated by strand. Aligning families and consensus calling The reads in each single-stranded family are aligned to themselves using a script calling the MAFFT multiple sequence aligner [27]. These alignments were used to call the SSCSs. First, a threshold is applied, requiring a specified number (default = 3) of reads to produce a consensus. Then, the consensus calling is performed by determining the majority base at each position. If no base is in the majority, “N” is used. Positions with gaps are considered in the same way as bases. Quality filtering is done at this stage: bases with a PHRED quality score lower than a user-given threshold are not counted (default = 20). For positions with gaps, a quality score is calculated by considering the quality scores of the eight nearest bases. The calculated score is a weighted average, with the weight decreasing linearly with distance from the gap. Finally, duplex consensus sequences are called using the two SSCSs. The two sequences are aligned using the Smith–Waterman algorithm (using an existing C implementation from https://code.google.com/archive/p/swalign/) and then each pair of bases is compared. If the bases agree, that base is used in that position. If they disagree, the IUPAC ambiguity code for the two bases is used. Gap and non-gap characters produce an “N”. If a SSCS has no matching opposite strand consensus, the user may choose to include the single-stranded consensus in the output, direct it to a separate file, or discard it. In silico mixture experiment We randomly generated 21 heteroplasmies and inserted them into the human mitochondrial genome (Revised Cambridge Reference Sequence (rCRS), NC_012920.1) at a spacing of at least 600 bp from each other and from the chromosome ends (the genome is circular but its textual representation is not). This in silico mutated sequence is referred to as mt-mut to distinguish it from the unmodified reference, which we would call mt-ref. Next, 600-bp fragments were randomly generated using wgsim (version 0.3.1-r13) with the error and mutation rate set to 0. For mt-mut and mt-ref we generated 2500 and 25,000,000 of such fragments, respectively. Each of the fragments was tagged on each end with a random, 12-bp barcode and a 5-bp linker sequence. Each was then subjected to in silico PCR and sequencing to create a family of reads descended from the same fragment. To determine the size of the family, a random number was chosen from an empirically determined distribution, with a peak at nine reads. A phylogenetic tree was simulated for the reads by starting at the last PCR cycle and coalescing backward, randomly joining branches based on the probability of two reads sharing an ancestor at that cycle (2-cycle). Thirty cycles of PCR were simulated. Then, PCR polymerase errors were simulated by introducing random errors at each cycle, accumulating errors from each parent molecule. The error rate was 0.001 probability of an error per base, per cycle. Indels were given a 0.15 fraction of the errors and a 0.3 probability of extension per base. Finally, a pair of 250-bp reads was generated from each final fragment sequence. Sequencing polymerase errors were introduced at the same rates as PCR polymerase errors. Quality scores were not simulated and set to a PHRED value of 40. The strandedness of each read pair was determined according to the initial two potential daughters of the original fragment it was descended from. Duplex consensus reads were created from these simulated reads using Du Novo with three reads required per single-stranded consensus and base quality filtering turned off (PHRED threshold of 0). The reads were aligned to the mitochondrial genome (rCRS) with BWA-MEM and filtered for alignments with a minimum mapping quality (MAPQ) of 20. Error rate-singleton correlation simulation In silico duplex sequencing of the human mitochondrial reference sequence (rCRS) was performed as described above but with 10,000 400-bp fragments and 100-bp final reads to save computational time. Then, the reads were processed with the first part of the Du Novo pipeline, creating a strand-independent barcode from each read pair. Then, the total number of unique barcodes was counted and the fraction of those that were present only once. This was performed once for each error rate setting. Additional files Additional file 1: Figure S1. Receiver operating characteristic (ROC) for Du Novo detecting 21 artificial heteroplasmies in a simulated duplex sequencing experiment. Shown are true positives versus false positives detected using different minor allele frequency thresholds, in steps of 0.00001 (the depth of coverage threshold was held constant at 10,000×). At the bottom left, no heteroplasmies at all are detected at a threshold MAF of 0.00016. The first variant is detected at a MAF of 0.00015, with no false positives. Continuing upward, no false positives are detected while increasing true positives are found until the upper left corner at a MAF of 0.00008, with 20 true positives and no false positives. Then, increasing false positives are found with no gain in true positives until the last true single-nucleotide variant (SNV) is found at a MAF of 0.00004, with 46 false positives also observed at that threshold. (PNG 23 kb) Additional file 2: Figure S2. Distribution of reads per family in ABL1 (a) and SC8 (b) datasets. (JPEG 114 kb) Additional file 3: Figure S3. Here there are two distinct types of mitochondrial genomes: carrying A and G. Because the population of genomes is enriched via PCR, heteroduplex formation takes place, skewing frequency estimates performed using DCSs. If this PCR-derived DNA is now used as the starting material for a duplex sequencing experiment, the heteroduplex molecules will manifest themselves as having an N base at this site (because Du Novo interprets disagreements as Ns during consensus generation). So, DCSs produced from this dataset will have A, G, and N at the polymorphic site. Yet, SSCSs will only have A and G. Thus, SSCS will give a more accurate estimate of the allele frequency at this site in this particular case. (JPEG 196 kb) Additional file 4: Figure S4. Effect of errors on the number of single-read families. Duplex sequencing was simulated using different values for the PCR/sequencing polymerase error rates. In each case, 10,000 400-bp fragments were generated from the mitochondrial reference sequence. After simulating the duplex method, the number of reads observed for each unique barcode was counted. Shown are the fraction of families with only one read versus the polymerase error rate. (PNG 26 kb) Abbreviations DCSduplex consensus sequence MAFminor allele frequency mtDNAmitochondrial DNA NVCNaïve Variant Caller PCRpolymerase chain reaction SBstrand bias SNPSingle nucleotide polymorphism SSCSsingle-stranded consensus sequence Acknowledgements We are grateful to Kristin Eckert, Suzanne Hile, and Howard Fescemeyer for their advice regarding establishing duplex sequencing in our laboratory. Marcia Shu-Wei Su has performed the enrichment of mitochondrial DNA via long range PCR. Nathan Coroar provided assistance with tuning our software for the optimal utilization of high performance compute infrastructure. Michael Schmitt provided advice on the organization of the ABL1 dataset. Funding This project was supported by NIH Grants U41 HG005542 to AN and R01 GM116044 to KDM. Additional funding is provided by Huck Institutes for the Life Sciences at Penn State and, in part, under a grant with the Pennsylvania Department of Health using Tobacco Settlement Funds. The Department specifically disclaims responsibility for any analyses, interpretations, or conclusions. Availability of data and materials All software described here is accessible from http://usegalaxy.org/duplex under the terms of the BSD Open Source License. This publication is based on version 0.4 of Du Novo accessible from https://github.com/galaxyproject/dunovo/releases/tag/v0.4 (doi:10.5281/zenodo.57256). The ABL1 dataset from Schmitt and colleagues is available from the Short Read Archive under accession SRR1799908. The SC8 dataset is available from the Short Read Archive under accession SRR3749606. Authors’ contributions NS developed software and performed initial analyses; BA performed the duplex sequencing procedure; WG performed data analysis; KM and AN wrote the paper and designed the Galaxy workflow. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Ethics approval and consent to participate All experimental procedures described herein comply with the principles of the Helsinki Declaration. Buccal swabs from subject SC8C1 have been collected under IRB protocol number 30432EP. SC8C1 is a part of a larger study performed within the framework of this IRB protocol. In this study patients at the pediatric outpatient clinic (located the Pennsylvania State University Hershey Medical Center in Harrisburg, Pennsylvania) were approached by experienced coordinators and invited to participate. Samples were collected from participants providing verbal consent in the clinic. Only date of birth and date of collection were associated with each sample. No personally identifiable information was recorded. ==== Refs References 1. Hodgkinson A Idaghdour Y Gbeha E Grenier J-C Hip-Ki E Bruat V Goulet J-P de Malliard T Awadalla P High-resolution genomic analysis of human mitochondrial RNA sequence variation Science. 2014 344 413 5 10.1126/science.1251110 24763589 2. Zhang P Samuels DC Lehmann B Stricker T Pietenpol J Shyr Y Guo Y Mitochondria sequence mapping strategies and practicability of mitochondria variant detection from exome and RNA sequencing data Brief Bioinform. 2016 17 224 32 10.1093/bib/bbv057 26249222 3. Schirmer M Ijaz UZ D’Amore R Hall N Sloan WT Quince C Insight into biases and sequencing errors for amplicon sequencing with the Illumina MiSeq platform Nucleic Acids Res. 2015 43 e37 7 10.1093/nar/gku1341 25586220 4. Li M Stoneking M A new approach for detecting low-level mutations in next-generation sequence data Genome Biol. 2012 13 R34 10.1186/gb-2012-13-5-r34 22621726 5. Li M Schröder R Ni S Madea B Stoneking M Extensive tissue-related and allele-related mtDNA heteroplasmy suggests positive selection for somatic mutations Proc Natl Acad Sci U S A. 2015 112 2491 6 10.1073/pnas.1419651112 25675502 6. Goto H Dickins B Afgan E Paul IM Taylor J Makova KD Nekrutenko A Dynamics of mitochondrial heteroplasmy in three families investigated via a repeatable re-sequencing study Genome Biol. 2011 12 R59 10.1186/gb-2011-12-6-r59 21699709 7. Rebolledo Jaramillo B Su MS-W Stoler N McElhoe JA Dickins B Blankenberg D Korneliussen TS Chiaromonte F Nielsen R Holland MM Paul IM Nekrutenko A Makova KD Maternal age effect and severe germ-line bottleneck in the inheritance of human mitochondrial DNA Proc Natl Acad Sci U S A. 2014 111 15474 9 10.1073/pnas.1409328111 25313049 8. Quail MA Smith M Coupland P Otto TD Harris SR Connor TR Bertoni A Swerdlow HP Gu Y A tale of three next generation sequencing platforms: comparison of Ion Torrent. Pacific Biosciences and Illumina MiSeq sequencers BMC Genomics. 2012 13 1 10.1186/1471-2164-13-341 22214261 9. Ross MG Russ C Costello M Hollinger A Lennon NJ Hegarty R Nusbaum C Jaffe DB Characterizing and measuring bias in sequence data Genome Biol 2013 14 5 R51 10.1186/gb-2013-14-5-r51 23718773 10. Wallace DC Chalkia D Mitochondrial DNA genetics and the heteroplasmy conundrum in evolution and disease Cold Spring Harb Perspect Biol. 2013 5 a021220 10.1101/cshperspect.a021220 24186072 11. Dimond R Social and ethical issues in mitochondrial donation Br Med Bull. 2015 115 173 82 10.1093/bmb/ldv037 26351372 12. Jabara CB Jones CD Roach J Anderson JA Swanstrom R Accurate sampling and deep sequencing of the HIV-1 protease gene using a Primer ID Proc Natl Acad Sci U S A. 2011 108 20166 71 10.1073/pnas.1110064108 22135472 13. Schmitt MW Fox EJ Prindle MJ Reid-Bayliss KS True LD Radich JP Loeb LA Sequencing small genomic targets with high efficiency and extreme accuracy Nat Methods. 2015 12 423 5 10.1038/nmeth.3351 25849638 14. Kim SY Lohmueller KE Albrechtsen A Li Y Korneliussen T Tian G Grarup N Jiang T Andersen G Witte D Jorgensen T Hansen T Pedersen O Wang J Nielsen R Estimation of allele frequency and association mapping using next-generation sequencing data BMC Bioinformatics. 2011 12 231 10.1186/1471-2105-12-231 21663684 15. Hindson CM Chevillet JR Briggs HA Gallichotte EN Ruf IK Hindson BJ Vessella RL Tewari M Absolute quantification by droplet digital PCR versus analog real-time PCR Nat Methods. 2013 10 1003 5 10.1038/nmeth.2633 23995387 16. Miotke L Lau BT Rumma RT Ji HP High sensitivity detection and quantitation of DNA copy number and single nucleotide variants with single color droplet digital PCR Anal Chem 2014 86 5 2618 24 10.1021/ac403843j 24483992 17. Schmitt MW Kennedy SR Salk JJ Fox EJ Hiatt JB Loeb LA Detection of ultra-rare mutations by next-generation sequencing Proc Natl Acad Sci U S A. 2012 109 14508 13 10.1073/pnas.1208715109 22853953 18. Ewing B Green P Base-calling of automated sequencer traces using phred. II. Error probabilities Genome Res. 1998 8 186 94 10.1101/gr.8.3.186 9521922 19. Smith TF Waterman MS Identification of common molecular subsequences J Mol Biol. 1981 147 195 7 10.1016/0022-2836(81)90087-5 7265238 20. Blankenberg D Von Kuster G Bouvier E Baker D Afgan E Stoler N Galaxy Team Taylor J Nekrutenko A Dissemination of scientific software with Galaxy ToolShed Genome Biol 2014 15 403 10.1186/gb4161 25001293 21. Garrison E, Marth G. Haplotype-based variant detection from short-read sequencing. arXiv:1207.3907 arXiv.org 2012. 22. Guo Y Li J Li C-I Long J Samuels DC Shyr Y The effect of strand bias in Illumina short-read sequencing data BMC Genomics. 2012 13 1 1 10.1186/1471-2164-13-1 22214261 23. Kennedy SR Schmitt MW Fox EJ Kohrn BF Salk JJ Ahn EH Prindle MJ Kuong KJ Shen J-C Risques R-A Loeb LA Detecting ultralow-frequency mutations by Duplex Sequencing Nat Protoc. 2014 9 2586 606 10.1038/nprot.2014.170 25299156 24. Li H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. arXiv:1303.3997 arXiv.org 2013. 25. Li H, Durbin R. Fast and accurate short read alignment with Burrows-Wheeler Transform. Bioinformatics. 2009;1–7. 26. Dickins B Rebolledo Jaramillo B Su MS-W Paul IM Blankenberg D Stoler N Makova KD Nekrutenko A Controlling for contamination in re-sequencing studies with a reproducible web-based phylogenetic approach Biotechniques. 2014 56 134 6 10.2144/000114146 24641477 27. Katoh K Standley DM MAFFT multiple sequence alignment software version 7: improvements in performance and usability Mol Biol Evol. 2013 30 772 80 10.1093/molbev/mst010 23329690

PMC005xxxxxx/PMC5000404.txt

==== Front Cardiovasc DiabetolCardiovasc DiabetolCardiovascular Diabetology1475-2840BioMed Central London 44110.1186/s12933-016-0441-2Original InvestigationCommon dysregulated pathways in obese adipose tissue and atherosclerosis Moreno-Viedma V. veronica.morenoviedma@meduniwien.ac.at 1Amor M. melina.amor@meduniwien.ac.at 1Sarabi A. a.sarabi@gmx.net 1Bilban M. martin.bilban@meduniwien.ac.at 2Staffler G. guenther.staffler@affiris.com 3Zeyda M. maximilian.zeyda@meduniwien.ac.at 14Stulnig T. M. +43 (0)1 40400 61027thomas.stulnig@meduniwien.ac.at 11 Christian Doppler Laboratory for Cardio-Metabolic Immunotherapy and Clinical Division of Endocrinology and Metabolism, Department of Medicine III, Medical University of Vienna, Waehringer Guertel 18-20, 1090 Vienna, Austria 2 Department of Laboratory Medicine & Core Facility Genomics, Core Facilities, Medical University of Vienna, Vienna, Austria 3 AFFiRiS AG, Vienna, Austria 4 Department of Pediatrics and Adolescent Medicine, Clinical Division of Pediatric Pulmonology, Allergology and Endocrinology, Medical University of Vienna, Vienna, Austria 26 8 2016 26 8 2016 2016 15 1 1202 7 2016 17 8 2016 © The Author(s) 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background The metabolic syndrome is becoming increasingly prevalent in the general population that is at simultaneous risk for both type 2 diabetes and cardiovascular disease. The critical pathogenic mechanisms underlying these diseases are obesity-driven insulin resistance and atherosclerosis, respectively. To obtain a better understanding of molecular mechanisms involved in pathogenesis of the metabolic syndrome as a basis for future treatment strategies, studies considering both inherent risks, namely metabolic and cardiovascular, are needed. Hence, the aim of this study was to identify pathways commonly dysregulated in obese adipose tissue and atherosclerotic plaques. Methods We carried out a gene set enrichment analysis utilizing data from two microarray experiments with obese white adipose tissue and atherosclerotic aortae as well as respective controls using a combined insulin resistance-atherosclerosis mouse model. Results We identified 22 dysregulated pathways common to both tissues with p values below 0.05, and selected inflammatory response and oxidative phosphorylation pathways from the Hallmark gene set to conduct a deeper evaluation at the single gene level. This analysis provided evidence of a vast overlap in gene expression alterations in obese adipose tissue and atherosclerosis with Il7r, C3ar1, Tlr1, Rgs1 and Semad4d being the highest ranked genes for the inflammatory response pathway and Maob, Bckdha, Aldh6a1, Echs1 and Cox8a for the oxidative phosphorylation pathway. Conclusions In conclusion, this study provides extensive evidence for common pathogenic pathways underlying obesity-driven insulin resistance and atherogenesis which could provide a basis for the development of novel strategies to simultaneously prevent type 2 diabetes and cardiovascular disease in patients with metabolic syndrome. Electronic supplementary material The online version of this article (doi:10.1186/s12933-016-0441-2) contains supplementary material, which is available to authorized users. Keywords Cardiovascular diseasesDiabetes mellitus, type 2Insulin resistanceAtherosclerosisPathway analysisFederal Ministry of Economy, Family and Youth and the National Foundation for Research, Technology and Developmentissue-copyright-statement© The Author(s) 2016 ==== Body Background The metabolic syndrome is a worldwide public health challenge with a prevalence above 20 % within adults in Western societies [1]. This disorder is based on several factors including visceral obesity, hypertension, dyslipidemia and hyperglycemia conferring a fivefold increased risk for type 2 diabetes and twofold for cardiovascular disease compared to the non-affected population [2, 3]. A chronic low-grade inflammation in response to obesity originating from the white adipose tissue has been identified as the link between obesity, insulin resistance, type 2 diabetes and cardiovascular disease [4–7]. Due to the simultaneous occurrence of insulin resistance and atherosclerosis, a considerable number of pathogenic pathways might be shared in the development of both conditions. In the past decades traditional approaches have been confined to identify changes in the expression levels of individual genes between two different conditions, however the integration and comprehension of large amounts of data remained a challenge [8–10]. Recently several methods and bioinformatic tools have been developed to perform pathways analyses out of gene expression data enabling to manage, integrate and interpret them with a more holistic view and a biological meaning [11, 12]. Gene set enrichment analysis (GSEA) provides the possibility to compare data with different gene set databases of interest and reports group of genes associated with the same biological function or common pathways [12, 13]. Hence such analyses allow a more general picture on dysregulation compared to analyses focusing on individual genes. Despite a number of investigations focusing on alterations leading to the development of either insulin resistance or atherosclerosis, there is no record in the literature systematically looking for dysregulated pathways common to insulin resistance and atherosclerosis in the same individual. Due to the concurring risk of type 2 diabetes and cardiovascular disease, the elucidation of dysregulated pathways in adipose tissue and atherosclerotic plaques should be based on an animal model that mirrors human disease by simultaneously developing adipose tissue inflammation/insulin resistance and atherosclerosis. Therefore, the aim of this study was the identification and analysis of common dysregulated pathways in obesity-induced adipose tissue inflammation and atherosclerotic plaque formation to elucidate interrelations in the concurrent development of type 2 diabetes and cardiovascular disease. Increasing our understanding on simultaneous dysregulation may indicate common molecular mechanisms that underlie type 2 diabetes and cardiovascular disease to facilitate novel preventive and therapeutic strategies in patients with metabolic syndrome. In this study we performed a pathway analysis using GSEA software with data from an own microarray experiment carried out with gonadal white adipose tissue (AT) and aortae (AO) samples from a combined insulin resistance/atherosclerosis mouse model established in our lab [14]. With this inbred mouse model, we identified common pathways in the onset of adipose tissue inflammation/insulin resistance and atherosclerosis taking advantage of the simultaneous development of both of them in individual mice while avoiding genetic variation. In conclusion, this study provides a highly valuable set of information which may be used by multiple researchers to generate hypotheses on the common development of insulin resistance and atherosclerosis. Methods Animals and diets A combined insulin resistance/atherosclerosis mouse model established in our laboratory was used as described [14]. At 9 weeks of age, male LDL-receptor knockout mice (Ldlr−/−) were placed for 20 weeks either on normal chow (NC; V1126-000, Ssnif, Soest, Germany) or diabetogenic diet (DDC; D09071704, Research Diets Inc.) (Additional file 1: Table S1). Animals were sacrificed and AT and AO were collected and immediately snap frozen in liquid nitrogen. All mice were housed in a specific pathogen-free facility with a 12 h light/dark cycle. Mice had free access to food and water. The protocol fully complied with the guidelines on accommodation and care of animals formulated by the European Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes and was approved by the local ethics committee for animal studies and the Austrian Federal Ministry for Science and Research. Atherosclerosis quantification En-face staining was used to determine atherosclerotic plaque formation as described [15]. Briefly, after sacrificing the mice the thorax was opened and the aorta was removed and cleaned removing all fat and connective tissue. Subsequently, the aorta was excised 2 mm above aortic root and below iliac bifurcation, opened longitudinally, pinned to silicone plates with acupuncture needles (asia-med, Suhl, Germany) and fixed overnight in 4 % paraformaldehyde, 5 % sucrose, 20 μM EDTA (pH 7.4). Atherosclerotic plaques were stained with Sudan IV for 15 min and destained with 75 % ethanol. Pictures were taken with a Sony Z-1000 camera and atherosclerotic lesion area was assessed by a person blinded to the samples by using ImageJ software. Microarray analysis The frozen tissue samples were homogenized in TRIzol® reagent (Invitrogen/Life Technologies, Carlsbad, CA, USA) and processed based on manufacturer’s instructions for the RNA isolation. Total RNA (1 μg) was used for GeneChip analysis, for AT preparation six individual samples were used, whereas in the case of AO each of the three individual samples were pooled in three groups due to a limited quantity of material. Terminal labeled cDNA, hybridization to genome-wide Mouse Gene 2.0 ST Gene Chips and scanning of the arrays were carried out according to the manufacture’s indications (Affymetrix). Robust Multiarray Average (RMA) signal extraction, normalization and filtering were performed as described (http://www.bioconductor.org) [16, 17]. The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus (TM. Stulnig et al. 2016) and are accessible through GEO Series accession number GSE76812 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE76812). Pathways analysis For the GSEA analysis, output primary raw data from the AT or AO microarray experiments was set up as the expression data set in accordance with GSEA indications and uploaded to the software collectively with the phenotype labels, chip annotations and either Biocarta, KEGG, Reactome or Hallmark gene sets downloaded from the Molecular Signatures Database (MSigDB) from the GSEA website. Subsequently the program was run with 1000 permutations and gene set as a permutation type, obtaining all dysregulated pathways, their respective normalized enrichment score (NES) and the enrichment plot for each microarray experiment. Statistical significances were set at a nominal p < 0.05 and false discovery rate q < 25 %. According to the GSEA directions, the pathways upregulated by NC were taken as pathways downregulated by DDC. Afterwards, the leading edge analysis function was executed to determine all those genes that significantly contribute to the dysregulation of the pathway of interest, also called leading genes. GSEA software was also used for perform the leading genes heat maps, where higher expression values are represented with red and lower expression values with dark blue. All Venn diagrams were made with the free access Venn Diagram Plotter from the Pacific Northwest National Laboratory. Statistical analyses Data are given as mean ± SEM. Dietary treatment differences were estimated by unpaired two-tailed Student t test. Results Diabetogenic diet induces obesity and atherosclerosis in Ldlr−/−mice Male Ldlr−/− mice were fed DDC or NC following the combined cardiometabolic mouse model previously established and characterized in detail in our lab [14]. The key parameters of this model were reevaluated for the mice in this study. We observed a significantly higher body weight at all time points (Fig. 1a) as well an elevated final AT weight (Fig. 1 b) in Ldlr−/− mice fed DDC for 16 weeks compared to those on NC animals. Atherosclerotic plaque formation as analyzed by en-face staining revealed markedly enhanced atherosclerotic lesions in Ldlr−/− mice fed with DDC (Fig. 1c, d). Together these results point to the simultaneous development of considerable obesity and atherosclerosis in Ldlr−/− mice fed with DDC used in this study, reflecting published findings in this mouse model.Fig. 1 Diabetogenic diet induces obesity and atherosclerosis in Ldlr−/− mice. Ldlr−/− mice were placed on either diabetogenic diet (DDC) or normal chow (NC) for 16 weeks: a mean body weight during dietary treatment (n = 12); b mean gonadal white adipose tissue (AT) weight after 16 weeks on dietary treatment (n = 12); c representative images of en-face Sudan IV stained aortae after 16 weeks of indicated treatments; d atherosclerotic lesion quantification after 16 weeks of dietary treatment (n = 3) GSEA revealed common dysregulated pathways in AT and AO from obese mice In this study, the output primary raw data from the AT or AO microarrays were uploaded into the GSEA with the aim to determine significantly dysregulated pathways in either tissue by the effect of the supplied diets (DDC or NC) and possible overlaps in terms of dysregulated pathways between both tissues. The analyses were compared with Biocarta, KEGG, Reactome or Hallmark as gene sets, obtaining upregulated pathways by DDC or NC for either AT or AO (Fig. 2). To identify common dysregulated pathways between AT and AO, we proceeded to match the upregulated or downregulated pathways for both tissues in each of the four analyses finding a considerable overlap in all the gene sets. The highest proportion of overlapping pathways was obtained for the Hallmark gene set (Fig. 3). Hence, the Hallmark gene set analysis was chosen for more comprehensive examinations, which allowed us to elucidate common pathways related to obesity induced AT inflammation and atherosclerosis.Fig. 2 Study design and pathways GSEA results overview. The data from gonadal white adipose tissue (AT) and atherosclerotic aorta (AO) microarrays were analyzed by GSEA software with Biocarta, KEGG, Reactome or Hallmark as gene sets. Upregulated or downregulated pathways in AT or AO by the effect of the diabetogenic diet (DDC) are shown Fig. 3 Venn diagrams plots for the dysregulated pathways according to GSEA. Upregulated or downregulated pathways for obese gonadal white adipose tissue (AT) and atherosclerotic aorta (AO) are shown in the Venn diagrams for each of the four analyses: a biocarta, b KEGG, c reactome, d Hallmark. Common pathways and percentage of overlap are included in the central region of each diagram Hallmark gene set involves many well characterized biological processes combining numerous pathways. The dysregulated pathways from the GSEA with the Hallmark gene set are shown in Table 1 and were ranked by AT microarray NES, the main statistic to evaluate gene set enrichment. Several well-known as well as novel pathways potentially involved in type 2 diabetes and atherosclerosis were dysregulated in AT and AO from obese and atherosclerotic animals, respectively, compared with healthy counterparts. Upregulation of inflammatory genes with positive NES occurred in both AT and AO in obesity and atherosclerosis. With respect to a pathway downregulated in obese adipose tissue (negative NES), we selected the oxidative phosphorylation pathway due to experimental and clinical evidence linking mitochondrial alterations to type 2 diabetes and atherosclerosis.Table 1 Dysregulated pathways from obese adipose tissue and atherosclerotic aortae after Hallmark-GSEA analysis DPa NESd NOM p-Vale FDRf IRPg MRPh Name ATb AOc AT AO AT AO AT AO x x 2.43 1.35 <0.001 0.047 0.000 0.098 E2F targets x x 2.39 1.74 <0.001 <0.001 0.000 0.004 G2 M checkpoint x x 2.15 2.27 <0.001 <0.001 0.000 0.000 x Allograft rejection x x 2.12 2.33 <0.001 <0.001 0.000 0.000 x Inflammatory response x x 2.08 1.99 <0.001 <0.001 0.000 0.001 x IL6 JAK STAT3 signaling x x 1.99 1.90 <0.001 <0.001 0.000 0.001 Epithelial mesenchymal transition x x 1.97 1.97 <0.001 <0.001 0.000 x TNFA signaling via NFKB x x 1.88 1.75 <0.001 <0.001 0.000 Mitotic spindle x x 1.86 2.10 <0.001 <0.001 0.000 0.000 Complement x x 1.86 2.10 <0.001 <0.001 0.000 x Kras signaling up x x 1.81 1.41 0.002 0.045 0.001 Protein secretion x x 1.72 1.89 <0.001 <0.001 0.004 0.001 Apoptosis x x 1.63 –1.34 0.001 0.026 0.008 0.091 x MTORC1 signaling x x 1.60 2.35 0.001 <0.001 0.011 0.000 x Interferon gamma response x x 1.58 1.61 0.021 0.021 0.013 0.011 x Angiogenesis x x 1.50 1.48 0.006 0.012 0.027 0.035 Coagulation x 1.48 0.012 0.030 Unfolded protein response x 1.43 0.015 0.047 MYC targets V1 x 1.41 0.024 0.051 x Androgen response x 1.36 0.027 0.073 x IL2 STAT5 signaling x x 1.34 1.68 0.024 <0.001 0.084 P53 pathway x x −1.45 −1.72 0.016 <0.001 0.036 0.004 x Peroxisome x −1.56 <0.001 0.000 Myogenesis x x −1.65 −2.14 <0.001 <0.001 0.000 0.000 x Bile acid metabolism x x −1.86 –1.46 <0.001 0.003 0.000 0.038 Xenobiotic metabolism x x –2.17 –2.40 <0.001 <0.001 0.000 0.000 x Fatty acid metabolism x x –2.44 −2.15 <0.001 <0.001 0.000 0.000 x Oxidative phosphorylation x x −3.06 −2.38 <0.001 <0.001 0.000 0.000 x Adipogenesis x 2.08 <0.001 0.000 x Interferon alpha response x 1.72 <0.001 0.005 Heme metabolism x 1.63 0.001 0.010 Apical junction x 1.50 0.007 0.031 x Hypoxia x –1.77 <0.001 0.002 Spermatogenesis Commonly dysregulated pathways in AT and AO are shown in italics Dysregulated pathways are sorted by descending NES in AT aDysregulated pathways bObese white adipose tissue cAtherosclerosic aortae dNormalized enrichment score eNominal p value fFalse discovery rate gInflammation related pathways hMetabolism related pathways Single gene analysis of inflammatory response and oxidative phosphorylation pathways To investigate changes at the single gene level involved in the inflammatory response and oxidative phosphorylation pathways, we first proceeded with a leading edge analysis, which allowed us to identify those genes that are significantly affecting the dysregulation of each pathway, called leading genes. The leading genes are represented in the heatmaps (Figs. 4, 5) and in the pathway enrichment plots (Additional files 2, 3). The enrichment score that gives an idea about the overall regulation of the pathway is represented by the score at the peak of the enrichment plot. We carried out a leading edge analysis for the inflammatory response pathway in obese adipose tissue (Fig. 4a) and in atherosclerotic aortae (Fig. 4b) as well as for the oxidative phosphorylation pathways also in obese adipose tissue (Fig. 5a) and in atherosclerotic aortae (Fig. 5b). For the obese adipose tissue, a definite leading genes regulation profile was observed in all the samples in both analyzed pathways, denoted by the clear color patterns in the heatmaps (Figs. 4a, 5a). In contrast to the AT results, the leading gene regulation profile either in the inflammatory response or in the oxidative phosphorylation pathways was not quite homogenous in all the AO samples (Figs. 4b, 5b). Nevertheless, the oxidative phosphorylation pathway was clearly downregulated in obese/atherosclerotic animals as shown in Table 1, heatmaps and enrichment plots.Fig. 4 Leading genes heat map expression for the inflammatory response pathway in AT and AO. Leading genes of the inflammatory response pathway in: a gonadal white adipose tissue (AT) and b aorta (AO) are represented in the heatmaps. Upregulated genes are represented in red and downregulated genes are represented in blue Fig. 5 Leading genes heat map expression for the oxidative phosphorylation pathway in AT and AO. Leading genes of the oxidative phosphorylation pathway in: a gonadal white adipose tissue (AT) and b aorta (AO) are represented in the heatmaps. Upregulated genes are represented in red and downregulated genes are represented in blue We next tested for a possible overlap between the leading genes of each pathway in AT and AO. The percentage of matched leading genes from AT and AO in the inflammatory response pathway was 36 % (Fig. 6a), while it was even 54 % in oxidative phosphorylation (Fig. 6b). Additionally, the lists of common leading genes between both tissues for the inflammatory response pathway (Table 2) or oxidative phosphorylation pathway (Table 3) were obtained together with the corresponding rank metric score per gene which represents their importance in the dysregulation of the pathway with the uppermost listed genes being the most influential in the pathway in AT (Upregulated, Table 2; Downregulated, Table 3). For the inflammatory response pathway Il7r, C3ar1, Tlr1, Rgs1 and Semad4d were the highest ranked genes. Among the most highly ranked in the oxidative phosphorylation pathway were Maob, Bckdha, Aldh6a1, Echs1 and Cox8a. Non-overlapping genes regulated solely in AT or AO for inflammatory response and oxidative phosphorylation pathways are listed in the Additional file 1: Tables S2, S3, respectively. In addition, the common leading genes of the other pathways dysregulated in both analyzed tissues that have not been further analyzed in this study are presented in the Additional file 1: Table S4.Fig. 6 Venn diagrams plot of leading genes involved in the enrichment score of inflammatory response and phosphorylation pathways. Percentage of the overlap between the leading genes from the inflammatory response (a) and oxidative phosphorylation (b) pathways in obese white adipose tissue (AT) and in atherosclerotic aortae (AO) Table 2 Common genes from obese adipose tissue and atherosclerotic aortae involved in the inflammatory response pathway Rank metric score Gene symbol Gene name ATa AOb 4.554 2.000 Il7r Interleukin 7 receptor 3.713 2.000 C3ar1 Complement component 3a receptor 1 3.652 0.928 Emr1 Adhesion G protein-coupled receptor E1 3.489 0.579 Tlr1 Toll-like receptor 1 3.298 1.000 Rgs1 Regulator of G-protein signalling 1 3.285 0.576 Semad4d Semaphorin 4D 3.205 0.940 Cybb Cytochrome b-245 heavy chain 3.175 0.445 Kcnj2 Inward rectifier potassium channel 2 3.111 0.779 C5ar1 C5a anaphylatoxin chemotactic receptor 1 3.037 1.000 Cd48 CD48 antigen 2.961 1.000 Pik3r5 Phosphoinositide-3-kinase, regulatory subunit 5, p101 2.626 0.992 Ptafr Platelet-activating factor receptor 2.422 0.894 Ptpre Protein tyrosine phosphatase, receptor type, E 2.364 0.716 P2rx4 Purinergic receptor P2X, ligand-gated ion channel, 4 2.346 2.000 Timp1 Tissue inhibitor of metalloproteinase 1 2.197 2.000 Il10ra Interleukin 10 receptor, alpha 1.692 0.519 Plaur Plasminogen activator, urokinase receptor 1.594 0.496 Clec5a C-type lectin domain family 5, member a 1.535 0.361 Gna15 Guanine nucleotide binding protein, alpha 15 1.452 0.922 Ccl2 Chemokine (C–C motif) ligand 2 1.303 0.273 Ifnar1 Interferon (alpha and beta) receptor 1 1.129 0.957 Ccl7 Chemokine (C–C motif) ligand 7 1.090 1.000 Csf3r Colony stimulating factor 3 receptor (granulocyte) 0.921 0.460 Il1r1 Interleukin 1 receptor, type I 0.905 0.697 Ccl5 Chemokine (C–C motif) ligand 5 0.882 0.782 Icam1 Intercellular adhesion molecule 1 (CD54), human rhinovirus receptor 0.806 0.277 Slc11a2 Solute carrier family 11, member 2 Dysregulated genes are sorted by descending rank metric score in AT aObese gonadal white adipose tissue bAtherosclerotic aortae Table 3 Common genes from obese adipose tissue and atherosclerotic aortae involved in the oxidative phosphorylation pathway Rank metric score Gene symbol Gene name ATa AOb −2.2950 −0.4060 Maob Monoamine oxidase B −2.0570 −0.3260 Bckdha Branched chain keto acid dehydrogenase E1, alpha polypeptide −2.0250 −0.8910 Aldh6a1 Aldehyde dehydrogenase 6 family, member A1 −2.0020 −0.2060 Echs1 Enoyl CoA hydratase, short chain, 1, mitochondrial −1.7080 −0.3710 Cox8a Cytochrome c oxidase subunit 8a −1.4980 −0.7180 Phyh Phytanoyl–CoA hydroxylase −1.4240 −0.1680 Cox4i1 Cytochrome c oxidase subunit IV isoform 1 −1.3880 −0.3800 Aco2 Aconitase 2, mitochondrial −1.3400 −0.2640 Acat1 Acetyl-Coenzyme A acetyltransferase 1 −1.3350 −0.2220 Ndufb8 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 8 −1.3200 −0.1620 Ndufv1 NADH dehydrogenase (ubiquinone) flavoprotein 1 −1.3060 −0.3470 Atp5d ATP synthase, H+ transporting, mitochondrial F1 complex, delta subunit −1.2750 −0.2330 Ndufs7 NADH dehydrogenase (ubiquinone) Fe–S protein 7 −1.2670 −0.4190 Ndufa4 NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 4 −1.2370 −0.2430 Gpx4 Glutathione peroxidase 4 −1.2260 −0.4390 Pdha1 Pyruvate dehydrogenase E1 alpha 1 −1.2060 −0.4440 Sdhc Succinate dehydrogenase complex, subunit C, integral membrane protein −1.2030 −0.3650 Cs Citrate synthase −1.1660 −0.3270 Ndufa1 NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 1 −1.1510 −0.2410 Ndufa8 NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 8 −1.1450 −0.1990 Cox6b1 Cytochrome c oxidase subunit Vib polypeptide 1 −1.1290 −0.3140 Mrps15 Mitochondrial ribosomal protein S15 −1.1150 −0.2280 Idh3 g Isocitrate dehydrogenase 3 (NAD+) gamma −1.0870 −0.1990 Acadvl Acyl-Coenzyme A dehydrogenase, very long chain −1.0820 −0.1850 Ndufa9 NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 9 −1.0620 −0.3370 Cox6a1 Cytochrome c oxidase subunit VIa polypeptide 1 −1.0570 −0.2870 Ndufc1 NADH dehydrogenase (ubiquinone) 1, subcomplex unknown, 1 −1.0470 −0.1760 Mfn2 Mitofusin 2 −1.0230 −0.5940 Pdhb Pyruvate dehydrogenase (lipoamide) beta −1.0140 −0.3930 Idh3b Isocitrate dehydrogenase 3 (NAD+) beta −0.9750 −0.5050 Atp5g3 ATP synthase, H + transporting, mitochondrial F0 complex, subunit C3 (subunit 9) −0.9440 −0.1990 Uqcrfs1 Ubiquinol-cytochrome c reductase, Rieske iron-sulfur polypeptide 1 −0.9060 −0.2240 Mrpl34 Mitochondrial ribosomal protein L34 −0.8870 −0.2150 Uqcrh Ubiquinol-cytochrome c reductase hinge protein −0.8840 −0.3650 Grpel1 GrpE−like 1, mitochondrial −0.8440 −0.2340 Uqcrc2 Ubiquinol-cytochrome c reductase core protein 2 −0.8300 −0.1600 Ndufb6 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 6 −0.8010 −0.5270 Ldhb Lactate dehydrogenase B −0.7460 −0.2100 Phb2 Prohibitin 2 −0.7320 −0.2170 Mrpl11 Mitochondrial ribosomal protein L11 −0.7270 −0.4530 Cox6c Cytochrome c oxidase subunit VIc −0.7160 −0.6190 Cox5b Cytochrome c oxidase subunit Vb −0.7090 −0.1590 Suclg1 Succinate-CoA ligase, GDP-forming, alpha subunit −0.7080 −0.2300 Cox7b Cytochrome c oxidase subunit VIIb −0.7030 −0.1930 Atp5j ATP synthase, H+ transporting, mitochondrial F0 complex, subunit F −0.6970 −0.3770 Cox5a Cytochrome c oxidase subunit Va −0.6920 −0.2190 Dlst Dihydrolipoamide S-succinyltransferase −0.6890 −0.2280 Atp5 h ATP synthase, H+ transporting, mitochondrial F0 complex, subunit D −0.6880 −0.3350 Mdh1 Malate dehydrogenase 1, NAD (soluble) −0.6860 −0.3760 Ndufb3 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 3 −0.6670 −0.2560 Mrpl35 Mitochondrial ribosomal protein L35 −0.6340 −0.2650 Prdx3 Peroxiredoxin 3 −0.6270 −0.2690 Ndufa5 NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 5 −0.6050 −0.2480 Ndufb5 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 5 −0.5830 −0.5650 Sdhd Succinate dehydrogenase complex, subunit D, integral membrane protein −0.5760 −0.1720 Ndufs8 NADH dehydrogenase (ubiquinone) Fe–S protein 8 Dysregulated genes are sorted by descending rank metric score in AT aObese gonadal white adipose tissue bAtherosclerotic aortae Discussion Type 2 diabetes and cardiovascular disease are common risks inherent with presence of metabolic syndrome and are triggered by insulin resistance and atherosclerosis, respectively. Several immunological and metabolic changes in serum levels from factors such as TNF-α [18], IL-6 [19], plasminogen activator inhibitor (PAI-1) [20], C-reactive protein (CRP) and fibrinogen [21], have been reported to be altered during insulin resistance and adipose tissue inflammation, both consequences of obesity, as well as atherosclerotic plaque formation during dyslipidemias [22, 23]. Even though diverse studies have been carried out to determine common variations at gene or protein levels between different metabolic tissues during obesity, the common pathways involved in the simultaneous development of adipose tissue inflammation and atherosclerosis remained elusive. Identification of common dysregulated pathways, however, could be a key in the development of both conditions and provide a chance for novel strategies to simultaneously prevent type 2 diabetes and cardiovascular disease in subjects at risk, i.e. those with metabolic syndrome [24, 25]. Pathway-based analysis is a powerful tool that allows to detect changes at a higher biological level than individual genes or molecules, complementing single gene level approaches with biological interactions, and thus contributing to a better understanding of the complexity of diseases [13]. Taking this together with the high reliability of the microarray technology nowadays [26–28], the pathways analyses based on microarray data can be considered as remarkably robust studies at the meta level. There are several publications showing single pathways to be involved in type 2 diabetes [29, 30] or atherosclerosis [31–33], such as focal adhesion [34], angiotensin II-NFκB [35], electron carrier activity, PPAR signaling and protein secretion [36]. However, in this study, we combined adipose tissue inflammation and atherosclerosis, in order to identify common dysregulated pathways in a systematical and unbiased manner, merging own and published data and applying different bioinformatic approaches. Initially, we took advantage of a recently established combined insulin resistance/atherosclerosis inbred mouse model and analyzed microarray raw data from AT and AO with the GSEA software. Using four different gene sets, Biocarta, KEGG, Reactome and Hallmark (Fig. 2) we obtained considerably overlaps of dysregulated pathways in each of the four analyses. We choose Hallmark results to conduct a deeper look. However, to consider data from Biocarta, KEGG and Reactome analyses remains an interesting aim for further studies (Fig. 3). Interestingly, the Hallmark analysis results included novel as well as already described common dysregulated pathways related with insulin resistance and atherosclerosis such as inflammatory response, IL6 JAK STAT3 signaling, TNFA signaling via NFKB, interferon gamma response, fatty acid metabolism, oxidative phosphorylation and adipogenesis (Table 1). This list of pathways could be of a great tool for researchers who are interested in the common mechanisms behind adipose tissue inflammation and atherosclerosis, since common genes enclosed in the pathways could be new targets to investigate mechanism as well as possible future treatments. After evaluation of the common pathways altered in both metabolic processes, we selected the inflammatory and oxidative phosphorylation pathways as representatives of up- and down-regulated pathways, respectively, due to the importance of the involved genes in the pathogenesis of type 2 diabetes and atherosclerosis [37–41]. We could show a clear expression pattern for all genes that contribute to the dysregulation of the pathway in each tissue (Figs. 4, 5). The upregulation of the inflammatory response pathway in inflamed adipose tissue as well as in atherosclerotic aorta in the animal model used in this study complements the recent findings showing after a long term high fat diet (16 weeks) that the homeostasis of the immune system is altered from a physiological immune response to a pathological state [42]. Moreover, the down-regulation of the oxidative phosphorylation pathway in AT supports the observations indicating decreased expression of the genes implicated in the mitochondria electron chain in visceral adipose tissue in type 2 diabetes [39]. However, our study shows the down-regulation of the oxidative phosphorylation pathway also in the atherosclerotic aorta. Additionally we also identified a considerable number of common genes between both analyzed tissues that significantly contribute to the dysregulation of the inflammatory response or the oxidative phosphorylation pathways and could be potentially involved in the pathogenesis of both type 2 diabetes and cardiovascular disease (Fig. 6; Tables 2, 3). A large amount of studies have been performed to identify potential treatment strategies for multifactorial diseases which are caused by a collective dysregulation of many genes. Recently, identification of dysregulated pathways was proposed as possible biomarkers for such disorders [43, 44]. This study conferred a strong evidence of similarities on the pathogenesis between insulin resistance and atherosclerosis, which could support diagnostic processes and drug design in a more effective manner facilitating a more personalized medicine in patients with metabolic syndrome. Conclusion In conclusion, we describe analogies at the pathway as well as individual gene level between obese adipose tissue and atherosclerotic aortae potentially to be considered as a basis to achieve novel therapeutic approaches for the simultaneous prevention of type 2 diabetes and atherosclerotic cardiovascular disease. Additional files 10.1186/s12933-016-0441-2 Additional tables. 10.1186/s12933-016-0441-2 Enrichment plots for inflammatory response pathway in AT and AO. The enrichment plots for inflammatory response pathway in: (a) obese white adipose tissue (AT) and (b) atherosclerotic aorta (AO) are shown, representing at the top the enrichment score and the leading edge subset, at the middle the genes that appear in the rank list and in the bottom the ranking metric that measures the correlation between the gene expression and the phenotype. 10.1186/s12933-016-0441-2 Enrichment plots for oxidative phosphorylation pathway in AT and AO. The enrichment plots for oxidative phosphorylation pathway in: (a) obese gonadal white adipose tissue (AT) and (b) atherosclerotic aorta (AO) are shown, representing at the top the enrichment score and the leading edge subset, at the middle the genes that appear in the rank list and in the bottom the ranking metric that measures the correlation between the gene expression and the phenotype. Abbreviations AOaortae ATgonadal white adipose tissue DDCdiabetogenic diet FDRfalse discovery rate GSEAgene set enrichment analysis IRPinflammation related pathways Ldlr−/−lDL-receptor knockout mice MRPmetabolism related pathways NCnormal chow diet NESnormalized enrichment score V. Moreno-Viedma and M. Amor contributed equally to this work Authors’ contributions VMV and MA conceived the study, researched, interpreted data and wrote the manuscript. AS researched and interpreted data. MB and GS contributed to intellectual content of the manuscript and reviewed and edited the manuscript. MZ contributed to data interpretation, made significant contributions to intellectual content of this manuscript, reviewed and edited the manuscript. TMS designed and supervised the study, critically reviewed and edited the manuscript, and rose funding. VMV and MA contributed equally to this work. All authors read and approved the final manuscript. Acknowledgements Not applicable for this section. Competing interests The authors declare that they have no competing interests. Availability of data and material The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus (TM. Stulnig et al., 2016) and are accessible through GEO Series accession number GSE76812 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE76812). Funding This work was supported by the Federal Ministry of Economy, Family and Youth and the National Foundation for Research, Technology and Development. (to T.M.S.). ==== Refs References 1. Balkau B Vernay M Mhamdi L Novak M Arondel D Vol S Tichet J Eschwege E The incidence and persistence of the NCEP (National Cholesterol Education Program) metabolic syndrome. The French D.E.S.I.R. study Diabetes Metab 2003 29 5 526 532 10.1016/S1262-3636(07)70067-8 14631330 2. Alberti KG Eckel RH Grundy SM Zimmet PZ Cleeman JI Donato KA Fruchart JC James WP Loria CM Smith SC Jr Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity Circulation 2009 120 16 1640 1645 10.1161/CIRCULATIONAHA.109.192644 19805654 3. Kaur J A comprehensive review on metabolic syndrome Cardiol Res Pract 2014 2014 943162 24711954 4. Pal S Ellis V The chronic effects of whey proteins on blood pressure, vascular function, and inflammatory markers in overweight individuals Obesity (Silver Spring) 2010 18 7 1354 1359 10.1038/oby.2009.397 19893505 5. Grundy SM Adipose tissue and metabolic syndrome: too much, too little or neither Eur J Clin Invest. 2015 45 1209 1217 10.1111/eci.12519 26291691 6. Lumeng CN Saltiel AR Inflammatory links between obesity and metabolic disease J Clin Invest 2011 121 6 2111 2117 10.1172/JCI57132 21633179 7. O’Rourke RW Inflammation in obesity-related disease Surgery 2009 145 3 255 259 10.1016/j.surg.2008.08.038 19231576 8. Curtis RK Oresic M Vidal-Puig A Pathways to the analysis of microarray data Trends Biotechnol 2005 23 8 429 435 10.1016/j.tibtech.2005.05.011 15950303 9. Draghici S Khatri P Tarca AL Amin K Done A Voichita C Georgescu C Romero R A systems biology approach for pathway level analysis Genome Res 2007 17 10 1537 1545 10.1101/gr.6202607 17785539 10. Khatri P Sirota M Butte AJ Ten years of pathway analysis: current approaches and outstanding challenges PLoS Comput Biol 2012 8 2 e1002375 10.1371/journal.pcbi.1002375 22383865 11. Sun S Ji Y Kersten S Qi L Mechanisms of inflammatory responses in obese adipose tissue Annu Rev Nutr 2012 32 261 286 10.1146/annurev-nutr-071811-150623 22404118 12. Subramanian A Tamayo P Mootha VK Mukherjee S Ebert BL Gillette MA Paulovich A Pomeroy SL Golub TR Lander ES Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles Proc Natl Acad Sci USA 2005 102 43 15545 15550 10.1073/pnas.0506580102 16199517 13. Mootha VK Lindgren CM Eriksson KF Subramanian A Sihag S Lehar J Puigserver P Carlsson E Ridderstrale M Laurila E PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes Nat Genet 2003 34 3 267 273 10.1038/ng1180 12808457 14. Neuhofer A Wernly B Leitner L Sarabi A Sommer NG Staffler G Zeyda M Stulnig TM An accelerated mouse model for atherosclerosis and adipose tissue inflammation Cardiovascu Diabetol 2014 13 23 10.1186/1475-2840-13-23 15. Messner B Zeller I Ploner C Frotschnig S Ringer T Steinacher-Nigisch A Ritsch A Laufer G Huck C Bernhard D Ursolic acid causes DNA-damage, p53-mediated, mitochondria- and caspase-dependent human endothelial cell apoptosis, and accelerates atherosclerotic plaque formation in vivo Atherosclerosis 2011 219 2 402 408 10.1016/j.atherosclerosis.2011.05.025 21703625 16. Huber W Carey VJ Gentleman R Anders S Carlson M Carvalho BS Bravo HC Davis S Gatto L Girke T Orchestrating high-throughput genomic analysis with bioconductor Nat Methods 2015 12 2 115 121 10.1038/nmeth.3252 25633503 17. Gentleman RC Carey VJ Bates DM Bolstad B Dettling M Dudoit S Ellis B Gautier L Ge Y Gentry J Bioconductor: open software development for computational biology and bioinformatics Genome Biol 2004 5 10 R80 10.1186/gb-2004-5-10-r80 15461798 18. Zahorska-Merkiewicz B Janowska J Olszanecka-Gliniarowicz M Majewski T Serum concentration of turnout necrosis factor in obese women J Endocrinol Invest 1999 22 66 66 10.1007/BF03345481 19. Park HS Park JY Yu R Relationship of obesity and visceral adiposity with serum concentrations of CRP, TNF-alpha and IL-6 Diabetes Res Clin Pract 2005 69 1 29 35 10.1016/j.diabres.2004.11.007 15955385 20. Festa A D’Agostino R Jr Tracy RP Haffner SM Elevated levels of acute-phase proteins and plasminogen activator inhibitor-1 predict the development of type 2 diabetes: the insulin resistance atherosclerosis study Diabetes 2002 51 4 1131 1137 10.2337/diabetes.51.4.1131 11916936 21. Kaptoge S Di Angelantonio E Pennells L Wood AM White IR Gao P Walker M Thompson A Sarwar N Caslake M C-reactive protein, fibrinogen, and cardiovascular disease prediction N Engl J Med 2012 367 14 1310 1320 10.1056/NEJMoa1107477 23034020 22. Pradhan AD Ridker PM Do atherosclerosis and type 2 diabetes share a common inflammatory basis? Eur Heart J 2002 23 11 831 834 10.1053/euhj.2001.3052 12042000 23. Donath MY Shoelson SE Type 2 diabetes as an inflammatory disease Nat Rev Immunol 2011 11 2 98 107 10.1038/nri2925 21233852 24. Ridker PM Luscher TF Anti-inflammatory therapies for cardiovascular disease Eur Heart J 2014 35 27 1782 1791 10.1093/eurheartj/ehu203 24864079 25. Shoelson SE Lee J Goldfine AB Inflammation and insulin resistance J Clin Invest 2006 116 7 1793 1801 10.1172/JCI29069 16823477 26. Allanach K Mengel M Einecke G Sis B Hidalgo LG Mueller T Halloran PF Comparing microarray versus RT-PCR assessment of renal allograft biopsies: similar performance despite different dynamic ranges Am J Transplant 2008 8 5 1006 1015 10.1111/j.1600-6143.2008.02199.x 18416738 27. Wurmbach E Yuen T Ebersole BJ Sealfon SC Gonadotropin-releasing hormone receptor-coupled gene network organization J Biol Chem 2001 276 50 47195 47201 10.1074/jbc.M108716200 11581274 28. Etienne W Meyer MH Peppers J Meyer RA Comparison of mRNA gene expression by RT-PCR and DNA microarray Biotechniques. 2004 36 4 618 620 15088380 29. Sato M Dehvari N Oberg AI Dallner OS Sandstrom AL Olsen JM Csikasz RI Summers RJ Hutchinson DS Bengtsson T Improving type 2 diabetes through a distinct adrenergic signaling pathway involving mTORC2 that mediates glucose uptake in skeletal muscle Diabetes 2014 63 12 4115 4129 10.2337/db13-1860 25008179 30. Munasinghe PE Riu F Dixit P Edamatsu M Saxena P Hamer NS Galvin IF Bunton RW Lequeux S Jones G Type-2 diabetes increases autophagy in the human heart through promotion of Beclin-1 mediated pathway Int J Cardiol 2015 202 13 20 10.1016/j.ijcard.2015.08.111 26386349 31. King JY Ferrara R Tabibiazar R Spin JM Chen MM Kuchinsky A Vailaya A Kincaid R Tsalenko A Deng DX Pathway analysis of coronary atherosclerosis Physiol Genomics 2005 23 1 103 118 10.1152/physiolgenomics.00101.2005 15942018 32. Fu Y, Ma D, Liu Y, Li H, Chi J, Liu W, Lin F, Hu J, Zhang X, Zhu M et al. Tissue factor pathway inhibitor gene transfer prevents vascular smooth muscle cell proliferation by interfering with the MCP-3/CCR2 pathway. Lab Invest. 2015. 33. Xing SS Yang XY Zheng T Li WJ Wu D Chi JY Bian F Bai XL Wu GJ Zhang YZ Salidroside improves endothelial function and alleviates atherosclerosis by activating a mitochondria-related AMPK/PI3 K/Akt/eNOS pathway Vascul Pharmacol 2015 72 141 152 10.1016/j.vph.2015.07.004 26187353 34. Zhang L Cui Y Fu F Li Z Pan X Li H Li L An insight into the key genes and biological functions associated with insulin resistance in adipose tissue with microarray technology Mol Med Rep 2015 11 3 1963 1967 25385661 35. Zhou MS Liu C Tian R Nishiyama A Raij L Skeletal muscle insulin resistance in salt-sensitive hypertension: role of angiotensin II activation of NFkappaB Cardiovasc Diabetol 2015 14 45 10.1186/s12933-015-0211-6 25928697 36. Shim U Kim HN Sung YA Kim HL Pathway analysis of metabolic syndrome using a genome-wide association study of Korea Associated Resource (KARE) cohorts Genom Inform 2014 12 4 195 202 10.5808/GI.2014.12.4.195 37. O’Rourke RW Inflammation, obesity, and the promise of immunotherapy for metabolic disease Surg Obes Relat Dis 2013 9 5 609 616 10.1016/j.soard.2012.07.010 22959472 38. Rocha VZ Libby P Obesity, inflammation, and atherosclerosis Nat Rev Cardiol 2009 6 6 399 409 10.1038/nrcardio.2009.55 19399028 39. Dahlman I Forsgren M Sjogren A Nordstrom EA Kaaman M Naslund E Attersand A Arner P Downregulation of electron transport chain genes in visceral adipose tissue in type 2 diabetes independent of obesity and possibly involving tumor necrosis factor-alpha Diabetes 2006 55 6 1792 1799 10.2337/db05-1421 16731844 40. Rochette L Zeller M Cottin Y Vergely C Diabetes, oxidative stress and therapeutic strategies Biochim Biophys Acta 2014 1840 9 2709 2729 10.1016/j.bbagen.2014.05.017 24905298 41. Kondo T Hirose M Kageyama K Roles of oxidative stress and redox regulation in atherosclerosis J Atheroscler Thromb 2009 16 5 532 538 10.5551/jat.1255 19749495 42. Kim J, Kwon EY, Park S, Kim JR, Choi SW, Choi MS, Kim SJ. Integrative systems analysis of diet-induced obesity identified a critical transition in the transcriptomes of the murine liver and epididymal white adipose tissue. Int J Obes (Lond). 2015. 43. Liu KQ Liu ZP Hao JK Chen L Zhao XM Identifying dysregulated pathways in cancers from pathway interaction networks BMC Bioinformatics 2012 13 126 10.1186/1471-2105-13-126 22676414 44. Liu Y Li Z Zhang M Deng Y Yi Z Shi T Exploring the pathogenetic association between schizophrenia and type 2 diabetes mellitus diseases based on pathway analysis BMC Med Genom 2013 6 Suppl 1 S17 10.1186/1471-2350-14-17

PMC005xxxxxx/PMC5000405.txt

==== Front BMC Womens HealthBMC Womens HealthBMC Women's Health1472-6874BioMed Central London 33810.1186/s12905-016-0338-yResearch ArticleSocio-economic and demographic factors influencing nutritional status among early childbearing young mothers in Bangladesh Islam Ashraful drashraf@um.edu.my 1Islam Nurul mnurul58@yahoo.com 2Bharati Premananda pbharati@gmail.com 3Aik Saw sawaik@hotmail.com 4Hossain Golam +8801914254013hossain95@yahoo.com 21 Research Management Centre, Faculty of Medicine, University of Malaya, Kuala Lumpur, 50603 Malaysia 2 Department of Statistics, University of Rajshahi, Rajshahi, 6205 Bangladesh 3 Biological Anthropology Unit, Indian Statistical Institute, 203 SH1 State highways 1, Kolkata, 700108 India 4 Department of Orthopaedic Surgery, Faculty of Medicine, University of Malaya, Kuala Lumpur, 50603 Malaysia 26 8 2016 26 8 2016 2016 16 1 5824 11 2015 19 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Early childbearing influences women’s health. This study aims to examine the effects of socio-demographic factors on nutritional status of early childbearing mothers in Bangladesh based on Body Mass Index (BMI) as the indicator. Methods Data was extracted from Bangladesh Demographic and Health Survey (BDHS)-2011. The survey was performed on 17,842 married women aged 15–49. We focused on early childbearing mothers (age ≤ 24, and who had delivered their first child ≤ 20). Mothers who were underweight (BMI ≤ 18.5 kg/m2) would be further classified into various grades of chronic energy deficiency (CED): mild (17.0 ≤ BMI < 18.5 kg/m2), moderate (16.0 ≤ BMI <17.0 kg/m2), and severe (BMI < 16.0 kg/m2). Multiple logistic regression model was used to examine the effect of socio-demographic factors on nutritional status. Results Mean age of the mothers was 20.49 ± 2.37 years (ranged 15–24 years). The prevalence of underweight among early childbearing mothers was 32.1 % (urban 25 % and rural 35.1 %). Most of the underweight mothers had mild (62.2 %) CED, while the remaining had either moderate (25.9 %) or severe (11.9 %) CED. Multiple logistic regression analysis demonstrated that young mothers from rural areas, poor families, and those who were illiterate or with low level of education, working, and married to unemployed husband were at higher risk for being underweight. Young mothers who had non-caesarean delivered, delivered at home, or married at early age and had more than two children were also at higher risk for being underweight. Conclusions The prevalence of underweight among early childbearing mothers in Bangladesh is very high (32.1 %), associated with the still common practice of teenage marriage. Education level, wealth index, occupation, place of residence, age at first marriage and parity were important predictors for their nutritional status. The government and non-government organizations should take initiatives to reduce the prevalence of underweight mothers in Bangladesh. Keywords Early ChildbearingYoung MotherUnderweightBDHS-2011Logistic Regressionissue-copyright-statement© The Author(s) 2016 ==== Body Background Women who become pregnant before age 20 are considered as early childbearing mothers. General wellbeing and nutritional requirements of these young mothers have recently received more attention especially in developing and under-developed countries [1]. It has been shown that early childbearing mothers were at higher risk of prenatal morbidities such as gestational diabetes, gestational hypertension and preterm labour compared to the general population [2]. The risk of pregnancy related mortality for mothers aged 15 to 19 was twice as high compared to those aged 20 and older [3]. The mortality risk would be 5–7 times higher for mothers who became pregnant before age 15 [4]. Some anthropometric and socio-economic factors had been associated with adverse health consequences among adolescent mothers [5]. The pelvic bone of young mothers may not have fully developed to accommodate the passage of the babies, increasing the risk of obstructed labour [6]. A recent longitudinal study on African-American community in Chicago reported that adolescent mothers were more likely to be unemployed, live in poverty and dependent on social welfare [7]. Body mass index (BMI) is calculated from a person’s weight and height. It is an important indicator of the nutritional status for a population. BMI value of less than or equal to 18.5 kg/m2 is considered as underweight, and this is a common finding among people suffering from chronic energy deficiency. Underweight women were associated with higher risk of adverse health outcomes like hip fractures [8, 9]. Pregnant mothers who were underweight have higher risk of perinatal mortality, and delivering low birth weight babies [10]. The mean age of first marriage for Bangladeshi women was 15.69 ± 2.97 years [11]. Based on a report by Kamal in 2012, the practice of teenage marriage and early childbearing were still common in Bangladesh despite substantial improvements in various Human Development indicators (HDI) [12]. In the urban areas of Bangladesh, researchers have reported that underweight was very common among ever-married non-pregnant women [13]. Prevalence of chronic energy deficiency (CED, with BMI <18.5 kg/m2) was very high among women from poor families in both rural and urban areas of this country (38.8 % rural, 29.7 % urban poor) [14]. More recently, Hossain et al. studied the association between BMI and socio-demographic factors among ever-married non-pregnant Bangladeshi women (aged 15–49 years) based on birth year cohorts from 1957 to 1992. They reported an increasing trend of BMI during the first sixteen years from 1957 to 1972, but a decreasing trend thereafter of Bangladeshi ever married women aged 15–49 years [15]. Information on the importance of adequate nutritional for childbearing mothers would not only help the community to understand their needs, but also indirectly help to promote the well being on their children who would eventually grow up and contribute as leaders or workforce for the nation. This is especially true in Asian communities because of the traditional role of mothers in the family and the community. Studies have already shown that nutritional status of married and unmarried women was influenced by various demographic and socio-economic factors such as age, education level, wealth index, age at first marriage, number of ever-born children, residence, religion, occupation, place of delivery and method of delivery [13, 15–17]. Two of these studies were focused on married Bangladeshi women in their reproductive age [13, 15], and two others were conducted on Indian women [16, 17]. In Bangladesh, special attention should be given to early childbearing mothers because this subgroup of the population was associated with high risk of adverse health outcome compared to mothers of older age [1]. Moreover, practice of teenage marriage remained common in Bangladesh. There has not been any published study on the general health for this group of population in Bangladesh. We therefore decided to use the nutritional status as proxy for general health, and study risk factors related to the general health status of early childbearing mothers in this country. Methods Information for this cross sectional study was extracted from a sample of 17,842 married Bangladeshi women (age 15–49 years) from the Bangladesh Demographic and Health Survey (2011- BDHS). From this initial data set, we identified married Bangladeshi women age 24 and below, and selected those who had been pregnant and delivered before age 20. We considered age 24 as the cut off point for inclusion because the practice of early childbearing might still influence the health status of these young women. We excluded women who were pregnant at the time of survey, and those who delivered their first baby after age 20. We also excluded those with incomplete information that was required for our study. Following the selection process, we obtained data set of 2,808 women. We subsequently checked the available data for outliers using an informal technique [18] (checking for abnormal BMI values based on a scatter diagram), and removed some subjects from analysis because their data may affect the interpretation of results [19, 20]. We eventually came up with a list of 2,743 young women who had delivered at least one child before age 20. BDHS-2011 was conducted by way of two-stage stratified cluster sampling. In the first stage, the researchers selected 600 enumeration areas (EAs) with probability proportional to the EA size. Of these clusters, 207 EAs were from urban areas and 393 from rural areas. In the second stage, a systematic sampling method was used to select an average of 30 households for each EA. This would provide statistically reliable estimates of key demographic and health variables for the whole country, including all the seven divisions, covering both the urban and rural areas. The BDHS is a part of the worldwide Demographic and Health Surveys program. The BDHS-2011 was conducted under the authority of the National Institute of Population Research and Training (NIPORT) of the Ministry of Health and Family Welfare. The project was conducted by Mitra and Associates, a Bangladeshi research firm based in Dhaka. ICF International from Maryland, United States (US), provided technical support as part of its international Demographic and Health Surveys program (MEASURE DHS). United States Agency for International Development (USAID) provided the financial support for the project. The sampling technique, survey design, survey instruments, measuring system, quality control, ethical approval and subject consent for the 2011 BDHS have been described elsewhere [21]. BDHS-2011 collected socio-demographic, health and lifestyle information from each of their selected subject from July 8, 2011 to December 27, 2011. In addition, body height and weight were measured for all subjects. Outcome variable The outcome variable of this study was nutritional status, and it was measured by BMI. BMI was defined and calculated as the ratio of weight in kilograms to height in meters squared. The BMI was classified according to most widely used categories for the adults: (i) underweight (under-nutrition) (BMI ≤ 18.5 kg/m2), (ii) normal weight (18.5 < BMI <25 kg/m2), (iii) overweight (25 ≤ BMI < 30 kg/m2) and (iv) obese (BMI ≥ 30 kg/m2) [16, 22, 23]. The subjects were also classified on the basis of chronic energy deficiency (CED) grades as follows: (i) grade III (severe thinness) (BMI < 16.0 kg/m2), (ii) grade II (moderate thinness) (16.0 ≤ BMI <17.0 kg/m2), (iii) grade I (mild thinness) (17.0 ≤ BMI < 18.5 kg/m2) [17, 24]. Independent variables Various socio-economic and demographic factors were used in this study as independent variables, and they included : type of place of residence, respondent’s (woman) educational level, partner’s (husband) educational level, respondent’s occupation, partner’s occupation, wealth index, type of delivery, place of delivery, total number of children ever born, age at first marriage, and respondent age at first birth. More detail on the definition of these variables is available in the BDHS-2011 survey report [21]. Statistical analysis Descriptive statistics was done for calculating prevalence of underweight, normal weight, overweight and obese among early childbearing mothers. Chi-square test was utilized in this study for selecting significant independent factors for logistics regression models. Finally, binary multiple logistic regression was used to examine the relative importance of socio-demographic factors on early childbearing mothers’ health. In this model, category of body size (BMI) was considered as a dependent variable coded as 0 = normal weight and 1 = underweight. The underlying multiple logistic regression models corresponding to each variable is: logp1−p=β0+β1X1+β2X2+β3X3+β4X4+β5X5+β6X6+β7X7+β8X8+β9X9+β10X10+β11X11 where, p = the probability of underweight (coded 1) 1-p = the probability of normal (coded 0) X1 = place of residence (coded; urban = 0, rural = 1) X2 = respondent’s (woman) educational level (coded; no education = 0, school education = 1, higher education = 2) X3 = partner’s (husband) educational level (coded; no education = 0, school education = 1, higher education = 2) X4 = respondent’s occupation (coded; housewife = 0, hard labor = 1) X5 = partner’s occupation (coded; employed = 0, farmer/worker = 1) X6 = wealth index (coded; poorest = 1, poorer = 2, middle = 3, richer = 4, richest = 5) X7 = type of delivery (coded; non-caesarian = 0, caesarian = 1) X8 = place of delivery (coded; hospital/clinic = 0, home = 1) X9 = total number of children ever born X10 = age at first marriage X11 = respondent age at first birth, and β0 = intercept term, and β1, β2,…,β11 are unknown coefficients. Multicollinearity problem among the predictor variables were checked by standard error (SE) [25]. If the magnitude of the SE is approximately 0.001–5.0 [25], suggested that there is no evidence of Multicollinearity problem. All the statistical analyses were carried out using Statistical Package for Social Scientists (IBM SPSS version 22.0) software. Results After we analysed the data set of 2,743 young non-pregnant young Bangladeshi mothers (age 24 or younger) who delivered their first child before age 20, we noted that the mean age of the mothers was 20.49 ± 2.37 years (95 % CI: 20.40–20.58). The mean number of children per mother was 1.60 ± 0.76. More than half (53.8 %) of them had one child, about one third (34.6 %) had two children, and the rest (11.6 %) had two and more children. Mean weight of all mothers was 45.97 ± 8.01 kg (95 % CI: 45.67–46.27), ranging from 28.00 to 104.20 kg. Mean height of the mothers was 150.87 ± 5.48 cm (95 % CI: 150.66–151.07), ranging from 112.10 to 197.20 cm. BMI varied from 13.09 kg/m2 to 38.13 kg/m2, with a mean of 20.16 ± 3.07 kg/m2 (95 % CI: 20.04–20.27) (Table 1). The age at first childbirth in these women varies from 12 to 19 years, with the mean being 16.54 ± 1.67 years (95 % CI: 16.48–16.60).Table 1 Descriptive statistics for age, age of first birth, weight, height and BMI of early childbearing young mothers in Bangladesh (n = 2743) Variable Mean SD SE 95 % CI for mean Mini-mum Maxi-mum Lower Upper Age (year) 20.49 2.377 0.045 20.40 20.58 13 24 Age of first birth (year) 16.54 1.679 0.032 16.48 16.60 12 19 Weight(kg) 45.97 8.014 0.153 45.67 46.27 28.00 104.20 Height (cm) 150.87 5.482 0.104 150.66 151.07 112.10 197.20 BMI (kg/m2) 20.16 3.075 0.058 20.04 20.27 13.09 38.13 Based on BMI categories, more than half (60.5 %) of the mothers had normal weight, and about one third (32.1 %) of them were underweight. There were relatively few mothers who were overweight (6.2 %) or obese (1.2 %) (Table 2), and they were mostly from the urban areas (Fig. 1). When we considered mothers who had CED (same criteria as underweight based on BMI categories: BMI ≤ 18.5 kg/m2), 11.9 % were grade III (severe), 25.9 % were grade II (moderate), and 62.2 % were mild (grade I) (Table 2).Table 2 Frequency distribution of BMI category and chronic energy deficiency of early childbearing mothers (n = 2743) Variables n (%) BMI category  Underweight (BMI ≤ 18.5 kg/m2) 879 (32.1 %)  Normal weight (18.5 < BMI < 25 kg/m2) 1660 (60.5 %)  Overweight (25 ≤ BMI < 30 kg/m2) 171 (6.2 %)  Obese (BMI ≥30 kg/m2) 33 (1.2 %) CED category  CED grade III (severe thinness) (BMI < 16.0 kg/m2) 104 (11.9 %)  CED grade II (moderate thinness) (16.0 kg/m2 ≤ BMI < 17.0 kg/m2) 228 (25.9 %)  CED grade I (mild thinness) (17.0 kg/m2 ≤ BMI ≤ 18.5 kg/m2) 547 (62.2 %) Fig. 1 Difference between urban and rural in the percentage of underweight and obese mothers in Bangladesh Women from the rural areas were more likely to be underweight (38.8 %) compared to those from urban areas (25.5 %). Those with lower level of education (39.9 %) were more likely to be underweight compared to those who had higher level of education (30.3 %). Level of education for the husband also showed similar pattern of influence on the nutritional status of these women. Family income was another factor that influenced the BMI. Our study showed that women from poor households were more likely (39 %) to be underweight compared with those not from poor households (26.1 %). We also noted that women who were not working (housewives) were less likely to be underweight compared to those who had to work (32.4 % vs 49.0 %). On the contrary, women with husbands who were unemployed were more likely to be underweight compared to those with working husbands (37.7 vs 30.8 %). Chi-square test showed that the association of all these factors with low BMI (being underweight) were statistically significant (p < 0.001). When we look at the method and place of delivery their child, we noted that mothers who delivered at home had higher risk of being underweight compared to those who delivered in hospitals or clinics (36.4 % vs 28.7 %). The association of place of delivery and underweight was significant (p < 0.001). Those who delivered naturally (non-Caesarean section) were more likely to be underweight (35.3 % vs 27.7 %) since most of them were home deliveries. The association of type of delivery and underweight was statistical significant (p < 0.05). Women who were married before age 17 were more likely to be underweight compared to those who were married later (35.2 % vs 31.1 %). In addition, more than one third of women who delivered their first child before age 17 were underweight, and this was much higher than those who delivered their first child at an older age (36.8 % vs 32.1 %). We also noted that women who had more than two children were more likely to be underweight than those who had only one or two children (41.4 % vs 33.5 %). The association between all these factors and underweight was statistically significant (p < 0.05) (Table 3).Table 3 Association between underweight and socio-economic, and demographic factors of early childbearing Bangladeshi mother (BDHS-2011) (n = 2539) Variable Group Underweight n (%) χ 2-value p-value Place of residence Urban 215 (25.5) 44.09 <0.001 Rural 659 (38.8) Mother education level Up to primary 431 (39.9) 25.04 <0.001 Secondary and higher 443 (30.4) Partner education level Up to primary 560 (38.5) 25.52 <0.001 Secondary and higher 314 (28.9) Respondent’s occupation Housewives 722 (32.4) 33.39 <0.001 Involved with hard labor 152 (49.0) Partner’s occupation Employee 373 (30.8) 13.25 <0.001 Non-employee 501 (37.7) Wealth index Poor 638 (39.0) 43.39 <0.001 Rich 236 (26.1) Place of delivery Respondent home 688 (36.4) 12.38 <0.001 Hospital/Clinic 186 (28.7) Type of delivery Non-caesarean 793 (35.3) 6.33 0.011 Caesarean 81 (27.7) Mother age at first marriage Lowest through 16 years 730 (35.2) 4.27 0.039 17 years and above 144 (31.1) Mother age at first birth Lowest through 16 years 460 (36.8) 6.31 0.012 17 years and above 414 (32.1) Total children ever born 1–2 751 (33.5) 7.28 0.007 3 and above 123 (41.4) Since there were very few overweight and obese women in our sample population (6.2 % and 1.2 % respectively), we decided to exclude them from χ2-test and logistic regression model. We considered underweight as 1 (reference case) and normal as 0 (non-reference case), and used them as dependent variable in this binary model. Only variables that demonstrated significant association were considered as independent variables in this model. The logistic regression coefficient and odds ratio showed that women who came from rural areas had a 1.478 times [95 % CI: 1.23–1.78; p < 0.01] higher chance to be underweight, as compared to those who came from urban areas. Uneducated mothers were 4.169 [95 % CI: 2.10–8.26; p < 0.01] times more likely to be underweight compared to educated mothers, and those with only secondary education were 2.997 times [95 % CI: 1.57–5.72; p < 0.01] more likely to be underweight compared to mothers with higher levels of education. Similar observation was noted when we analysed the risk of these women being underweight with education levels of their partner or husband (Table 4). Working mothers would have 2.085 times [95 % CI: 1.58–2.55; p < 0.01] higher risk of being underweight compared to mothers who were housewives. On the other hand, women whose partner was unemployed would have 1.204 times [95 % CI: 1.02–1.47; p < 0.05] higher risk of being underweight compared to women whose partner were working as farmer or worker. When we look at the economic background, risk of being underweight for women from poorest, poor, middle and rich families would be 3.303 times [95 % CI: 2.44–4.47; p < 0.01], 2.104 times [95 % CI: 1.55–2.86; p < 0.01], 1.944 times [95 % CI: 1.43–2.65; p < 0.01], and 1.870 times [95 % CI: 1.36–2.56; p < 0.01] higher than those from richest family respectively. The risk of being underweight for mothers who delivered naturally was 1.258 times [95 % CI: 1.03–1.85; p < 0.05] higher than those who delivered by Caesarian sections. Mothers who delivered at home had 1.290 times [95 % CI: 1.06–1.57; p < 0.05] higher risk of being underweight compared to mothers who delivered at hospitals or clinics. Young mothers with more than 2 children would have 1.41 times [95 % CI: 1.11–1.79; p < 0.01] higher risk to be underweight. On the other hand, age at first marriage [OR = 0.953, 95 % CI: 0.91–0.96; p < 0.05] and age at first childbirth [OR = 0.960, 95 % CI: 0.92–0.98; p < 0.05] were negatively related to being underweight, and the risk would reduce with the increasing age (Table 4).Table 4 Effect of selected socio-economic and demographic characteristics on young mothers’ nutritional status (n = 2539) Variable * p-value OR a (95 % CI for odds ratio) Place of Residence  Rural (Ref. Urban) p < 0.01 1.478 (1.23, 1.78) Respondent education level  No educated (Ref. Higher) p < 0.01 4.169 (2.10, 8.26)  Secondary (Ref. Higher) p < 0.01 2.997 (1.57, 5.72) Partner education level  No educated (Ref. Higher) p < 0.01 2.417 (1.68, 3.48)  Secondary (Ref. Higher) p < 0.01 1.851 (1.32, 2.61) Respondent’s Occupation  Hard labor mother (Ref. Housewife) p < 0.001 2.085 (1.58, 2.55) Partner’s Occupation  Farmer/Worker (Ref. Employed) p < 0.05 1.204 (1.02, 1.47)  Businessman (Ref. Employed) 0.302 0.878 (0.86, 1.13) Wealth Index  Poorest (Ref. Richest) p < 0.01 3.303 (2.44, 4.47)  Poor (Ref. Richest) p < 0.01 2.104 (1.55, 2.86)  Middle (Ref. Richest) p < 0.01 1.944 (1.43, 2.65)  Rich (Ref. Richest) p < 0.01 1.870 (1.36, 2.56) Type of delivery  Non-caesareans (Ref. Cesarean) p < 0.05 1.258 (1.03, 1.85) Place of Delivery  Home (Ref. Hospital/Clinic) p < 0.05 1.290 (1.06, 1.57) Total Children Ever Born  3 & more children (Ref. 1–2 children) p < 0.01 1.141 (1.11, 1.79) Age at First Marriage p < 0.05 0.953 (0.91, 0.96) Age at First Birth p < 0.023 0.960 (0.92, 0.98) Underweight was considered as reference category for dependent variable (underweight =1, normal weight =0) Hosmer-Lemeshow test, (p = 0.96), Pearson Chi-square & Sig., (p < 0.001) and classification table (overall correctly classified percentage = 87) were applied to check the model fitness aAdjusted odds ratio Note: *p-value (p < 0.01, 1 % and p < 0.05, 5 % level of significance) Discussion Our study showed that the mean BMI of young childbearing women in Bangladesh was 20.16 kg/m2, and 32.1 % of these women were underweight. Very few young women in our study were noted to be overweight (6.2 %) or obese (1.2 %). Similar outcome pattern was observed in few other studies based on all Bangladeshi women [13, 15]. A large population study conducted in India reported a relatively similar finding where 31.2 % of their women were underweight, with only 9.4 % in overweight and 2.6 % in obese categories [17]. One local study that focused on women living in the slum areas in Bangladesh capital (Dhaka) reported that more than half (54 %) of these women were underweight, and concluded that the most likely contributing factor would be extreme poverty [26]. Among the underweight mothers (BMI ≤ 18.5), most of them had mild CED (grade I), more than one-fourth had moderate CED (grade II), and a relatively high percentage (11.9 %) suffered severe CED (grade III). Chronic malnutrition can retard the physiological developments of an individual, and this negative effect may be more prominent in young women in their reproductive age. It has been reported that although the difference in mortality rates between normal adult women and those with grade I CED was only about 1 % per year, but the percentage increased significantly in those with grade II and grade III CED (BMI <17.0 kg/m2) [27]. Our study showed that the risk of CED among rural women was higher than urban women, and this would most likely due to poverty and lack of resources on maternal health or general health education [21]. We noted that level of education was associated with nutritional status in young women of Bangladesh. Multivariate analyses indicated that the risk of being underweight for women who were either illiterate or with only primary education was higher than those with secondary or higher education. On the other hand, early childbearing was also linked with lower level of education, being undernourished and severe CED. Educated women were generally more conscious about the timing of childbearing, and many of them were well informed on adverse outcomes associated with early motherhood. Our study demonstrated that level of education of both these women and their husbands were important factors for nutritional status of our study subjects. From the logistic regression analysis, we noted that the risk of being underweight of young women from rural area was higher than those from urban area. In Indian subcontinent, teenage marriage and early pregnancy are more common among rural compared to urban communities [28]. In addition to culture and religious factors, poverty has also been associated with teenage pregnancy [29]. Our analysis showed that wealth index influenced the BMI of Bangladeshi women, where poorer mothers had the higher chance to be undernourished. Poor countries such as Niger and Bangladesh had higher rates of teenage mothers compared with richer countries such as Switzerland and Japan [30]. Since more than 35 % of Bangladeshi women were poor, and 32 % were illiterate [21], we would expect these factors to contribute towards higher rates of early childbearing mothers in this country. The present study demonstrated that age at first marriage, age at first birth, place of childbirth, and total number of children were important factors associated to nutritional status of young Bangladeshi women (Table 4). A local study by Hossain et al. also listed poor economic conditions, illiteracy, large family size, early age at marriage, early age at first delivery and lack of medical facilities as main factors contributing to the low BMI for married non-pregnant women staying in the rural areas [15]. By using the BDHS-2004 database, Khan and Kraemer [13] reported that age, education level, type of occupation, place of stay and marital status influenced the BMI of married non-pregnant women from the urban areas of Bangladesh. Based on the Nutritional Surveillance Project (NSP) conducted between 2000 and 2004, Shafique et al. reported that age, level of education, wealth index, place of residence were important factors that influence the BMI of women in this country [14]. Data used in this study was gathered by the BDHS-2011 that covered both the rural and urban populations, and our study indicated that early childbearing is still common and the practice is deeply entrenched in Bangladeshi culture. There were two other studies on the association of BMI with various socio- demographic factors in Bangladeshi women aged between 15 and 49 [13, 14]. In the current study, we focused on the subgroup aged 24 and below because they would represent the younger generation who will most likely respond to corrective measures by the authorities and public health organizations. According to the 65th World Health Assembly report [31], there were about 2 million girls under the age of 15 and 16 million girls aged between 15 to 19 years giving birth every year worldwide. About 95 % of all adolescent births occurred in low and middle income countries, and they were more likely staying in the poor rural areas and had little education. In developing countries, about 90 % of births in adolescents occur within marriage. The percentage was about 70–80 % in South America and in sub-Saharan Africa, but close to 100 % in Western Asia/Northern Africa, Central Asia, and South-Central and South-Eastern Asia [32]. The National Campaign to Prevent Teen Pregnancy of World Health Organization reported that career opportunity for many young women was reduced by earlier childbearing and failure to complete high school education [30]. Despite these obvious problems and undesirable consequences, most of the adolescent pregnancies in Bangladesh were pre-planned and highly valued [33]. With poverty and lack of education, the young families would not know the medical and health problems associated with their behaviour. Limitations of study This study used secondary data derived from a national level cross-sectional survey where some of pertinent health indicator variables (like blood pressure, mortality as well as morbidity pattern) were not available. Despite the careful design and stratification of sample population, selection bias and reporting error might still be possible, and it was not known what women’s height and weight were taken before the pregnancy or immediate after the delivery. We only investigate the association between selected socio-economic and demographic factors with nutrition status that was represented by BMI. We were not able to cover physiological factors that may also be influenced by nutritional status of the women, like age at menarche [34], level of physical activities, level of energy intake [35], and behaviour patterns like dietary habits, smoking habits, weight goals, methods of weight-loss and body-shape perceptions [36]. Subsequent researchers may consider including these variables. Conclusion About one third (32.1 %) of early childbearing mothers in Bangladesh were underweight, and 37.4 % of these underweight mothers were considered to have moderate (grade II) and severe (grade III) CED. Low level of education of both the women and their spouses, and poverty were factors associated with poor nutritional status, especially for those from the rural areas. Women who delivered naturally and those who delivered at home were more likely to ne underweight, and this was also applicable for those with more than two children at the time of study. Our study also noted that early marriage and early childbearing were still commonly practiced in Bangladesh. Our study indicated that poor nutrition remained a serious problem among young childbearing mothers in Bangladesh, and we hope that information derived from this study would be able to help relevant authorities to plan remedial actions. Abbreviations BDHSBangladesh Demographic and Health Survey BMIBody mass index CEDChronic Energy Deficiency EAsEnumeration areas HDIHuman Development Indicators NIPORTNational Institute of Population Research and Training NSPNutritional Surveillance Project USAIDUnited States Agency for International Development Acknowledgements We would like to thank the Bangladesh Demographic and Health Survey (BDHS) for providing nationally representative based data collected 2011. We express our special thanks to Professor Wah Yun Low, Research Management Centre, Faculty of Medicine, University of Malaya for her critical review and comments on this paper. Funding The authors have no support or funding to report. Availability of data and material 2011-BDHS datasets are freely available at [http://dhsprogram.com/data/dataset/ Bangladesh_Standard-DHS_2011.cfm?flag = 0]. Authors’ contributions AI and GH created concept, participated in the design of the study, performed the statistical analysis and drafted the manuscript. NI, PB and SA made revision of the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable for this study. Ethics approval and consent to participate The 2011 BDHS received ethics approval from the Ministry of Health and Family Welfare, Bangladesh. The 2011 BDHS received written consent from each individual. ==== Refs References 1. Goli S Rammohan A Singh D The Effect of Early Marriages and Early Childbearing on Women’s Nutritional Status in India Matern Child Health J 2015 25656721 2. Palacios J Kennedy HP Reflections of Native American teen mothers J Obstet Gynecol Neonatal Nurs 2011 39 4 425 434 10.1111/j.1552-6909.2010.01149.x 3. United Nations We the children: end-decade review of the follow-up to the World Summit for Children. Report of the Secretary-General (A/S-27/3) 2001 New York United Nations 4. UNESCO Case Studies. Malaysia Demographic Charact eristic of Adolescents. Adolescent Health Reprodu ctive and Sexual Health 2004 Bangkok UNESCO 5. Nonnemaker JM Morgan-Lopez AA Pais JM Youth BMI trajectories: Evidence from the NLSY97 Obesity 2009 17 1274e80 19584884 6. Nour NM Health Consequences of Child Marriage in Africa Emerg Infect Dis 2006 12 11 1644 1649 10.3201/eid1211.060510 17283612 7. Assini-Meytin LC Kerry MS Green M Long-Term Consequences of Adolescent Parenthood Among African-American Urban Youth: A Propensity Score Matching Approach J Adolesc Health 2015 56 5 529 535 10.1016/j.jadohealth.2015.01.005 25769478 8. Gnudi S Sitta E Lisi L Relationship of body mass index with main limb fragility fractures in postmenopausal women J Bone Miner Res 2009 27 4 479 484 10.1007/s00774-009-0056-8 9. Morin S Tsang JF Leslie WD Weight and body mass index predict bone mineral density and fractures in women aged 40 to 59 years Osteoporos Int 2009 20 3 363 370 10.1007/s00198-008-0688-x 18633665 10. Hosegood V Campbell OM Body mass index, height, weight, arm circumference, and mortality in rural Bangladeshi women: a 19-y longitudinal study Am J Clin Nutr 2003 77 2 341 347 12540392 11. Hossain MG Mahumud RA Aik S Prevalence of child marriage among Bangladeshi women and trend of change over time J Biosoc Sci 2015 26286142 12. Kamal SMM Adolescent motherhood in Bangladesh: Evidence from 2007 BDHS data Can Stud Popul 2012 39 1–2 63 82 13. Khan MM Kraemer A Factors associated with being underweight, overweight and obese among ever-married non-pregnant urban women in Bangladesh Singapore Med J 2009 50 8 804 813 19710981 14. Shafique S Akhter N Stallkamp G de Pee S Panagides D Bloem MW Trends of urder-and overweight among rural and urban poor women indicate the double burden of malnutrition in Bangladesh Int J Epidemiol 2007 36 2 449 457 10.1093/ije/dyl306 17242024 15. Hossain MG Bharati P Aik S Lestrel PE Abeer A Kamarul T Body mass index of married Bangladeshi women: trends and association with socio-demographic factors J Biosoc Sci 2012 44 4 385 399 10.1017/S002193201200003X 22340969 16. Subramanian SV Perkins JM Khan KT Do burdens of underweight and overweight coexist among lower socioeconomic groups in India? Am J Clin Nutr 2009 90 2 369 376 10.3945/ajcn.2009.27487 19515733 17. Bharati S Pal M Bhattacharya BN Bharati P Prevalence and causes of chronic energy deficiency and obesity in Indian women Hum Biol 2007 79 4 395 412 10.1353/hub.2007.0048 18075004 18. Dunn OJ Clark VA Applied Statistics: Analysis of Variance and Regression 1974 Toronto John Wiley & Sons 19. Stevens J Applied Multivariate Statistics for the Social Sciences 1996 Mahwah Lawrence erlbaum associates 20. Godfrey K Bailar JC Mosteller F Simple linear regression in medical research Medical uses of statistics 1992 2 Boston NEJM Books 201 232 21. NIPORT ‘National Institute of Population Research and Training, Bangladesh Demographic and Health Survey Dhaka, Mitra & Associates and ORC Macro, Bangladesh and Calverton, MD, USA 2007 22. Flegal KM Graubard BI Williamson DF Gail MH Excess deaths associated with underweight, overweight, and obesity JAMA 2005 293 15 1861 1867 10.1001/jama.293.15.1861 15840860 23. World Health Organization (WHO) Diet, Nutrition and the prevention of Chronic diseases 2003 Geneva Report of a Joint WHO/FAO Expert Consultation 24. Khongsdier R Body mass index of adult males in 12 populations of northeast India Ann Hum Biol 2001 28 4 374 383 10.1080/03014460010004610 11459235 25. Chan YH Biostatistics 202: Logistic regression analysis Singapore Med J 2004 45 4 149 153 15094982 26. Pryer JA Rogers S Rahman A Factors affecting nutritional status in female adults in Dhaka slums, Bangladesh Soc Biol 2003 50 3–4 259 69 16382815 27. Adak DK Gautam RK Bharati S Gharami AK Pal M Bharati P Body Mass Index and Chronic Energy Deficiency of Adult Males of Central Indian Populations Hum Biol 2006 78 2 161 178 10.1353/hub.2006.0032 17036924 28. Mehta S Groenen R Roque F United Nations Social and Economic Commission for Asia and the Pacific, 1998. Adolescents in Changing Times: Issues and Perspectives for Adolescent Reproductive Health in the ESCAP Region 2006 29. Laura F Do Mothers’ Opinions Matter in Teens’ Sexual Activity? J Fam Issues 2005 26 7 947 974 10.1177/0192513X04272758 30. The National Campaign to Prevent Teen Pregnancy. Not Just Another Single Issue: Teen Pregnancy Prevention’s Link to Other Critical Social Issues, Report Published by, The National Campaign to Prevent Teen Pregnancy, Washington D.C., 2002. http://www.teepregnancy.org/product/pdf/NotSingleIssue.pdf. 31. World Health Organization (WHO) Sixty-fifth World Health Assembly 2012 32. World Health Organization (WHO) MPS Notes, Department of making pregnancy safer 2008 33. Henry EG Lehnertz NB Alam A Ali NA Williams EK Rahman SM Ahmed S Arifeen SE Baquia AH Peter J Winch ‘Sociocultural factors perpetuating the practices of early marriage and childbirth in Sylhet District, Bangladesh Int Health 2014 25294844 34. Hossain MG Islam S Aik S Kamarul TK Lestrel PE Age at menarche of university students in Bangladeshi: secular trends and association with adult anthropometric measures and socio-demographic factors J Biosoc Sci 2010 42 5 677 687 10.1017/S0021932010000210 20529410 35. Najat Y Alice A Abbass A Rizk S Eating habits and obesity among Lebanese university students Nutr J 2008 7 32 10.1186/1475-2891-7-32 18973661 36. Cilliers J Senekal M Kunneke E The association between the body mass index of first-year female university students and their weight-related perceptions and practices, psychological health, physical activity and other physical health indicators Public Health Nutr 2006 9 2 234 243 10.1079/PHN2005846 16571178

PMC005xxxxxx/PMC5000406.txt

==== Front Int J Behav Nutr Phys ActInt J Behav Nutr Phys ActThe International Journal of Behavioral Nutrition and Physical Activity1479-5868BioMed Central London 42210.1186/s12966-016-0422-6ResearchA 5-year longitudinal analysis of modifiable predictors for outdoor play and screen-time of 2- to 5-year-olds http://orcid.org/0000-0002-9488-8121Xu Huilan +61 2 9515 9074Huilan.Xu@sswahs.nsw.gov.au 12Wen Li Ming lmwen@email.cs.nsw.gov.au 12Hardy Louise L Louise.hardy@sydney.edu.au 3Rissel Chris Chris.rissel@sydney.edu.au 11 Sydney School of Public Health, University of Sydney, Sydney, NSW 2006 Australia 2 Health Promotion Unit, Sydney Local Health District, Level 9, King George V Building, Missenden Road, Camperdown, NSW 2050 Australia 3 Prevention Research Collaboration, Sydney School of Public Health, University of Sydney, Sydney, NSW 2006 Australia 26 8 2016 26 8 2016 2016 13 1 9616 6 2016 17 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Early childhood is a critical time for establishing physical activity and sedentary behaviours. Identifying modifiable predictors of physical activity and sedentary behaviours in the early life stages can inform the development of early intervention programs. The aim of this study was to identify modifiable predictors of outdoor play (a proxy of physical activity) and screen-time in 2- to 5-year-olds. Methods A longitudinal data analysis was conducted using 5-year follow-up data from the Healthy Beginnings Trial undertaken in Sydney, Australia from 2007 to 2013. A total of 667 pregnant women were recruited for the study. Information on mothers’ demographics, physical activity, screen-time, knowledge of child development, and awareness of childhood obesity during pregnancy (at baseline); children’s tummy time (a colloquial term describing the time when a baby is placed on his or her stomach while awake and supervised) at 6 months old and screen-time at 1 year old was collected via interviews with participating mothers as potential modifiable predictors. Main outcomes were children’s outdoor playtime and screen-time at ages 2, 3.5, and 5 years. Mixed linear and logistic regression models were built to determine these modifiable predictors. Results Mothers’ screen-time during pregnancy (β = 2.1, 95 % CI 0.17–4.12; P = 0.030) and children’s daily screen-time at age 1 year (β = 15.2, 95 % CI 7.28–23.11; P < 0.0001) predicted children’s daily screen-time across ages 2 to 5 years after controlling for confounding factors. Practising tummy time daily (β = 13.4, 95 % CI 1.26–25.52; P = 0.030), mother’s physical activity level (β = 3.9, 95 % CI 0.46–7.28; P = 0.026), and having been informed about playing with child at baseline (β = 11.6, 95 % CI 1.56–21.54; P = 0.023) predicted children’s outdoor playtime across ages 2 to 5 years. Conclusions Mothers played an important role in their children’s outdoor play and screen-time in the first years of live. Children’s early exposure to screen devices could be associated with their later screen-time. Early interventions to improve young children’s physical activity and sedentary behaviour should focus on improving pregnant women’s physical activity, awareness of playing with their child, reducing their own screen-time as well as practicing daily tummy time for infants after giving birth. Trial registration The Healthy Beginnings Trial is registered with the Australian Clinical Trial Registry (ACTRNO12607000168459). Registered 13 March 2007. Prospectively registered. Keywords Physical activityOutdoor playScreen-timePredictorthe Australian National Health and Medical Research CouncilPhase 1 ID number: 393112, Phase 2 ID number: 1003780Wen Li Ming issue-copyright-statement© The Author(s) 2016 ==== Body Background Physical activity and sedentary behaviours of young children have been gaining public health attention with increasing early onset and high prevalence of childhood overweight and obesity [1, 2]. Indeed, recommendations for how long children should spend in physical activity and sedentary behaviours each day have been developed in many countries to guide parents and carers of young children. The National Association for Sport and Physical Education in the U.S.A recommends that 3- to 5-year-olds should accumulate at least 60 min daily of structured physical activity; engage in at least 60 min daily unstructured physical activity; and should not be sedentary for more than 60 min at a time except when sleeping [3]. In Australia, the Department of Health recommends that 2- to 5-year-olds should be physically active for at least 3 h per day (accumulated throughout the day including light-, moderate-, and vigorous-intensity physical activity); spend less than one hour per day watching television and using other electronic media; and not be sedentary, restrained, or kept inactive for more than one hour at a time with the exception of sleeping [4]. The United Kingdom and Canada endorsed similar guidelines of physical activity and sedentary behaviour for young children [5–7]. Many people believe that young children are naturally physically active [8]. However, many 3- to 5-year-olds are not as physically active as they need to be for good health, with one-in-twenty and less than one-in-seven Australian preschool children meeting physical activity and screen-time recommendations respectively [9]. Previous systematic reviews examining objectively measured physical activity within child-care centers also concluded that physical activity levels of preschool children are typically very low and levels of sedentary behaviour are high [10, 11]. Ensuring adequate physical activity and preventing excessive sedentary behaviour, in particular, screen-time, not only provides benefits for weight control in children but also benefits their physical development and psychological well-being [12–14]. Therefore, interventions in improving children’s physical activity behavior during their early years will be important in preventing childhood overweight and obesity and improving children’s well-being [15]. In recent years, several systematic reviews have summarized the factors associated with physical activity and sedentary behaviours in young children [16–19]. Most studies included in systematic reviews were cross-sectional studies and identified correlates of physical activity and sedentary behaviours, many of which are not modifiable, such as children’s sex, age, ethnicity, socioeconomic status. Longitudinal studies that examine predictors of physical activity and sedentary behaviours in young children are scarce. Although identifying demographic and socioeconomic factors that are associated with physical activity and sedentary behaviours are important for targeted population interventions, identifying modifiable predictors can lead to more effective interventions. Among young children, physical activity occurs predominantly during active play. Outdoor playtime is often used as a proxy of physical activity of young children [20]. Therefore, the present study aimed to determine modifiable factors in early life that predict outdoor play and screen-time in 2- to 5-year-olds. Methods Study design A longitudinal data analysis was conducted using data extracted from the Healthy Beginning Trial (HBT), which was a 5-year randomised controlled trial undertaken in south-western Sydney, Australia, during 2007–2013. The details of the research protocol and the main outcomes of the HBT have been reported elsewhere [21, 22]. Briefly, the HBT assess the effectiveness of a staged home based early intervention in reducing early childhood obesity. The intervention comprises eight home visits from a specially trained community nurse promoting healthy feeding and physical activity. The intervention started from the gestation age of 30–36 weeks to children reaching 2 years old. Five-year follow up data were obtained from mothers at the late pregnancy to children’s ages at 6 months, 1, 2, 3.5, and 5 years old. The study was approved by the Ethics Review Committee of Sydney South West Area Health Service (Royal Prince Alfred Hospital Zone, X10-0312 & HREC/10/RPAH/546). Since the HBT was a randomised controlled trial the group allocation (intervention or control) was taken into account in the data analysis. Study participants A total of 667 first-time mothers at 24–34 weeks of pregnancy were recruited from antenatal clinics at Liverpool and Campbelltown Hospitals, located in south-western Sydney, Australia. The analysis included 497, 415, and 369 mother-child dyads that were retained at ages 2, 3.5, and 5 years. Data collection and measures Face-to-face interviews with each mother at their home were conducted by trained research nurses at baseline (30 to 36 weeks of pregnancy), 1, 2, 3.5, and 5 years follow-up. A telephone survey was conducted when each child was 6 months old. Potential modifiable predictors were assessed at baseline, 6 months, and 1 year follow-up. Children’s outdoor play and screen-time Children’s outdoor playtime was measured at ages 2, 3.5, and 5 years. To assess outdoor playtime, the mother was asked how much time her child spent playing outdoors on a typical weekday and on a typical weekend day with validated survey questions [20]. The mean outdoor playtime per day was calculated as (hours/weekday × 5 + hours/weekend day × 2)/7. Outdoor playtime was then categorised into ‘<2 h/day’ and ‘≥2 h/day’ (based on the median outdoor playtime around 2.3 h/day). Children’s screen-time was measured at ages 1, 2, 3.5, and 5 years. Survey questions from a national-wide study called Growing Up in Australia: The Longitudinal Study of Australian Children were used to assess screen-time [23]. The mother was asked to provide the total time her child spent on each of the activities including (i) watching TV programmes; (ii) watching DVDs or videos; (iii) using a computer; and (iv) playing with an electronic game system from Monday to Friday and also on weekends [23]. Screen-time was summed and the mean screen-time (hours per day) was calculated. Screen-time was then categorised into ‘<1 h/day’ and ‘≥1 h/day’ based on the screen-time recommendation for children aged 2–5 years [4]. For full-time employed mothers, they were asked to estimate their child’s outdoor playtime and screen-time based on the time when they were with the child. Mothers’ physical activity and screen-time Questions from the Active Australia Survey questionnaire 2003 [24] were used to assess mothers physical activity level before pregnancy, at 2, 3.5, and 5 years follow-up. Mother’s total physical activity time was calculated as hours per day and further categorised into ‘<150 min/week’ and ‘≥150 min/week’ based on the recommendation of physical activity for adults [25]. Mother’s screen-time was assessed by a question ‘Currently, on average, how many hours per day or per week do you spend sitting watching TV, videos, DVDs, playing computer or video games, or surfing the Internet for pleasure?’ It was reported by mothers at baseline, 2, 3.5 and 5 years. Screen-time was calculated as hours per day. Baseline screen-time was further categorised as ‘<3 h/day’ and ‘≥3 h/day’ based on the mean value of 3.1 h/day from the study participants at baseline. Mothers’ knowledge and awareness Mothers’ knowledge of child development at baseline was assessed by three questions including 1) How informed do you feel about child development e.g. age a child typically crawls, walks, runs? 2) How informed do you feel about playing with children? 3) How able do you feel you are to give your child activities that will keep them occupied while you are doing other things? The response was chosen from a 4-point Likert-type scale (very informed/able, somewhat informed/able, a little informed/able, and not at all informed/able). The responses were categorized into ‘yes’ or ‘no’, with ‘yes’ referring to ‘very informed/able’. Mothers’ awareness of childhood obesity at baseline was assessed by a question ‘How worried should adults be about their children being overweight or obese?’ The response was chosen from a 4-point Likert-type scale (Extremely worried, Very worried, A little worried, Not at all worried). The responses were categorized into ‘yes’ or ‘no’, with ‘yes’ referring to ‘Extremely/very worried’. Tummy time and playgroup Children’s tummy time is when a baby is placed on his or her stomach while awake and supervised. Children’s tummy time frequency and starting time were reported by mothers when children were 6 months old. Tummy time frequency was assessed by a question ‘How often does your child spend time on their tummy when they are awake?’ [26] Response options were ‘Not at all’, ‘1–2 days a week’, ‘3–4 days a week’, ‘5–6 days a week’, ‘Daily’. The responses were categorized into ‘daily’ and ‘not daily’. Age of starting tummy time was assessed by a question ‘At what age did your child start spending time on [his/her] tummy when [he/she] was awake?’ It was calculated as days of age and then categorised into within or after 1 month (30 days) of birth according to the mean starting tummy time 30 days. Attendance of childhood program or activity was measured at age 1 year by a question ‘Does your baby currently attend any play group or other early childhood program or activity?’ The responses were ‘yes’ or ‘no’. Mothers’ demographics Mothers’ demographic and socioeconomic information were collected at baseline, 1, 2, 3.5, and 5 years using questions from the NSW Child Health Survey 2001 [26]. All mothers’ demographic and socioeconomic information were categorized into groups (see Table 1).Table 1 Baseline characters of mothers and children at baseline, 2, 3.5, and 5 years Baseline demographics Baseline 2 years 3.5 years 5 years n (%) n (%) n (%) n (%) Mother’s age (years)a  16–24 279 (42) 185 (37) 140 (34) 117 (32)  25–29 226 (34) 176 (36) 153 (37) 139 (38)   ≥ 30 162 (24) 136 (27) 122 (29) 113 (30) Mother’s country of birth  Other 237 (36) 175 (35) 151 (36) 130 (35)  Australia 429 (64) 321 (65) 263 (64) 238 (65) Mother’s education statusb  Completed primary school/School Certificate 137 (21) 82 (16) 57 (14) 46 (12)  HSC to TAFE certificate or diploma 364 (55) 280 (57) 234 (56) 208 (57)  University 163 (24) 133 (27) 123 (30) 114 (31) Mother’s employment statusa  Employed 363 (54) 295 (59) 261 (63) 238 (65)  Other 303 (46) 201 (41) 153 (37) 131 (35) Mother’s marital statusa  Married/partner 584 (88) 452 (91) 382 (92) 343 (93)  Other 81 (12) 45 (9) 33 (8) 26 (7) Annual household income ($AUD)a   < 40,000 208 (31) 130 (26) 96 (23) 82 (22)  40,000–< 80,000 215 (32) 159 (32) 130 (31) 114 (31)   ≥ 80,000 244 (37) 208 (42) 189 (46) 173 (47) Child sex  Girls 333 (50) 249 (50) 202 (49) 183 (50)  Boys 330 (50) 248 (50) 213 (51) 186 (50) Note: sample size is not always 667, 497, 415, and 369 at baseline, 2, 3.5, and 5 years due to missing values aCompare to baseline distribution, the distribution of the baseline variable at ages 3.5 and 5 years were significantly different bCompare to baseline distribution, the distribution of the baseline variable at ages 2, 3.5 and 5 years were significantly different Other covariates Children’s night sleep duration, child-care/school attendance, TV time and programme viewing rules, whether TV is on all the time and during meals were measured at ages 2, 3.5 and 5 years. Children’s night sleep duration (hours per night) was measured using questionnaire items from the Prevention of Overweight in Infancy study developed from the consensus opinion of the researchers [27]. Child-care/school attendance was measured using questions from the state population health survey [26], the mother was asked “Is your child currently having any type of formal or informal child-care on a regular basis?” and “Does your child attend school yet?” Response options were “yes” or “no”. To assess TV rules for children, the mother was asked ‘Are there rules about what TV programmes your child can watch?’ and ‘Are there rules about how many hours of TV your child can watch?’ The mother was also asked ‘How often is a TV on when no one is watching?’ and ‘How often is a TV on during meals?’ Response options were ‘always’, ‘often’, ‘sometimes’, ‘rarely’ or ‘never’ [23]. The responses were regrouped as ‘yes’ or ‘no’ with ‘yes’ referring to ‘always’, ‘often’ or ‘sometimes’. Data analysis Statistical analyses were carried out using Stata 13 [28]. Mean and standard deviation or number and percentage were reported to summarise children’s outdoor play and screen-time, mothers’ demographics and other study factors. Children’s outdoor play and screen-time were analysed as both continuous and binary outcome variables. Considering the longitudinal design, random-intercept mixed models were built to take into account correlations between repeated measures. Also by building mixed models, to some extent, missing data can be implicitly imputed. Therefore, participants with partial missing data were still able to contribute information. Since children’s outdoor playtime and screen-time were included as both continuous and dichotomous outcomes, mixed linear and logistic models were built respectively. In order to examine whether predictors varied on weekday and weekend day, also overcome the limitation that working mothers and mothers whose children attending child-care might underestimated or overestimate children’s outdoor play and screen-time on weekday, mixed models were built for outdoor play or screen-time on weekdays, weekend days, and overall daily respectively. Two steps were taken to identify predictors of outdoor play and screen-time. First, the relationship between each potential predictor or confounder (such as demographics, correlates etc.) and children’s outdoor playtime and screen-time was examined by including each potential predictor or confounder in a mixed linear or logistic model that adjusted for time and quadratic slopes for time. The reason for including quadratic slopes for time was to improve model fit. Second, all variables significant in the first analysis with P < 0.25 were entered into a multivariable mixed model. All multiple mixed models were adjusted for allocation of intervention to control the intervention effect, time and quadratic slopes for time. The least significant variables were progressively dropped until only those with P < 0.05 remained. Child-care attendance, mothers’ employment status at 2, 3.5, and 5 years, variables that significantly predicted outdoor play and screen-time on either weekdays or weekend-days were also remained in models. Subsequently, the variables which were not included in the model were given an extra chance to enter the final model one by one to see whether they were predictors or confounders. Interaction between time and the potential predictors and confounding factors were also included in all models to test whether their effect varied across three time points. The interaction was excluded from the final model if the interaction was not statistically significant (P > 0.05). Visual inspection of residual plots did not reveal any obvious deviations from linearity or normality, indicating that a random-intercept mixed linear model was appropriate. Results There was no significant difference regarding mothers’ country of birth and children’s sex among mother-child dyads that were retained at ages 2, 3.5, and 5 years. Mothers lost-to-follow-up were typically young, unmarried, lower educated, unemployed, and had lower household income (see Table 1). The main study factors are summarised in Table 2 and show that mothers’ mean screen-time was 3.12 (SD 2.48) hours per day with nearly half of them having ≥3 h screen-time per day at baseline. Mothers’ mean physical activity before pregnant was 1.16 (SD 1.44) hours per week with 71 % mothers met physical activity recommendation. Most children practiced tummy time daily (79 %) and started tummy time (71 %) within 1 month of birth. The mean children’s screen-time at 1 year was 0.64 h per day (SD 0.82) with 74 % children met screen-time recommendation. Study outcomes were summarised in Table 3. Across ages 2 to 5 years, both children’s outdoor playtime (from 2.28 to 2.64 h per day) and screen-time (from 1.37 to 2.25 h per day) significantly increased.Table 2 Main study factors Mean (SD) n (%) Baseline variables  Mother’s screen-time 3.12 (2.48)    < 3 h/day 347 (52)    ≥ 3 h/day 318 (48)  Mother’s PA time 1.16 (1.44)    < 150 min/week 191 (29)    ≥ 150 min/week 476 (71)  Informed about playing with child   No 336 (51)   Yes 329 (49)  Informed about child development   No 538 (81)   Yes 127 (19)  Able to give child activities to keep them occupied   No 332 (50)   Yes 333 (50)  Awareness of childhood obesity   No 142 (21)   Yes 523 (79) 6 months variables  Start tummy time   After 1 month of birth 158 (29)   Within 1 month of birth 392 (71)  Tummy time frequency   Not daily 115 (21)   Daily 442 (79) 1 year variables  Child screen-time (hours/day) 0.64 (0.82)    ≥ 1 h/day 93 (26)    < 1 h/day 259 (74)  Play group attendance   No 359 (68)   Yes 166 (32) Note: sample size is not always 667, 561, and 527at baseline, 6 months, and 1 year due to missing values Table 3 Children’s screen-time and outdoor playtime at 2, 3.5, and 5 years Outcomes 2 years 3.5 years 5 years n (%) n (%) n (%) Child screen-time   ≥ 1 h/day 310 (64) 378 (91) 333 (90)   < 1 h/day 175 (36) 37 (9) 36 (10) Child outdoor playtime   < 2 h/day 181 (37) 135 (33) 120 (33)   ≥ 2 h/day 305 (63) 379 (67) 246 (67) Mean (SD) Mean (SD) Mean (SD) Child screen-time (hours/day) 1.37 (1.02) 2.48 (1.49) 2.25 (1.27) Child outdoor playtime (hours/day) 2.28 (1.21) 2.48 (1.28) 2.64 (1.37) Note: sample size is not always 497, 415, and 369 at 2, 3.5, and 5 years due to missing values Predictors of children’s outdoor play and screen-time across ages 2 to 5 years The group allocation was not significantly associated with either children’s outdoor playtime or screen-time. Results of multiple mixed linear models are shown in Table 4. Mothers’ daily screen-time at baseline and children’s daily screen-time at 1 year were positively associated with children’s screen-time across ages 2 to 5 years after adjusting for time, time2, intervention allocation, mothers’ country of birth, children’s night sleep duration, child-care attendance, TV is on all the time, TV time rules, mothers’ employment status and screen-time at 2, 3.5, and 5 years. Each one hour increase in mothers’ daily screen-time at baseline was associated with 2 min more children’s daily screen-time (95 % CI 0.17–4.12) and screen-time on a weekday (95 % CI 0.26–4.41). Each one hour increase in children’s daily screen-time at 1 year was associated with 15 min (95 % CI 6.21–22.90) and 18 min (95 % CI 6.40–28.83) more screen-time on a weekday and weekend day respectively. Compared with children whose mothers had less than 3 h screen-time per day at baseline, children whose mothers had ≥3 h daily screen-time at baseline had 12 min (95 % CI 1.93–21.74) and 15 min (95 % CI 2.27–27.60) more screen-time on a weekday and weekend day respectively.Table 4 Predictors of children’s screen-time and outdoor play time at 2, 3.5 and 5 years Variables Children’s screen-time (minutes/day)a Week day Weekend day Daily β (95 % CI) β (95 % CI) β (95 % CI) Mothers’ baseline screen-time (hours/day) 2.3 (0.26 to 4.41) 1.7 (-1.02 to 4.34) 2.1 (0.17 to 4.12) Children’s screen-time at 1 year (hours/day) 14.6 (6.21 to 22.90) 17.6 (6.40 to 28.83) 15.2 (7.28 to 23.11) Mothers’ baseline screen-time   < 3 h/day _ _ _   ≥ 3 h/day 11.8 (1.93 to 21.74) 14.9 (2.27 to 27.60) 12.9 (3.43 to 22.3) Children’s outdoor playtime (minutes/day)b Mothers’ baseline PA (hours/day) 2.7 (-0.70 to 6.18) 6.5 (2.06 to 11.04) 3.9 (0.46 to 7.28) Baseline informed about playing with children  No _ _ _  Yes 9.4 (-0.65 to 19.47) 14.9 (1.75 to 28.05) 11.6 (1.56 to 21.54) Tummy time frequency  Not daily _ _ _  Daily 12.6 (0.37 to 24.81) 16.6 (0.60 to 32.53) 13.4 (1.26 to 25.52) Mothers’ baseline PA   < 150 min/week _ _ _   ≥ 150 min/week −0.8 (-11.78 to 10.13) 5.3 (-9.15 to 19.67) 0.37 (-10.53 to 11.28) PA physical activity aAll models are adjusted for time, time2, intervention allocation, mothers’ country of birth; children’s night sleep duration, child-care attendance, TV time rules, TV is on all the time, mothers’ employment status and screen-time at 2, 3.5, and 5 years bAll models are adjusted for time, time2, intervention allocation, child sex, mothers’ country of birth and education status at baseline; child-care attendance, mothers’ employment status, TV time rules, interaction of TV time rules and time at 2, 3.5 and 5 years Mothers’ physical activity level and being informed about playing with children at baseline and children’s tummy time frequency were positively associated with children’s outdoor playtime across ages 2 to 5 years after adjusting for time, time2, intervention allocation, child sex, mothers’ country of birth and education status at baseline, child-care attendance, mother’s employment status, TV time rules, interaction of TV time rules and time at 2, 3.5 and 5 years. Each one hour increase in mother’s physical activity time before pregnancy was associated with 6 min (95 % CI 2.06–11.04) more children’s outdoor playtime on a weekend day and 4 min (95 % CI 0.46–7.28) daily outdoor playtime. Children whose mothers having been informed about playing with children at baseline had 12 min (95 % CI 1.56–21.54) more daily outdoor playtime and 15 min (95 % CI 1.75–28.05) more outdoor playtime on a weekend day. Children having tummy time daily had 13 (95 % CI 0.37–24.81) and 17 min (95 % CI 0.60–32.53) more outdoor playtime on a weekday and weekend day respectively. Whether children started tummy time within one month of birth was not associated with children’s outdoor playtime. Whether mothers met the physical activity recommendation and whether children attended play group or other early childhood program were not associated with children’s outdoor playtime. Mothers’ knowledge of child development, belief of being able to give child activities, and high awareness of childhood obesity were not associated with either outdoor play or screen-time. When examining predictors of children meeting the screen-time recommendation across ages 2 to 5 years, only children’s daily screen-time at age 1 year was a significant predictor. Children who had longer screen-time at age 1 year were less likely to meet the screen-time recommendation at ages 2 to 5 years with adjusted odds ratio (AOR) 0.55 (95 % CI 0.34–0.90) for daily screen-time recommendation after adjusting for potential confounders listed earlier. Children whose mothers having been well informed about playing with children at baseline were more likely to play outdoor ≥ 2 h per day on a weekend day with AOR 1.65 (95 % CI 1.06–2.55); children who started tummy time within 1 month of birth were more likely to play outdoor ≥ 2 h per day on a weekday with AOR 1.44 (95 % CI 1.00–2.08) after controlling for confounding factors. Mothers’ physical activity time at baseline and children’s tummy time frequency were not predictors of whether children played outdoor ≥ 2 h per day. Correlates of children’s outdoor play and screen-time across ages 2 to 5 years Multiple mixed linear models also showed that children with Australia born and employed mother had around 16 min (95 % CI 6.2–25.8) and 14 min (95 % CI 5.0–23.2) less screen-time per day; children who had longer night sleep duration, attended child-care, and had TV time rules had 6 min (95 % CI 1.6–10.2), 11 min (95 % CI -0.6–22.5), 9 min (95 % CI 0.8–18.1) less screen-time per day respectively; children whose mothers had lower than mean screen-time (2 h per day) at 2, 3.5, and 5 years had 19 min (95 % CI 10.2–27.6) less screen-time per day; children from family that TV is on all the time and TV is on during meals had 20 min (95 % CI 10.8–28.9) and 23 min (95 % CI 14.5–31.2) more screen-time per day. Boys were more active than girls accumulating an additional 16 min (95 % CI 5.6–25.3) more outdoor playtime per day. Children whose mothers were Australian born had 25 min (95 % CI 14.3–36.2) more outdoor playtime per day. Children from families carrying TV time rules had 22 min (95 % CI 8.8–35.5) more outdoor playtime per day at age 2 years but not at ages 3.5 and 5 years. Discussion To our knowledge, this was the first study identifying modifiable predictors of outdoor play and screen-time of young children using 5-year longitudinal data. We found that mothers’ own screen-time at baseline and children’s daily screen-time at age 1 year predicted children’s daily, weekday, and weekend day screen-time across ages 2 to 5 years; practising tummy time daily predicted children’s daily, weekday, and weekend day outdoor playtime across ages 2 to 5 years; mother’s baseline daily physical activity time and having been well informed about playing with children significantly predicted children’s daily and weekend day outdoor playtime. Although the effect sizes found in this study were relatively small, the findings could potentially have public health significance at a population level. The findings reinforce that mothers own physical activity and screen-time behaviours are important to their children’s physical activity and screen-time in the first 5 years of live; and children’s early exposure to screen devices can lead to longer screen-time at ages 2 to 5 years. The findings also suggested that young children’s outdoor play and screen-time were influenced by different factors. Mothers’ screen-time and the home screen environment (‘TV is on all the time’ and ‘TV is on during meal’) at 2, 3.5 and 5 years were significantly associated with children’s screen-time on both weekdays and weekend days across ages 2 to 5 years after adjusting for mother’s baseline screen-time and other confounders. We found that the association between mothers’ and children’s screen-time are comparable with the findings from previous systematic reviews [18, 19, 29] that mothers’ screen-time serves as both a predictor and correlate of young children’s screen-time. This highlights the significant impact of mothers’ own screen viewing behaviour and practice on their children’s screen-time. Children from families with TV time rules, Australia born mothers and mothers who were employed had significantly less screen-time on weekdays but not on weekend days. This finding may indicate that independent of differences in culture, employment status, and family rules, most mothers would consider the weekend is time for entertainment, and therefore, allow children more screen-time. Mothers’ pre-pregnancy physical activity level and being informed about playing with their child had a significant impact on children’s outdoor play on weekend days but not on weekdays. This may reflect families having more free time on the weekend than weekdays and those mothers tend to arrange more outdoor play for their children or themselves as well. The timing of when tummy time was introduced was not significantly associated with children’s outdoor play while practising tummy time daily were associated with outdoor playtime on both weekdays and weekend days. It might suggest that frequency of having tummy time is more important in relation to young children’s physical activity. Almost all existing studies about tummy time focused on the relationship between tummy time and developmental milestones in infant. A systematic review found that infants who did tummy time when awake achieved developmental milestones significantly earlier than those who did not or who did limited tummy time when awake in the first 6 months of life [30]. Two other studies also found that tummy time is associated with certain motor milestones achieved during early life [31, 32]. However, whether tummy time has a long term impact on children’s physical activity has not been studied yet. The findings from the present study suggested that promoting tummy time during infancy may lead to increased physical activity in young children. It might be because mothers who practise tummy time earlier and more frequently are more likely to support and encourage their children to play more actively. Analogous with previous systematic reviews [16, 17], the present study also found boys were more physically active than girls. Children with Australia born mothers played more time outdoors on both weekdays and weekend days. These findings are also important because it helps to develop interventions for targeted population. Having TV time rules at ages 2, 3.5, and 5 years was associated with more children’s outdoor playtime at age 2 years but not at 3.5 and 5 years. The relationship between TV time rules and children’s outdoor playtime needs to be further explored. Mothers’ knowledge of child development, belief of being able to provide child activities, and high awareness of childhood obesity were not associated with outdoor play and screen-time in 2- to 5-year-olds after adjusting for confounding factors. It may suggest that intervention should be more focused on improving mothers’ skills, practice and own physical activity and sedentary behaviours not only knowledge. Several limitations of the study need to be considered when interpreting the findings. First, children’s outdoor play and screen-time was reported by mothers that may be subject to recall bias. Also mothers who were employed and those whose children attended child-care might underestimate or overestimate children’s outdoor play and screen-time on weekdays. In order to remedy this limitation, mothers’ employment status and children’s attendance of child-care at 2, 3.5 and 5 years were adjusted, and predictors of outdoor play and screen-time were examined on weekdays and weekend days separately. Second, some important factors were not included in analyses, such as the environment the children were in and mother-child interaction in physical activity. The environment where the children were in may influence the amount of outdoor play and screen-time. However, the environmental variables such as ‘park nearby’, or ‘type of accommodation a child lived in’ were not consistently collected through the 5 year data collection. Therefore, they were not included in analyses. Mother-child interaction in physical activity might be an important factor that promotes young children being physically active. Several previous systematic reviews found that parent-child physical activity interaction and parental encouragement of being physically active were strongly associated with young children’s physical activity level [16, 17, 33]. A recent study showed that the time spent being physically active with their mother at 9-months predicted children’s physical activity at 19-months of age [34]. Third, the study was conducted in South West Sydney, Australia, an area with a relatively low socio-economic level which could limit the generalizability of the study. In addition, this study used data from the Healthy Beginnings trial which was not designed for this purpose and potential bias could be introduced. To address this limitation, we included the intervention allocation in model building. Conclusion Young children’s screen-time was significantly influenced by their mothers’ screen-time and parenting practice on screen viewing. Also early exposure to screen devices could be associated with children’s later screen-time. The early introduction and frequency of tummy time, and informing mothers the importance of playing with their young children could be foundations for children’s future physical activity. Early interventions in improving young children’s physical activity and screen-time should focus on improving pregnant women’s physical activity, awareness of the benefits of playing with their child, reducing their own screen-time as well as providing their children with regular periods of daily tummy time after giving birth. Abbreviations 95 % CI95 % confidence interval AORAdjusted odds ratio HBTHealthy Beginnings Trial. Acknowledgements This study is part of the Healthy Beginnings Trial funded by the Australian National Health and Medical Research Council (Phase 1 ID number: 393112, Phase 2 ID number: 1003780). We wish to thank mothers and children for their participation in this study. We also wish to thank members of the project team based at Health Promotion Service, South Western Sydney & Sydney Local Health Districts. Funding This study is part of the Healthy Beginnings Trial funded by the Australian National Health and Medical Research Council (Phase 1 ID number: 393112, Phase 2 ID number: 1003780). Availability of data and materials Data would be available on request. Authors’ contributions LMW and CR conceived and designed the Healthy Beginnings Trial and contributed to the development of the trial from which the data were extracted. LMW and HX conceived and designed the idea for the paper. HX undertook the literature review, data analysis and interpretation, and wrote the original draft. LMW, LLH, and CR commented on the draft. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication The manuscript does not contain any individual data. Therefore, it is not applicable. Ethics approval and consent to participate The study was approved by the Ethics Review Committee of Sydney South West Area Health Service (Royal Prince Alfred Hospital Zone, X10-0312 & HREC/10/RPAH/546). Written consent was obtained from all participating women before they were entered into the study. ==== Refs References 1. de Onis M Blossner M Borghi E Global prevalence and trends of overweight and obesity among preschool children Am J Clin Nutr 2010 92 1257 1264 10.3945/ajcn.2010.29786 20861173 2. Commonwealth of Australia. 2007 Australian National Children’s Nutrition and Physical Activity Survey- Main Findings. 2008. https://www.health.gov.au/internet/main/publishing.nsf/Content/8F4516D5FAC0700ACA257BF0001E0109/$File/childrens-nut-phys-survey.pdf. Accessed 10 May 2016. 3. National Association for Sport and Physical Education Active start: a statement of physical activity guidelines for children birth to five years 2002 Reston American Alliance for Health, Physical Education, Recreation and Dance 4. The Department of Health Move and Play Every Day National Physical Activity Recommendations for Children 0–5 Years 2010 5. UK Department of Health Physical Activity Guidelines for Early Years (Under 5s) – For Children Who Are Capable Of Walking 2011 6. Canadian Society for Exercise Physiology Canadian physical activity guidelines for the early years 2012 7. Canadian Society for Exercise Physiology Canadian sedentary behaviour guidelines for the early years 2012 8. Adamo KB Papadakis S Dojeiji L Turnau M Simmons L Parameswaran M Cunningham J Pipe AL Reid RD Using path analysis to understand parents’ perceptions of their children’s weight, physical activity and eating habits in the Champlain region of Ontario Paediatr Child Health 2010 15 9 e33 e41 22043145 9. Hinkley T Salmon J Okely AD Crawford D Hesketh K Preschoolers’ physical activity, screen time, and compliance with recommendations Med Sci Sport Exer 2012 44 458 465 10.1249/MSS.0b013e318233763b 10. Reilly JJ Low levels of objectively measured physical activity in preschoolers in child care Med Sci Sports Exerc 2010 42 3 502 507 10.1249/MSS.0b013e3181cea100 20068499 11. Vanderloo LM Screen-viewing among preschoolers in childcare: a systematic review BMC Pediatr 2014 14 205 10.1186/1471-2431-14-205 25129567 12. Biddle SJ Gorely T Stensel DJ Health-enhancing physical activity and sedentary behaviour in children and adolescents J Sports Sci 2004 22 8 679 701 10.1080/02640410410001712412 15370482 13. Burdette HL Whitaker RC Resurrecting free play in young children: Looking beyond fitness and fatness to attention, affiliation, and affect Arch Pediatr Adolesc Med 2005 159 1 46 50 10.1001/archpedi.159.1.46 15630057 14. Carson V Kuzik N Hunter S Wiebe SA Spence JC Friedman A Tremblay MS Slater LG Hinkley T Systematic review of sedentary behavior and cognitive development in early childhood Prev Med 2015 78 115 122 10.1016/j.ypmed.2015.07.016 26212631 15. World Health Organization Report of the commission on ending childhood obesity 2016 16. Hinkley T Crawford D Salmon J Okely AD Hesketh K Preschool children and physical activity: a review of correlates Am J Prev Med 2008 34 5 435 441 10.1016/j.amepre.2008.02.001 18407012 17. Emma L Jarrett M How we play—cultural determinants of physical activity in young children 2010 18. Cillero IH Jago R Systematic review of correlates of screen-viewing among young children Prev Med 2010 51 1 3 10 10.1016/j.ypmed.2010.04.012 20417227 19. Duch H Fisher EM Ensari I Harrington A Screen time use in children under 3 years old: a systematic review of correlates Int J Behav Nutr Phys Act 2013 10 102 10.1186/1479-5868-10-102 23967799 20. Burdette HL Whitaker RC Daniels SR Parental report of outdoor playtime as a measure of physical activity in preschool-aged children Arch Pediat Adol Med 2004 158 353 357 10.1001/archpedi.158.4.353 21. Wen LM Baur LA Rissel C Wardle K Alperstein G Simpson JM Early intervention of multiple home visits to prevent childhood obesity in a disadvantaged population: a home-based randomised controlled trial (Healthy Beginnings Trial) BMC Public Health 2007 7 76 10.1186/1471-2458-7-76 17490492 22. Wen LM Baur LA Simpson JM Rissel C Wardle K Flood VM Effectiveness of home based early intervention on children’s BMI at age 2: randomised controlled trial BMJ 2012 344 1 11 10.1136/bmj.e3732 23. FAHCSIA, AIFS, ABS. Growing Up in Australia: The Longitudinal Study of Australian Children. Wave 2.5 Mailback Questionnaire (B). 2016. http://www.growingupinaustralia.gov.au/studyqns/wave2qns/w2-5labelledqns/W2p5BLabelled.pdf. Accessed 13 May 2016. 24. AIHW The Active Australia Survey: A Guide and Manual for Implementation, Analysis and Reporting 2003 Canberra AIHW 25. World Health Organization Global Recommendations on Physical Activity for Health 2010 26. Centre for Epidemiology Research. NSW child health survey 2001. NSW Public Health Bull. 2002;13(S–4):1-86. 27. Taylor BJ Heath AL Galland BCW Prevention of Overweight in Infancy (POI.nz) study: a randomised controlled trial of sleep, food and activity interventions for preventing overweight from birth BMC Public Health 2011 11 942 10.1186/1471-2458-11-942 22182309 28. StataCorp Stata Statistical Software (Version Release 13) 2013 College Station StataCorp LP 29. Craemer M de Decker E de Bourdeaudhuij I Vereecken C Deforche B Manios Y Cardon G ToyBox-study group Correlates of energy balance-related behaviours in preschool children: a systematic review Obes Rev 2012 13 s1 13 28 10.1111/j.1467-789X.2011.00941.x 22309062 30. Pin T Eldridge B Galea MP A review of the effects of sleep position, play position, and equipment use on motor development in infants Dev Med Child Neurol 2007 49 11 858 867 10.1111/j.1469-8749.2007.00858.x 17979866 31. Dudek-Shriber L Zelazny S The effects of prone positioning on the quality and acquisition of developmental milestones in four-month-old infants Pediatr Phys Ther 2007 19 48 55 10.1097/01.pep.0000234963.72945.b1 17304097 32. Kuo YL Liao HF Chen PC Hsieh WS Hwang AW The influence of TT on motor development in early life J Dev Behav Pediatr 2008 29 367 376 10.1097/DBP.0b013e3181856d54 18766114 33. Mitchell J Skouteris H McCabe M Skouterisa H McCabea M Ricciardellia LA Milgromb J Baurc LA Fuller-Tyszkiewicza M Dwyerc G Physical activity in young children: a systematic review of parental influences Early Child Dev Care 2012 182 11 1411 1437 10.1080/03004430.2011.619658 34. Hnatiuk J Salmon J Campbell KJ Ridgers ND Hesketh KD Early childhood predictors of toddlers’ physical activity: longitudinal findings from the Melbourne InFANT Program Int J Behav Nutr Phys Act 2013 10 123 10.1186/1479-5868-10-123 24188589

PMC005xxxxxx/PMC5000407.txt

==== Front BMC PsychiatryBMC PsychiatryBMC Psychiatry1471-244XBioMed Central London 101310.1186/s12888-016-1013-4DebateRecommendations for the transition of patients with ADHD from child to adult healthcare services: a consensus statement from the UK adult ADHD network Young Susan susan.young1@imperial.ac.uk 1Adamou Marios 2Asherson Philip 3Coghill David 4Colley Bill 5Gudjonsson Gisli 3Hollis Chris 6McCarthy Jane 3Müller Ulrich 7Paul Moli 8Pitts Mark 9Arif Muhammad 101 Centre for Mental Health, Imperial College London, Broadmoor Hospital, WLMHT, London, UK 2 School of Human and Health Sciences, University of Huddersfield, Queensgate, Huddersfield UK 3 Institute of Psychiatry, Psychology & Neuroscience (IoPPN), King’s College London (KCL), London, UK 4 University of Dundee, Dundee, UK 5 Education and Children’s Services, Perth, UK 6 Division of Psychiatry and Applied Psychology, Institute of Mental Health, University of Nottingham, Nottingham, UK 7 Cambridge University, Cambridge, UK 8 Warwick Medical School, Coventry, UK 9 South London and Maudsley NHS Foundation Trust, London, UK 10 Leicester Adult ADHD Service, Leicester, UK 26 8 2016 26 8 2016 2016 16 1 30129 1 2016 18 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.The aim of this consensus statement was to discuss transition of patients with ADHD from child to adult healthcare services, and formulate recommendations to facilitate successful transition. An expert workshop was convened in June 2012 by the UK Adult ADHD Network (UKAAN), attended by a multidisciplinary team of mental health professionals, allied professionals and patients. It was concluded that transitions must be planned through joint meetings involving referring/receiving services, patients and their families. Negotiation may be required to balance parental desire for continued involvement in their child’s care, and the child’s growing autonomy. Clear transition protocols can maintain standards of care, detailing relevant timeframes, responsibilities of agencies and preparing contingencies. Transition should be viewed as a process not an event, and should normally occur by the age of 18, however flexibility is required to accommodate individual needs. Transition is often poorly experienced, and adherence to clear recommendations is necessary to ensure effective transition and prevent drop-out from services. issue-copyright-statement© The Author(s) 2016 ==== Body Background Transition can be defined as ‘the process of change from one stage (state) to another stage (state)’. There are many transitions that an individual makes throughout life, be they through psychosocial stages of development [7] or stages of cognitive development [22], and perhaps the most important transition is from adolescence to adulthood. This is a period of immense significance both in social and psychological terms, as well as in legal terms (i.e. legally becoming an adult at age 18). This is a period where young people are likely to move away from home, either to go to college/university or for job opportunities, start new relationships and assume new roles and responsibilities. This is also a period when many experiment in life, including alcohol and illicit substances. Any disruption during this period is likely to have longer lasting adverse effects on the young person’s development, realisation of their potential and possibly affect their psychological wellbeing. Young people with mental health conditions are particularly vulnerable during the transition period, and disruption of care during transition adversely affects the health, wellbeing and potential of this vulnerable group [15, 25]. Poor transition leads to disruption in continuity of care, disengagement from services and is likely to lead to poor clinical outcomes [25]. Ideally, transition should be a planned, orderly and purposeful process, taking into account developmental and illness specific needs [24]. For the majority of patients transition is poorly planned, poorly executed and poorly experienced [25, 26]. ADHD and transition Attention deficit hyperactivity disorder (ADHD) is estimated to affect up to five per cent of school-age children and adolescents in the UK [20], with a peak incidence in those aged between six and 12 years [18]. Although ADHD was previously thought of as a childhood condition, meta-analysis of follow-up studies found that 15 % of children with ADHD continued to have clinical features that met formal diagnostic criteria at the age of 25 years, and a further 50 % continued to suffer significant impairment into adulthood [8, 31]. In addition, comorbid problems such as mood and anxiety disorders and substance misuse persist or develop in adulthood. Other associated problems that may beset adults include difficulties in interpersonal relationships, problems with further education, poor employment opportunities, a propensity for dangerous risk-taking behaviours, and high rates of offending [1, 32]. For a review of ADHD in adolescence, see [30]. This constellation of clinical problems and functional impairments emphasises the clear need for mental health services for adolescents and adults with ADHD. Yet data ascertained from 1,636 patients in the UK General Practitioner Research Database, containing data of over three million active patients between the eight year period 1999–2006, reported a great reduction in prescribing patterns during adolescence, in fact hardly any patients were receiving ADHD medication by age 21 [17]. Given the findings of Faraone et al. [8] and Cheung et al. [5], this drop in service utilisation cannot possibly reflect spontaneous remission of ADHD symptoms. Of greater concern is that it occurs at a time when young people need the support of these services the most, as adolescence is a risk period when mental health problems may become more complex and serious disorders emerge. NICE guidelines In 2008, the National Institute for Clinical Health and Excellence (NICE) produced guidance about the recognition and management of ADHD across the life span. They emphasised that ADHD was a life-long condition and provided clear recommendations that care should continue from childhood to adulthood if symptoms persist. They also provided, for the first time in the UK, clear guidance for the development of transition arrangements from child to adult services:Young people with ADHD who are receiving healthcare from child and adolescent mental health services (CAMHS) or paediatric services should normally be transferred to adult mental health services (AMHS) if they continue to have significant symptoms of ADHD or other coexisting conditions that require treatment. Transition should be planned in advance by both the referring service and the receiving service. If needs are severe and/or complex, use of the Care Programme Approach (CPA) should be considered. Patients should be reassessed at school-leaving age to establish the need for continuing treatment into adulthood. If treatment is necessary, arrangements should be made for a smooth transition to adult services, with details of the anticipated treatment and services that the young person will require made clear. Precise timing of the transition arrangements may vary locally, but the transfer should usually be completed by the time the patient is 18 years of age. During the transition to adult services, a formal meeting involving CAMHS and/or the paediatric service and AMHS should be considered, and full information about the adult service should be provided to the patient. For young people aged 16 years and older, the CPA should be used as an aid to transfer between services. The young person, and when appropriate the parent/carer, should be involved in the planning. After transition, a comprehensive patient assessment should be carried out. The assessment should look at personal, educational, occupational and social functioning, and it should evaluate any coexisting conditions, notably drug misuse, personality disorders, emotional problems and learning difficulties. Trusts should ensure that specialist ADHD teams for children, young people and adults jointly develop age-appropriate training programmes for the diagnosis and management of ADHD. Common practice in the UK Although publication of the NICE guidelines has raised awareness of the need for appropriate arrangements for the transfer of young people with ADHD from child to adult healthcare services, AMHS in the UK for ADHD remain patchy. Furthermore, there is significant variation in the ways that services are organized across the UK and it is likely that some of this variation impacts significantly on the effectiveness of transition between adolescent and adult services. Healthcare for children with ADHD is generally provided by either community paediatric services or by CAMHS, or sometimes a combination of the two, that may or may not work in a joined-up manner. The interface between these services and AMHS is often complex with the potential for significant gaps. A survey conducted by Hall et al. [13] reported the absence of joint working between CAMHS and AMHS to be 66 % and 59 % of child and adult services, respectively. Furthermore, all 24 respondents from adult services reported having no official transition policy in place, and 89 % of National Health Service (NHS) Mental Health Trusts reported an absence of dedicated staff to support the transition. Surveys of NHS Trusts have also indicated a nationwide lack of accurate data pertaining to both the number of adults with ADHD within individual trusts, and the number of children/adolescents transitioning to adult services [12]. There seems to be a difference in practice between community paediatric and child & adolescent mental health services. Traditionally, paediatric services have stopped providing care for medical problems early in adolescence, although most now continue to provide care until school-leaving age. In the past most CAMHS generally stopped providing care either at 16 years of age or at school-leaving age (whichever was later). However, there is now a shift towards CAMHS continuing to provide care for patients up to 18 years of age regardless of school status. It has been suggested that AMHS are often too rigid in their age criteria for transition, with reports of refusals to consider transition until the individual reaches the age of 18 [3]. To further complicate issues it is reported anecdotally that whatever the technical cut-off age may be, many paediatric and CAMHS teams continue to see young people well past this age, due to the difficulties in transferring care to adult services [33]. Hence there is a mismatch in the expectation about when care should be passed from adolescent to adult services. Such inconsistencies have been reported to be disruptive and unsettling for patients [28] and this disruption to continuing care is detrimental to overall clinical care from both an individual and service perspective. Expert workshop In order to facilitate further discussion about the transition from child to adult services and develop more explicit and comprehensive recommendations for clinicians, commissioners and policy makers, an expert workshop – “ADHD: Transition from Adolescence to Adult” – was convened in June 2012 in London by the UK Adult ADHD Network (UKAAN) [29]. UKAAN is an organisation founded in 2009 by a group of mental health specialists in response both to the NICE guidelines [19] and to recommendations from the British Association for Psychopharmacology (BAP) [21] that aims to provide support, education, research and training for mental health professionals working with adults with ADHD. The workshop was attended by experts in the field of ADHD, working across services, together with allied professionals and patients. The workshop consisted of a series of presentations summarising the transition process from the perspective of these experts. The workshop first considered the findings of the TRACK study [27] which sought to identify factors that facilitate or impede effective transition more generally within the healthcare system. The National Institute for Health and Care Excellence,[19]) guidelines for transitioning young people with ADHD and the extension to these recommendations proposed by Young et al. [33] were reviewed, and consideration was given to the broader clinical, educational, occupational, and social needs of young people in this age group. There followed an in depth discussion tabling a consensus for detailed recommendations on the provision of transition services for patients with ADHD. Setting the Context – Findings from the TRACK Study The expert workshop commenced with a summary of the findings from the Transitions of Care from CAMHS to AMHS (TRACK) study [27]. The specific aims of the TRACK study were to identify factors that impede or facilitate effective transitions and continuity of care across the mental health services for children and adolescents, covering all disorders and not limited to ADHD, and to make recommendations about the configuration of services. Healthcare transitions are commonly defined in the literature as a purposeful, planned process that addresses the medical, psychosocial and educational or vocational needs of young people and young adults with chronic physical and medical conditions as they move from child centred to adult orientated healthcare systems. An important context for considering healthcare transition for young people with long-term health conditions is that they typically face several types of transition simultaneously – developmental transitions from childhood to adolescence to adulthood, situational transitions from child services to adult services, and possibly even transitions from relative health to illness or, indeed, sometimes in the other direction. As young people make these transitions they generally take on more responsibility from their parents/carers for all aspects of their healthcare, as well as for other areas of their life. As a consequence it is inevitable that, at the same time, parents/carers face a transition process of their own. Hence the transition from child to adult health services brings service-related challenges associated with an inherent shift from a family-orientated service to an individual-orientated service. Whilst the TRACK study focused on those transitioning between child and adult services, not all patients with ADHD will necessarily be in contact with child services at the time of transition. Indeed many of those with ADHD will not have received specific healthcare for their problems during childhood and adolescence – and may be presenting formally for the first time at transition from adolescence to adulthood. The TRACK project was a multi-site mixed models study, undertaken in six mental health trusts within England. The key findings were:Whilst many of the identified transition protocols were based on policy documents, there was evidence of a policy-practice gap, as services differed considerably on their interpretation of both policy and the recommendations for the practical aspects of transition. Although most of the protocols identified the patient as central to the transition process, few, if any, specified ways of preparing them for it. Optimal transitions were defined as those demonstrating evidence of: information transfer across services; a period of joint working between CAMHS and AMHS; transition planning involving key professionals from both services, patients and/or their parent/carers; and continuity of care. Transitions were considered sub-optimal if they failed to meet one or more of these criteria. The main findings were that, whilst around 80 % of patients were considered suitable for transition, around one-third were not referred to adult services. Adult services accepted the vast majority of referrals, however of those accepted around 20 % were discharged without being seen and only four per cent experienced optimal transition. An organisational analysis showed that CAMHS were seen as being more person-centred, holistic and family orientated, and adult services as more medication focused and crisis orientated. Facilitators for transition included: dedicated transition posts, joint working, early communication and greater involvement of carers. Barriers for transition included variability of eligibility criteria, differing thresholds between AMHS and CAMHS, communication problems between services and lack of confidence amongst adult staff on managing young people in general, and specifically relating to those with neurodevelopmental disorders. In a qualitative study of patients’, carers’ and mental health professionals’ experiences of transitions, participants described the importance of transfer planning meetings, joint working between the two services, and good information transfer. Interestingly most young people preferred not having their parents/carers involved with their care at AMHS, and many said that they appreciated the break from this family orientated system because they wanted to be more autonomous. On the other hand parents/carers wanted greater involvement with adult services. Despite making up a large proportion of the caseloads of CAMHS services, children with ADHD were relatively under-represented within the TRACK sample, making up just under 10 % of the total sample. Only six out of the 15 cases reviewed were transitioned to AMHS and none of these were classified as being ‘optimally transitioned’. Overall those young people with neurodevelopmental disorders such as ADHD and autism spectrum disorders, and those with emotional disorders or emerging personality disorders, were more likely to fall through the gaps. Those with serious mental illness, such as psychosis, patients who are receiving pharmacological treatment and with history of inpatient treatment are more likely to be accepted for transition. Patients’ perspectives There followed representations from patients on their experience and knowledge of the transition process. The transition period is a time of changing responsibilities, in which the young person is assuming greater autonomy and increased responsibility for his or her own care. Nevertheless, it is important that parents/carers should feel that they are involved in the transition process. This can be challenging because young people do not always want their parents/carers to continue to be involved in their medical care or the management of their condition. However, AMHS need to be willing to work more closely with parents/carers and to recognise that ongoing parental support for the young person is often necessary. This of course can only be achieved with the patient’s permission. It was stressed that transition should start much earlier than is generally recognised. There was a clear call for psychoeducational material for patients and families, as they described wanting to be informed much earlier in the process about their condition and what to expect as they move into adulthood. Specifically they asked for greater detail about their diagnosis and its prognosis, treatment options, involvement in decision making and a specified care plan. In addition, they wanted the opportunity to share with professionals their feelings and their desire to take control of managing their condition and play an active role in their care plan. It was emphasised that transition needs to be seen as a “process” rather than an “event”. Specifically, young people need to know what can be expected of these services, what AMHS offer and what they do not offer. They thought that transition meetings attended by CAMHS and AMHS were a good idea as this would facilitate a smoother transition. The potential value of lifespan clinics in the management of ADHD was highlighted. Patients were mindful however of the support they need in this process due to organisation and planning deficits. Support groups play an important role in the process, providing support for the young person and their family; the latter for the first time have to take a step back with a less proactive role. The experience described at the workshop strongly resonates with perspectives reported in the Transition Into Adult Mental Health Services (TRAMS) study [28]. This was a qualitative study of the experiences of 10 young people with ADHD, and their parents/carers, in transition from CAMHS to AMHS. Four key themes were identified from the analysis:Clinician Qualities and Relationship: These were identified to be an integral component of patient satisfaction, with a particular focus on the need for continuity of listening throughout the transition process, non-judgemental support and a practical, solution-focused approach targeting achievable and realistic goals. Responsibility of Care: A number of issues arose regarding the question of who is responsible for providing care and ensuring continuity of care. From the young person’s perspective, a conflict can emerge between the greater autonomy expected in AMHS and the difficulties they experience in areas such as planning and organisation, for which they typically rely on parent/carer support. Inevitably for parents/carers, there is a responsibility to balance support and autonomy. Ultimate responsibility for providing a consistent transition lay with clinicians, with the importance of joint meetings between CAMHS and AMHS, alongside patients and their families, as well as an overall clear plan for handover suggested as integral components of a seamless transition. Nature and Severity of Problems: Acceptance by AMHS appeared to be contingent on the nature and severity of the young person’s problems. There was a suggestion that transition may go more smoothly for those presenting with more severe problems, such as self-harming behaviours, and higher levels of comorbidity. For those whose problems were well controlled at the time of transition, neither the AMHS nor the young person themselves recognised the need for the on-going support required to sustain this level of health. Expectations of AMHS: The expectations of AMHS from patients and their families were often high, although many of these are unlikely to be met (i.e. assistance with housing). It is essential therefore that CAMHS clinicians do not provide unrealistic expectations of AMHS, with clear information offered prior to transition concerning exactly what can and cannot be offered. The transition process Most AMHS are set up to provide care for patients with serious mental disorders such as psychosis. In the TRACK study, only one in five of the 17 year olds in CAMHS with neurodevelopmental disorders (including ADHD) made a successful transition to adult services [27]. This contrasts with the more successful transition of young people with psychosis and other conditions considered to be serious mental disorders. Moreover, experience suggested that patients with ADHD and comorbid mental health problems, such as depression and suicidal ideation, often have an easier transition to adult services than patients diagnosed with ADHD alone. Nevertheless, one-half of all patients fail to adhere to treatment guidelines [9] and discontinue treatment within 2–3 years of starting pharmacological therapy [11, 34]. Some common reasons for the discontinuation of treatment are: adverse effects, ineffectiveness/suboptimal response, poor adherence, patient/caregiver decision and symptom relief. Other less common, although important, reasons include a dislike of medication, dosing inconvenience and social stigma [10]. It needs to be emphasised that transition to adult care is a process, rather than an event. For their part, CAMHS should provide timely and accurate information to patients about adult services and develop specific mechanisms to overlap the transitional arrangements with adult services. The transition process can be used to prompt a review of what other organisations may be required in the patient’s care. Indeed, this is an opportunity to inform appropriate agencies and organisations, with the patient’s consent, of their current circumstances. For this to be achieved, a good working knowledge of the common partner organisations in the area is crucial. These may be statutory and/or local organisations, such as child protection services, the local authority (for assessment for social care), local educational facilities and resources, court diversion services, and youth offending teams. In addition, there are an increasing number of voluntary and third-sector agencies that provide help and support. It is also advisable to involve the patient’s general practitioner (GP) because often this is the most stable service provider during the period of transition. It was proposed that for the transition to adult care to work well, adult psychiatry needs to accept referrals from CAMHS and community paediatric services, and develop a better understanding of their needs. This will involve targeted continuing professional education on ADHD and, in particular, on the transition process. Of concern is that there is no specific provision to ensure ongoing care for ADHD patients who are not accepted by AMHS. There followed discussions regarding the best model for service provision and, in particular, whether adult-focused neurodevelopmental services need to be specifically developed. These already exist in some parts of the UK and were considered useful (see Table 1). To avoid excessive resource implications and an unnecessary division to the process, these are best located as an extension of CAMHS with a set agenda that includes psychoeducation, goal setting, risk reduction, coping strategies, employment and education. Additional consultations can be added on an ‘as needed’ basis, e.g. with a pharmacist, social worker, psychiatrist, psychologist and/or a member of the youth offending team.Table 1 Transition practice in Leicestershire Leicestershire have developed a robust transition protocol ensuring that all referrals from children’s services are carefully reviewed and appropriate action is taken. In Leicestershire, referral acceptance by the adult ADHD service is 100 %. Children services are advised to start the referral well in advance of the point of transition on the patient’s 18th birthday. They are also advised not to discharge the patient before they have been seen by the adult service. This ensures that the patient is never lost between the two services. In case the patient does not attend the appointment arranged by the adult ADHD clinic, the referring clinician is contacted for further advice and efforts are made to encourage the patient to attend. When dealing with complex cases or difficult to engage patients, a period of joint working has also been found to be useful. Recognising the vulnerability of patients through this period of transition and the need to ensure continuity of care and stability, a great degree of flexibility is exercised when dealing with this group of patients. Even when the patient is not keen on taking the medication or attending further reviews, they are not discharged from the service. Time is spent with patients providing psychoeducation about the disorder, its course and the consequences of untreated ADHD. During these discussions it is important to acknowledge the patient’s views and reservations, for example, stigma about a psychiatric diagnosis label, their wish for autonomy, their view that they can manage their symptoms without the help of medication etc. They are offered a further appointment and are encouraged to take their time before making a final decision. In the large majority of cases, patients decide in favour of continuity of care. ADHD, education and employment We can be fairly confident that key transitions, and in particular, those bridging secondary education and the world of further learning, independent living and work, constitute a period of heightened vulnerability in the life of any young person with ADHD. However, the absence of a reliable evidence-base to inform generalisations about the employment experience of young people transitioning from school or college to the world of work is complicated by the heterogeneity of the ADHD population in terms of cognitive ability, comorbidity, gender expression, and treatment uptake, as well as geographical variations in the labour market. Whilst it is fair to suggest that young people with ADHD suit better those occupations that provide novelty, physical activity and immediate feedback, and which are thus non-repetitive, highly structured and lacking in organisational/time-management demands, the prospects for any individual are likely to hinge on a number of other factors which may be either directly or indirectly related to the way that ADHD has manifested itself during their formative years. Sufficient anecdotal evidence exists to complement studies in USA and Europe to suggest that rates of economic activity and thus income levels are significantly lower for young adults with ADHD, and that such differentials often persist throughout the life span [4, 6, 16]. The treatment effect is poor for occupational outcomes [2], however there is an emerging body of data to support the positive impact of personalised assistance in negotiating the labour market by helping job-seekers with the application and interview process, providing structures in daily living to manage financial obligations and punctuality, and in averting negative tendencies towards substance misuse, social conflict and offending [14, 23]. Likewise, adjustments in the workplace, such as a re-framing of organisation demands or support with time-management, may allow workers with deficits in these areas to exploit relative strengths in others and thus provide a net gain to the employer [1]. Programmes to improve the employability of people with ADHD are likely to be most effective when they are sensitive to the current needs and future potential of each individual, multi-modal in delivery, and which address gaps in prior learning that impair recruitment prospects. Consensus Having considered all the available evidence and expert opinion, the expert group deliberated the recommendations that need to be made to achieve a smooth and effective transition from child to adult services for young people with ADHD, and prevent drop out from services. The earlier recommendations of Young et al. [33] were used as a starting point and framework for discussion. When using these guidelines and recommendations, local situations will need to be considered when designing transition services as there is a considerable variation in service design and delivery between regions. Consensus was reached for general recommendations (Table 2) and specific recommendations (Table 3).Table 2 General recommendations for transition of care from children’s to adult services 1. Clear transition protocols should be developed jointly by commissioners, CAMHS/paediatric services, AMHS, primary care, and other agencies as relevant to facilitate transition and ensure that standards of care are maintained during the transition period. 2. These protocols should specify timeframes, lines of responsibility, who should be involved, how the young person should be prepared, and what should happen if AMHS are not able to accept the referral. 3. Protocols should allow for flexibility in the age of transition so as to accommodate developmental needs and stages, but there should be explicit referral criteria and service provision. Ideally, transition should occur at a time of clinical stability. Patients should not have to relapse or have worsening mental health in order to continue to be able to access services. 4. Transition protocols should be available to all clinical teams and should include psychoeducational material that provides high-quality, comprehensive, impartial and appropriately written information for both young people and their parents and carers. There is a need for more age-appropriate psychoeducational material for patients at the transition stage. This material should include information about ways that young people can manage their own symptoms and problems, and access advice and support. Information should also be developed in a media format that is readily accessed by young people, e.g. use of phone applications and internet sites. 5. The needs and wishes of parents/carers should be respected and their ongoing involvement with the young person negotiated. Some parallel services that can provide information and support for parents/carers during the transition period may be required. 6. Efforts should be made to inform and educate allied professionals who may come into contact with young people with ADHD for the first time during the transition period, e.g. forensic medical examiners and those working in the probation services and in correctional units and prisons. 7. Healthcare jurisdictions should be encouraged to use similar care pathways and outcome measures across different patient age groups. Table 3 Specific recommendations for transition of care from children’s to adult services 1. A planned transfer to an adult service should be made if the young person continues to have significant symptoms of ADHD or other co-existing conditions that require treatment 2. Transition should be planned well in advance by both referring and receiving services. Timings of transition may vary but should ordinarily be completed by the age of 18 years. Transition between teams should be a gradual process and should be thought of as a ‘process’ and not a ‘single event’. 3. Patients should be involved in discussion about transition and informed of the outcome of any transition assessment. The transition process should proceed according to need in terms of future medical care (e.g. involvement of general practitioner [GP] services, specialist adult ADHD teams, adult learning disability services, adult physical health teams). Importantly, the GP should be involved throughout the process. 4. Discussion, and where necessary, joint meetings between child and adult services must ensure that the needs of the young person will be appropriately met. It is important to consider the presence of comorbid and/or related problems, which may involve further discussion and collaboration with educational, or occupational and social agencies. 5. CAMHS practitioners and paediatricians should foster engagement with AMHS through open discussion and psychoeducation about ADHD, the benefit of evidenced-based psychological and pharmacological treatment where appropriate, and the risks of disengagement. It is important to address concerns about stigma associated with referral to AMHS. 6. For young people aged 16 years or over in CAMHS, a CPA should be used to aid transfer. CPAs are not available in paediatric practice, and so a planned assessment of need with the young person and their parents/carers and a clearly documented plan of action is recommended. 7. Parents/carers need to be prepared and facilitated to aid their child’s gradual move towards independence and autonomy (with respect to the management and treatment of their ADHD). The referring and receiving healthcare teams should be mindful of possible parental ADHD and when this is present (or suspected) provide appropriate support. 8. Shared care arrangements between primary and secondary care services for the prescription and monitoring of ADHD medications should be continued into adulthood. 9. Direct psychological treatment should be considered (individual and/or group Cognitive Behavioural Therapy) to support young people during key transitional stages. This should have a skills development focus and target a range of areas including ADHD symptoms, social skills, interpersonal relationship problems (with peers and family), problem solving, self-control, dealing with and expressing feelings. Active learning strategies should be used. 10. Specific protocols need to be developed for young people who are not accepted by AMHS criteria, but whom the referring service strongly believe need ongoing support. Care needs to be taken that these patients are not left without the support they need during this very important transition period. 11. Separate care pathways should be developed for young people who drop out of CAMHS or paediatric services when they are under 18 years of age, and who later re-present in the healthcare system as adults. 12. Separate care pathways should be developed for patients who come to the attention of the healthcare system on account of ADHD for the first time as adults. 13. The referral letter from children’s services should provide a comprehensive account of the patient, including: diagnostic summary and formulation; treatment history; rationale and response; side effects, compliance, abuse and diversion issues, and ongoing treatment needs; any psychiatric and medical comorbidities, their impact on ADHD and treatment; any other ongoing needs - social, financial, accommodation or occupational and an updated risk assessment. 14. The adult service should acknowledge the receipt of the referral. The patient should not be discharged by the children’s services until they have been seen by the adult services and their care has formally been taken over by the adult services. This provides a safety net and reduces the likelihood of patients dropping out of the services during the transition period. 15. Following acceptance of the referral, the adult service should allocate a key worker/lead clinician who will coordinate the care needed. 16. When dealing with patients who are anxious about the transfer of care to adult services or those with complex needs, it may be necessary for children’s services to joint work with adult services for a few months to facilitate the transfer of care. Future research In the course of the workshop it became clear that little is known about the outcomes of individuals with ADHD who fall through care gaps, patients who remain under the care of Primary Care services throughout transition into adulthood, interventions that might improve the process, and patient and carer experiences. Thus recommendations were also made for future research into areas relevant to the transition of patients with ADHD from child to adult services, including:The reasons why patients stop taking medication during the transition period. The reasons why patients lose contact with their ADHD healthcare system during the transition period, including why adolescents chose to discontinue treatment and how to address issues of decreased efficacy of medication over time. Pharmacological strategies to identify patients, before their transition to adult services, who are at risk of developing comorbid psychiatric difficulties, with a view to preventing the development of these comorbid problems. The effectiveness of non-pharmacological interventions in this age group. Females with ADHD are underrepresented in the research literature, and their position through the transition process demands more attention. Conclusions Transition is a process, not an event, and normally occurs by the age of 18. However, flexibility is required to accommodate individual needs. Unfortunately, transition is often poorly planned, executed and experienced. Adherence to clear recommendations is necessary to ensure effective transition and prevent drop-out from services. Acknowledgements The authors wish to acknowledge the following individuals who attended the consensus meeting: Blanca Bolea, Helen Crimlisk, Helen Duncan, Susan Dunn Morua, Val Harpin, Claude Jousselin, Geoff Kewley, Angela Koni Moore, Clodagh Murphy, Mavis Nyakunengwa, Shorayi Nyamupanda, Fintan O'Regan, Sri Perecherla, Prem Shah. In addition we acknowledge the assistance of Ben Greer for the preparation of the manuscript. Funding No funding was obtained for this study. Availability of data and materials Not Applicable – not datasets were generated or analysed in the course of the expert workshop. Authors’ contributions All authors contributed to the planning and scientific input of this consensus statement. Early drafts of the manuscript were completed by SY, DC and MAr. All authors read and approved the final manuscript. Authors’ information Susan Young obtained a doctorate in clinical psychology from University College London in 1999 and a PhD in psychology from King’s College London in 1999. She is a Senior Lecturer in Forensic Clinical Psychology at Imperial College London, and Director of Forensic Research & Development at West London Mental Health NHS Trust. Previously, she was employed as a Clinical Neuropsychologist at the Maudsley Hospital, where she set up and developed the neuropsychology service at the first adult ADHD service in the United Kingdom, and has extensive clinical experience in the assessment and treatment of youths and adults with ADHD and in the assessment and treatment of offenders with mental illness and/or mental disorder. Her main research interests are in the areas of neurodevelopmental disorder, ADHD, and forensic psychology. Marios Adamou is a Consultant Psychiatrist in Neurodevelopmental Psychiatry (ADHD and ASD) at South West Yorkshire Partnership NHS Foundation Trust, and Professor at Huddersfield University, UK. His medical training was completed at the Aristotle University, Greece, and his psychiatric training at Guy’s, King’s, and St Thomas’s Hospitals, London. He holds numerous positions including fellow of the Higher Education Academy (FHEA), the Royal Society of Arts (FRSA), the Royal Society of Public Health (FRSPH) and the Society of Biology (FSB). Philip Asherson is Professor of Molecular Psychiatry at the MRC Social, Genetic and Developmental Psychiatry centre at the Institute of Psychiatry, Psychology and Neuroscience, King’s College London. He is also a Consultant Psychiatrist at the Maudsley Hospital. He earned his medical degree from The Royal London Hospital and his doctoral degree from the University of Wales. His current research interests focus upon clinical, quantitative genetic and molecular genetic studies of ADHD across the lifespan, investigation of mood instability and emotional regulation in ADHD, and treatment of ADHD in young offenders. David Coghill leads the Developmental Research Group at the University of Dundee. His group's research interests are focused on ADHD and other developmental disorders and include studies into the neuropsychology of ADHD and conduct disorder, neuropsychopharmacoloogy, genetics, pharmacogenetics, clinical trials, quality of life, and translational studies investigating the best ways to integrate scientific knowledge into clinical practice. Bill Colley is a writer, teacher trainer and educational consultant who spent much of his career in the independent sector before becoming the headmaster of a residential special school. He is currently working in local government as a Service Manager with responsibility for additional support needs and contributes at a local level to an ADHD support group, the NHS Autism Assessment Pathway, and to the National Autism Strategy Reference Group (Scotland). He is an advisor to Mindroom, a Scottish charity for people with learning difficulties and has a particular interest in improving educational outcomes for children who experience developmental disorders, and in supporting their families via multi-professional collaboration. His latest task is to establish a new provision, and an authority-wide service, for children who have disengaged from mainstream education. Gisli Gudjonsson CBE, obtained a PhD from the University of Surrey, England, in 1981. He is Emeritus Professor of Forensic Psychology at the Institute of Psychiatry, Psychology & Neuroscience King’s College London. His main research interests are in the areas of forensic psychology, including police interrogation, false confessions, and motivation for offending. In 2009 the British Psychological Society granted him a Lifetime Achievement Award for his exceptional and sustained contribution to the practice of psychology. He was appointed a Commander of the Order of the British Empire (CBE) in the Queen’s Birthday 2011 Honours List for services to clinical psychology. Chris Hollis is Professor of Child & Adolescent Psychiatry at the University of Nottingham and Consultant in Developmental Neuropsychiatry with Nottinghamshire Healthcare NHS Trust. Before moving to Nottingham, Chris trained in psychiatry at the Maudsley Hospital and Institute of Psychiatry, where he completed his PhD. His research interests include ADHD, Tourette syndrome, early onset schizophrenia and the development and implementation of digital technologies to enhance assessment and monitoring of mental health disorders. Over the last 5 years Chris has a research income of over £2.5 million. Chris was a member of the NICE Guideline Development Group (GDG) for ADHD (2005–8), NICE ADHD Quality Standard Advisory Committee (2012–13) and chaired the NICE GDG for schizophrenia and psychosis in children and young people (2011–13). Jane McCarthy is a Consultant Psychiatrist in Intellectual Disability with the East London NHS Foundation Trust and is a Visiting Senior Lecturer at King’s College London, England. She has worked for over 20 years as a senior psychiatrist in specialist services for adolescents and adults with intellectual disabilities and autism. Dr. McCarthy’s key research interests include the outcomes of psychiatric disorders in people with neurodevelopmental disorders, and in 2012 she was elected Chair of the International Association for Scientific Study of Intellectual and Developmental Disabilities (IASSIDD) Challenging Behavior & Mental Health Special Interest Research Group, which is the largest global research network in this field of research. She is also Vice Chair of the Psychiatry of Intellectual Disability Faculty, Royal College of Psychiatrists, UK. Ulrich Müller is a Senior Lecturer in the Department of Psychiatry, University of Cambridge, UK, and Honorary Consultant Psychiatrist with Cambridge University Hospitals NHS Foundation Trust and Cambridgeshire & Peterborough Foundation NHS Trust (CPFT). Previously he was trained in psychiatry, psychotherapy, neurological rehabilitation and cognitive neuroscience in Germany. His research focuses on adult ADHD and cognitive enhancing medication, with recent studies clarifying how drugs such as atomoxetine and methylphenidate modulate cognitive functions and related brain processes. Moli Paul is a principal teaching fellow at the University of Warwick and an honorary consultant child and adolescent psychiatrist with Coventry and Warwickshire Partnership Trust, UK. Her main research interests are health care decision-making by children, young people and their families; transitions between adolescent and adult health services; and empirical research on consent and rights in biomedical ethics and health care law. Mark Pitts is the Clinical Nurse Specialist at the Adult ADHD Service, an outpatient clinic based at the Maudsley Hospital, and one of a number of national specialist services offered by the South London & Maudsley NHS Foundation Trust. He has degrees in Psychology and Learning Disability Nursing, and worked for five years in a national low-secure service for people with mild/borderline learning disability at the Bethlem Royal Hospital prior to taking up his current position in 2005. His clinical interest are in the assessment of ADHHD in adults, pharmacological treatment and follow-up, and the development of adult ADHD services. Muhammad Arif trained in Psychiatry in Pakistan, where he gained experience in the assessment and treatment of children with ADHD. He is currently acting Consultant Psychiatrist working in Leicester, UK, where he formed the special interest Adult ADHD clinic in 2002. In January 2009 he was responsible for the commission of the Adult ADHD service in Leicester. Other interest include Trauma Psychiatry, CBT and EMDR. Competing interests SY has received honoraria for consultancy, sponsorship for attendance at scientific meetings, educational talks and /or research awards from Janssen, Lilly, Novartis, HB Pharma, Flynn Pharma and/or Shire. King’s College London departmental research support account for PA received honoraria for consultancy, educational talks and/or research awards from Shire, Lilly, Novartis, Vifor Pharma, GW Pharma, Janssen and/or QbTech. DC has received honoraria for consultancy, educational talks and/or research awards from Novartis, Sandoz, Vifor, Shire, Lilly and/or Janssen. GG has received consultancy fees, speaker fees and/or travel honoraria from Lilly, Janssen and Shire. The University of Nottingham has received research funding for CH from Shire for non-pharmacologicial ADHD research. UM has received honoraria for consultancy, educational talks and sponsorship for attendance at scientific meetings from Lilly, Flynn Pharma/Medice, Janssen and/or Shire. MPi has received honoraria for consultancy and educational talks from Shire. MA has received honoraria for consultancy, educational talks and sponsorship for attendance at scientific meetings from Lilly, Janssen Cilag, Otsuka, Servier, Flynn-Pharma. Other authors have no competing interests to declare. Consent for publication Not applicable. Ethics and consent to participate Not applicable. ==== Refs References 1. Adamou M Arif M Asherson P Aw TC Bolea B Coghill D Young S Occupational issues of adults with ADHD BMC psychiatry 2013 13 1 59 10.1186/1471-244X-13-59 23414364 2. Arnold LE, Hodgkins P, Caci H, Kahle J, Young S. Effect of treatment modality on long-term outcomes in attention-deficit/hyperactivity disorder: a systematic review. PloS one. 2015;10(2):e0116407. 3. Belling R McLaren S Paul M Ford T Kramer T Weaver T Singh SP The effect of organisational resources and eligibility issues on transition from child and adolescent to adult mental health services J Health Serv Res Policy 2014 19 3 169 176 10.1177/1355819614527439 24700210 4. Biederman J Faraone SV The effects of attention-deficit/hyperactivitydisorder on employment and household income Medscape General Medicine 2006 8 3 12 17406154 5. Cheung CH Rijdijk F McLoughlin G Faraone SV Asherson P Kuntsi J Childhood predictors of adolescent and young adult outcome in ADHD J Psychiatr Res 2015 62 92 100 10.1016/j.jpsychires.2015.01.011 25680235 6. Daley D Jacobsen RH Lange AM Sorensen A Walldorf J Private and social costs of ADHD: Cost analysis 2014 Copenhagen, Denmark University Press of Southern Denmark 7. Erikson EH Childhood and Society 1963 2 New York Norton 8. Faraone SV Biederman J Mick E The age-dependent decline of attention deficit hyperactivity disorder: a meta-analysis of follow-up studies Psychol Med 2006 36 2 159 165 10.1017/S003329170500471X 16420712 9. Ferrin M Taylor E Child and caregiver issues in the treatment of attention deficit-hyperactivity disorder: education, adherence and treatment choice Future Neurol 2011 6 3 399 413 10.2217/fnl.11.5 10. Gajria K Lu M Sikirica V Greven P Zhong Y Qin P Xie J Adherence, persistence, and medication discontinuation in patients with attention-deficit/hyperactivity disorder–a systematic literature review Neuropsychiatr Dis Treat 2014 10 1543 25187718 11. Garbe E Mikolajczyk RT Banaschewski T Petermann U Petermann F Kraut AA Langner I Drug treatment patterns of attention-deficit/hyperactivity disorder in children and adolescents in Germany: results from a large population-based cohort study J. Child Adolesc. Psychopharmacol. 2012 22 6 452 458 10.1089/cap.2012.0022 23234588 12. Hall CL Newell K Taylor J Sayal K Hollis C Services for young people with attention deficit/hyperactivity disorder transitioning from child to adult mental health services: A national survey of mental health trusts in England J Psychopharmacol 2015 29 1 39 42 10.1177/0269881114550353 25237121 13. Hall CL Newell K Taylor J Sayal K Swift KD Hollis C ‘Mind the gap-mapping services for young people with ADHD transitioning from child to adult mental health services BMC psychiatry 2013 13 1 186 10.1186/1471-244X-13-186 23842080 14. Kubik J Efficacy of ADHD coaching for adults with ADHD J Atten Disord 2010 13 5 442 453 10.1177/1087054708329960 19276311 15. Lamb, C., Hall, D., Kelvin, R., Van Beinum, M. (2008). Working at the CAMHS/Adult Interface: Good practice guidance for the provision of psychiatric services to adolescents/young adults. Retrieved from http://www.rcpsych.ac.uk/pdf/Transition_2008.pdf 16. Le HH Hodgkins P Postma MJ Kahle J Sikirica V Setyawan J Doshi JA Economic impact of childhood/adolescent ADHD in a European setting: the Netherlands as a reference case Eur. Child Adolesc. Psychiatry 2014 23 7 587 598 10.1007/s00787-013-0477-8 24166532 17. McCarthy S Asherson P Coghill D Hollis C Murray M Potts L Wong IC Attention-deficit hyperactivity disorder: treatment discontinuation in adolescents and young adults Br J Psychiatry 2009 194 3 273 277 10.1192/bjp.bp.107.045245 19252159 18. McCarthy S Wilton L Murray ML Hodgkins P Asherson P Wong IC The epidemiology of pharmacologically treated attention deficit hyperactivity disorder (ADHD) in children, adolescents and adults in UK primary care BMC Pediatr 2012 12 1 78 10.1186/1471-2431-12-78 22712630 19. National Institute for Health and Care Excellence Attention deficit hyperactivity disorder: Diagnosis and management of ADHD in children, young people and adults 2008 London NICE 20. National Institute for Health and Care Excellence Methylphenidate, atomoxetine and dexamfetamine for the treatment of attention deficit hyperactivity disorder (ADHD) in children and adolescents 2006 London NICE 21. Nutt DJ Fone K Asherson P Bramble D Hill P Matthews K Young S Evidence-based guidelines for management of attention-deficit/hyperactivity disorder in adolescents in transition to adult services and in adults: recommendations from the British Association for Psychopharmacology J Psychopharmacol 2007 21 1 10 41 10.1177/0269881106073219 17092962 22. Piaget J Piaget’s theory 1976 Berlin Springer 23. Ratey N Goldstein S Ellison AT Life coaching for adult ADHD Clinician's guide to adult ADHD: Assessment and Intervention 2002 London Academic 261 279 24. Royal College of Paediatrics and Child Health. (2003). Bridging the gaps: Health care for adolescents. The Royal College of Psychiatrists. Council report: CR114. 25. Singh SP Transition of care from child to adult mental health services: the great divide Curr. Opin. Psychiatry 2009 22 4 386 390 10.1097/YCO.0b013e32832c9221 19417667 26. Singh SP Evans N Sireling L Stuart H Mind the gap: the interface between child and adult mental health services Psychiatrist 2005 29 8 292 294 27. Singh SP Paul M Ford T Kramer T Weaver T McLaren S White S Process, outcome and experience of transition from child to adult mental healthcare: multiperspective study Br J Psychiatry 2010 197 4 305 312 10.1192/bjp.bp.109.075135 20884954 28. Swift KD Hall CL Marimuttu V Redstone L Sayal K Hollis C Transition to adult mental health services for young people with Attention Deficit/Hyperactivity Disorder (ADHD): a qualitative analysis of their experiences BMC psychiatry 2013 13 1 74 10.1186/1471-244X-13-74 23497082 29. UK Adult ADHD Network. (2012). Transition of ADHD from adolescence to adulthood. Retrieved from: //www.ukaan.org/ADHD-Adolescence-Adulthood/. Accessed 23 Aug 2016. 30. Young S. ADHD in Adolescence. 2014; Retrieved from http://www.psychology-services.uk.com/adhd-in-adolescence.htm. Accessed 23 Aug 2016. 31. Young S Gudjonsson G Growing out of Attention-Deficit/Hyperactivity Disorder: the relationship between functioning and symptoms J Atten Disord 2008 12 2 162 169 10.1177/1087054707299598 17494827 32. Young S Moss D Sedgwick O Fridman M Hodgkins P A meta-analysis of the prevalence of attention deficit hyperactivity disorder in incarcerated populations Psychol Med 2015 45 2 247 258 10.1017/S0033291714000762 25066071 33. Young S Murphy CM Coghill D Avoiding the ‘twilight zone’: recommendations for the transition of services from adolescence to adulthood for young people with ADHD BMC psychiatry 2011 11 1 174 10.1186/1471-244X-11-174 22051192 34. Zetterqvist J Asherson P Halldner L Långström N Larsson H Stimulant and non‐stimulant attention deficit/hyperactivity disorder drug use: total population study of trends and discontinuation patterns 2006–2009 Acta Psychiatr Scand 2013 128 1 70 77 10.1111/acps.12004 22943458

PMC005xxxxxx/PMC5000408.txt

==== Front BMC NeurolBMC NeurolBMC Neurology1471-2377BioMed Central London 67810.1186/s12883-016-0678-0Research ArticleAdjunctive dexamethasone therapy in unconfirmed bacterial meningitis in resource limited settings: is it a risk worth taking? Gudina Esayas Kebede esakgd@gmail.com 12Tesfaye Markos tesmarkos@yahoo.com 23Adane Aynishet ayne.2003@yahoo.com 4Lemma Kinfe kinfewudie@gmail.com 5Shibiru Tamiru drtamshib1@gmail.com 6Wieser Andreas wieser@mvp.uni-muenchen.de 789Pfister Hans-Walter hans-walter.pfister@med.uni-muenchen.de 10Klein Matthias matthias.klein@med.uni-muenchen.de 101 Department of Internal Medicine, Jimma University, Jimma, Ethiopia 2 Centre for International Health, Ludwig-Maximilians-University, Munich, Germany 3 Department of Psychiatry, Jimma University, Jimma, Ethiopia 4 Department of Internal Medicine, University of Gondar, Gondar, Ethiopia 5 Department of Internal Medicine, Hawassa University, Hawassa, Ethiopia 6 Department of Internal Medicine, Arba Minch Hospital, Arba Minch, Ethiopia 7 Division of Infectious Diseases and Tropical Medicine, Medical Center of Ludwig-Maximilians-University, Munich, Germany 8 Department of Bacteriology, Max von Pettenkofer Institute (LMU), Munich, Germany 9 German Center for Infection Research (DZIF), Partner Site Munich, Munich, Germany 10 Department of Neurology, Ludwig-Maximilians-University, Munich, Germany 26 8 2016 26 8 2016 2016 16 1 15328 5 2016 18 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Bacterial meningitis is associated with significant morbidity and mortality despite advances in medical care. The main objective of this study was to assess the association of adjunctive dexamethasone treatment with discharge outcome of patients treated as bacterial meningitis in low income setting. Methods A retrospective study was conducted at four teaching hospitals across Ethiopia. Patients of age 14 years and older treated as cases of bacterial meningitis between January 1, 2011 and April 30, 2015 were included in this study. Information regarding sociodemographic data, clinical presentations, laboratory data, treatments given and status at hospital discharge were retrieved from patients’ medical records using a structured questionnaire. Predefined outcome variables at discharge were analysed using descriptive statistics. Multivariable logistic regression was used to identify factors independently associated with poor outcome. Results A total of 425 patients treated with the presumptive clinical diagnosis of bacterial meningitis were included in this study (lumbar puncture done in 56 %; only 19 % had CSF findings compatible with bacterial meningitis, and only 3 % had proven etiology). The overall in hospital mortality rate was 20.2 %. Impaired consciousness, aspiration pneumonia, and cranial nerve palsy at admission were independently associated with increased mortality. Adjuvant dexamethasone, which was used in 50.4 % of patients, was associated with increased in-hospital mortality (AOR = 3.38; 95 % CI 1.87–6.12, p < 0.001) and low Glasgow outcome scale (GOS) at discharge (AOR = 4.46 (95 % CI 1.98–10.08). This association between dexamethasone and unfavorable outcome was found to be more pronounced in suspected but unproven cases and in those without CSF alterations compatible with bacterial meningitis. Conclusion Most patients treated for suspected bacterial meningitis did not receive proper diagnostic workup. Adjuvant dexamethasone use in clinically suspected but unproven cases of bacterial meningitis was associated with an increased mortality and poor discharge GOS. These findings show that there are potential deleterious effects in unconfirmed cases in this setting. Physicians practising under such circumstances should thus abide with the current recommendation and defer the use of adjuvant corticosteroid in suspected cases of bacterial meningitis. Electronic supplementary material The online version of this article (doi:10.1186/s12883-016-0678-0) contains supplementary material, which is available to authorized users. Keywords Bacterial meningitisOutcomeDexamethasoneEthiopiaEast-Africahttp://dx.doi.org/10.13039/501100003042Else Kröner-Fresenius-Stiftung2013_HA77Klein Matthias issue-copyright-statement© The Author(s) 2016 ==== Body Background Bacterial meningitis is a serious infection of the central nervous system that can progress rapidly and result in death or permanent debilitation [1]. It is associated with a high fatality rate despite advances in medical care [2] and a significant proportion of survivors suffer from long term neurologic sequelae [3]. Most of the cases of acute bacterial meningitis (ABM) occur in low income countries [4] where case fatality rates are higher than in countries with a high standard of medical care [5, 6]. The duration of disease [7], age [8], and immune status of the patient [9–11], timing of antibiotic initiation [12, 13], and type of microorganism [14, 15] were found to be important factors in determining the outcome of ABM. Significant controllable factors known to improve survival and neurologic recovery are rapid diagnosis and an early treatment [12, 16], both of which are difficult to achieve when laboratory support and treatment options are limited [4]. Corticosteroid as adjunctive treatment of ABM is one of the most thoroughly studied and widely discussed controversial issues in recent years [17–21]. Yet, its benefit in mortality and morbidity reduction is far from being settled [20–23]. The existing evidences indicate that the efficacy of dexamethasone varies with etiologic agents [24, 25], clinical circumstances and regions of the world [19, 24–27]. The current adult recommendations limit its use to pneumococcal meningitis in high income countries [24, 25]. Furthermore, the few studies from the developing world did not find any benefits of corticosteroid on mortality and neurologic sequelae [28–30]. We recently analysed the characteristics of 425 patients who were treated and discharged with the presumed diagnosis of bacterial meningitis. One of the main findings was that CSF analysis was done in only 56 % of these patients and the diagnosis could be proven by detection of a causative pathogen in as little as 3.3 % of patients [31]. Now we aimed to assess treatment outcomes and factors associated with poor outcome in these patients. We especially aimed to investigate the effect of adjunctive dexamethasone treatment on the outcome of patients treated for suspected ABM in the four study centres in Ethiopia. Methods Settings Ethiopia is a country located in East Africa with an estimated population of 87,952, 000 as of July 2014 [32]. This study was conducted at four teaching hospitals in Ethiopia – Jimma University Specialized Hospital, Hawassa University Referral Teaching Hospital, University of Gondar Hospital and Arba Minch Hospital. The first three are full-fledged university hospitals serving as referral hospitals. Arba Minch hospital is a general hospital affiliated with Arba Minch University’s medical school. All of these hospitals are located in the eastern end of meningitis belt of Africa in different regions of Ethiopia– Northwest (Gondar), southwest (Jimma), south (Arba Minch) and southcentral (Hawassa). The overall catchment population of the four hospitals is nearly 25 million – over a quarter of the Ethiopian population. Design A hospital based retrospective cohort study was conducted using medical record review of patients treated as cases of bacterial meningitis at the four hospitals during the period of 1 January, 2011 to 30 April, 2015. Clinical characteristics of the patients were recently published [31]. Study participants Patients included in the study were those of age 14 years and older treated with a presumptive diagnosis of bacterial meningitis and who had complete medical records regarding issues related to diagnosis, treatment and outcome of ABM. Patients whose antibiotic treatment was discontinued before ward admission because of confirmed alternative diagnosis and those with incomplete clinical records were excluded. Cases treated as bacterial meningitis were categorized based on the 2003 World Health Organization case definition used in WHO-recommended surveillance standards for surveillance of selected vaccine-preventable diseases [33]. These categories are:Suspected unproven cases of bacterial meningitis – Cases with acute onset (≤7 days) of fever (axillary temperature of ≥38.0 °C) PLUS any of: neck stiffness and altered consciousness PLUS no other alternative diagnosis PLUS no or incomplete CSF analysis. Possible bacterial meningitis – Cases with clinical signs as described for “suspected unproven ABM” PLUS CSF examination showing at least one of the following three – (1) turbid appearance (2) pleocytosis (>100 white cells/mm3) (3) pleocytosis (10–100 white cells/mm3) AND either an elevated protein (>100 mg/dl) or decreased CSF to serum glucose ratio (<40 %). Confirmed (proven) bacterial meningitis – Cases with detected microorganisms by culture, gram stain or agglutination test from CSF specimen. Non-cases (bacterial meningitis less likely) – Cases not fulfilling any of the above criteria and/or those with evidences suggesting other diagnoses. Data collection procedure Patients treated as cases of ABM were identified using the data from inpatient registration books of medical wards at each hospital. Their medical records were then retrieved from the archives to be reviewed according to a structured questionnaire prepared for the study (Additional file 1). The information gathered included socio-demographic profiles, clinical conditions at presentation, type of antibiotic treatment, adjunctive dexamethasone treatment, clinical course in the hospital, and discharge conditions (death and neurologic sequelae). Glasgow Outcome Scale (GOS) was interpreted from the discharge note (see below). Definitions of outcome variables Focal neurologic deficit (FND) – refers to (1) unilateral extremity weakness [monoparesis or hemiparesis] (2) unilateral hypaesthesia (3) localized cranial nerve palsies (III, IV and VII). Glasgow Outcome Scale (GOS) – was the interpretation of treating physician’s documentation of the patient status at discharge. 1 = if death was documented; 2 = if patient was in ‘coma’ or ‘unresponsiveness’ at leaving hospital; 3 = if document included ‘some improvement’ and any of ‘hemiplegia’, ‘paraparesis’, or ‘major disability’; 4 = if document included ‘improved’ with minor sequelae such as ‘facial palsy’ or ‘decreased hearing capacity’; 5 = if document included ‘full recovery’ or ‘discharge with complete improvement’. Level of consciousness – was stated using Glasgow Coma Scale (GCS) which ranges from score of 3 to 15. Patients with score of 15 were considered as fully conscious; 9–14 as impaired consciousness and as coma for scores between 3 and 8. Data processing, analysis and interpretation The data was checked for completeness and consistency. It was then entered to EpiData version 3.1 and was later transferred to SPSS® (IBM Corporation) version 20 for analysis. Bivariable analysis was done to identify association between dependent and independent variables. All independent variables with p < 0.25 in bivariate analysis were entered for multivariable analysis. Forward logistic regression analysis was done to identify the best fit model. Independent predictors were analysed for three outcome variables – death, Glasgow outcome scale and neurologic sequelae at discharge from the hospital. P-values of <0.05 were used as level of statistical significance. Results Background clinical characteristics Complete medical records were available for 425 patients treated as bacterial meningitis. The main clinical characteristics of the patients have been previously described in detail [31]. Briefly, the mean age at presentation was 32 ± 15.7 years (range 14 to 85); 52.7 % of them were men. Only about 30 % (127 patients) presented within 2 days of symptom onset. Fever and headache were major presenting symptoms. On presentation, 213 (50.1 %) of patients had impaired consciousness and 33 (7.8 %) had focal neurologic deficits. HIV infection was detected in 23 (5.4 %) patients, however, only 349 (82.1 %) were tested for it. Forty-four (10.4 %) patients had additional diagnosis of pneumonia, all of which were attributed to aspiration. Lumbar puncture was done for 236 (55.5 %) of patients; 220 (93.2 %) of them had microscopic examination of Gram-stained CSF specimen. Leukocyte count was done for 180 (76.3 %) of these patients but only 58.1 % had analysis for both protein and glucose. Culture was done for only 61(25.8 %) of these cases. Only 14 (3.3 %) of them had a confirmed etiology and classified as “proven ABM”; 82 (19.3 %) belonged to the category of possible ABM. About 46 % (196 patients) were classified as suspected unproven cases. In all cases in this category, CSF was either only partially analysed or not collected at all. The rest 133 (31.3 %) did not fulfil clinical or laboratory based definitions of ABM [31]. Antimeningeal dose of intravenous ceftriaxone (4 g/day) alone or in combination with other antibiotics was used in all patients except for one, who was given a combination of benzyl penicillin and chloramphenicol. Intravenous metronidazole was given to 44 (10.4 %) patients for suspected aspiration pneumonia. Adjunctive dexamethasone treatment was given to 214 (50.4 %) patients [31]. Discharge outcome One hundred fifty-six (36.7 %) were discharged with unfavorable outcome (GOS = 1–4); 86 patients (20.2 %) died in the hospital. The median time from hospital admission to death was 3 days; 55.8 % of the deaths occurred in the first 4 days of admission. Of those who left the hospital alive (339), 277 (81.7 %) were discharged and 57 (16.8 %) left against medical advice (LAMA). Among surviving patients, 70 (20.6 %) had unfavourable GOS (2 to 4); 38 (11.2 %) had documented neurologic sequelae. The average length of hospital stay (LOS) for discharged patients was 11.0 days (SD = 6.6). Those who left against medical advice had a LOS of 6.1 days (SD = 3.7) and 52.6 % of them left in the first 4 days of admission (Table 1).Table 1 Outcome at leaving hospital in patients treated as bacterial meningitis at teaching hospitals in Ethiopia, 2011–2015 Characteristics All patients Adjuvant dexamethasone No dexamethasone Status at leaving hospital (N = 425), n (%)  Discharged 277 (65.2) 125 (58.4) 152 (72.0)  Left against medical advice 57 (13.4) 26 (12.1) 31 (14.7)  Referred 5 (1.2) 1 (0.5) 4 (1.9)  Died 86 (20.2) 62 (29.0) 24 (11.4) Discharge condition of survivors (N = 339), n (%)  Complete recovery 230 (67.9) 98 (64.5) 132 (70.6)  Some improvement 78 (23.0) 41 (27.0) 37 (19.8)  The same as admission 31 (9.1) 13 (8.6) 18 (9.6) GOS (N = 425), n (%)  1 86 (20.2) 62 (29.0) 24 (11.4)  2 22 (5.2) 9 (4.2) 13 (6.2)  3 9 (2.1) 5 (2.3) 4 (1.9)  4 39 (9.2) 26 (12.1) 13 (6.2)  5 269 (63.3) 112 (52.3) 157 (74.4) Neurologic sequelae (N = 38 a), n (%)  Low GCS 18 (47.4) 5 (27.8) 13 (59.1)  Hemiparesis 8 (21.1) 4 (22.2) 4 (18.2)  Seizure 9 (23.7) 6 (33.3) 3 (13.6)  Paraparesis 3 (7.9) 3 (16.7) 0  Cranial nerve palsy 2 (5.3) 0 2 (9.1) Length of hospital stay in days, mean (SD)  Total 8.9 (6.4) 8.8 (7.0) 9 (5.8)  Discharged patients 11.0 (6.6)  LAMA 6.1 (3.7)  Referred 7.4 (5.2)  Died 4.0 (3.4) GOS Glasgow outcome scale, LAMA left against medical advice a2 Patients had multiple complications Factors associated with poor outcome Glasgow outcome scale (GOS) – by dichotomizing GOS into favorable (GOS = 5) and unfavorable (GOS =1 to 4) outcome, we found that admission GCS (AOR = 0.77; 95 % CI = 0.66–0.89) and dexamethasone treatment (AOR = 4.46; 95 % CI 1.98–10.08) were independently associated with unfavorable outcome. Note that GCS had reverse association with poor outcome; every increment from lowest of 3 to 15 resulted in improvement of outcome by 23 %. Fifty-two (12.2 %) of patients were additionally treated with presumptive diagnosis of tuberculous meningitis (TBM). These groups of patients had unfavourable outcome at discharge as compared to other groups (AOR = 2.78; 95 % CI 1.06–7.30). In hospital death – Admission Glasgow coma scale, presence of pneumonia and cranial nerve palsy during hospitalization were patient related factors independently associated with increased mortality. Accordingly, every drop of GCS from 15 was associated with increment of mortality by 21 % (AOR = 0.79; 95 % CI = 0.73–0.85). On the other hand, adjunctive dexamethasone therapy was found to be associated with over 3 times increment of mortality (AOR = 3.38; 95 % CI = 1.87–6.12) (Table 2). However, no association was seen between increased mortality and other conventional risk factors such as duration of illness, age of the patient and HIV infection.Table 2 Factors independently associated with poor outcomes at leaving hospital in patients treated as bacterial meningitis at teaching hospitals in Ethiopia, 2011–2015 Variable AOR 95 % CI P-value Death  Level of consciousness at presentation (for a point increase in GCS) 0.79 0.73–0.85 <0.001  Dexamethasone treatment 3.38 1.87–6.12 <0.001  Aspiration pneumonia at presentation 2.97 1.36–6.41 0.006  Cranial nerve palsy at presentation 4.73 1.45–15.50 0.010 Unfavorable outcome (GOS = 1–4)  Level of consciousness at presentation (for a point increase in GCS) 0.77 0.66–0.89 <0.001  Dexamethasone treatment 4.46 1.98–10.08 <0.001  TB suspected cases 2.78 1.06–7.30 0.038 Neurologic sequelae  Focal neurologic deficit at presentation 3.33 1.31–8.50 0.012  Seizure at presentation 2.20 1.03–4.67 0.041  Duration of illness before presentation 1.09 1.01–1.16 0.020  Impaired consciousness (GCS < 15) 2.65 1.21–5.81 0.015 This table presents output of Forward logistic regression analysis Neurologic sequelae – Focal neurologic deficits (AOR = 3.33; 95 % CI 1.31–8.50), seizures (AOR = 2.20; 95 % CI 1.03–4.67) and a low level of consciousness (AOR = 2.65; 95 % 1.21–5.81) at admission were associated with the occurrence of neurologic sequelae at discharge. This analysis showed also that a delay of one day from symptom onset to hospital presentation was associated with 9 % increment in risk of neurologic sequelae (AOR = 1.09; 95 % CI 1.01–1.16) (Table 2). As described above, 15 % of patients left hospital against medical advice or referred for better care. Separate analysis was done to assess if these patients differed clinically from discharged patients. Accordingly, they were found to have lower GOS, lower GCS and higher proportion of neurologic sequelae when leaving the hospital (Table 3).Table 3 Difference in secondary outcome variables between discharged patients and those who left the hospital against medical advice or were referred to other centres, in patients treated as bacterial meningitis at teaching hospitals in Ethiopia, 2011–2015 Discharge outcome GOS, N (%) Neurologic sequelae, N (%) Impaired GCS, N (%) 2 3 4 5 P Yes No P Yes No P LAMA/referred 21 (33.9) 6 (9.7) 20 (32.3) 15 (24.2) <0.001 16 (25.8) 46 (74.2) <0.001 13 (21.0) 49 (79.0) <0.001 Discharged 1 (0.4) 3 (1.1) 19 (6.9) 254 (91.7) 18 (6.5) 259 (93.5) 2 (0.7) 275 (99.3) Dexamethasone treatment and its association with discharge outcomes Chi-square test showed that dexamethasone was used more often in confirmed and probable cases of bacterial meningitis, those with turbid CSF and organism detected by Gram staining. On the other hand, it was found to be prescribed less often in HIV-positive patients as compared to non-HIV cases. There was also a clear difference in the pattern of dexamethasone treatment between hospitals ranging from 23.5 % at Hawassa to 73.9 % at Gondar (Table 4).Table 4 Comparison of background characteristics by dexamethasone treatment of patients treated as bacterial meningitis in Ethiopia, 2011–2015 Characteristics Dexamethasone No dexamethasone P value Mean age, year (SD) 30.1 (15.0) 33.9 (16.2) 0.116 Duration of illness, days (SD) 5.2 (4.3) 5.0 (4.3) 0.683 Diagnosis of meningitis  Confirmed 12 (85.7) 2 (14.3) 0.001*  Probable 51 (62.2) 31 (37.8)  Suspected 84 (42.9) 112 (57.1)  Non-cases 67 (50.4) 66 (49.6) Prior antibiotic treatment  Yes 45 (43.3) 59 (56.7) 0.096  No 169 (52.6) 152 (47.4) Impairment of consciousness  Yes 102 (52.0) 94 (48) 0.52  No 112 (48.9) 117 (51.1) Focal neurologic deficit  Yes 22 (66.7) 11 (33.3) 0.051  No 192 (49.0) 200 (51.0) CSF appearance  Turbid 37 (64.9) 20 (35.1) <0.001*  Normal 68 (38.2) 110 (61.8) Detection of organism by Gram stain  Yes 11 (84.6) 2 (15.4) 0.003*  No 89 ((43.0) 118 (57.0) HIV status  Positive 7 (30.4) 16 (69.6) 0.048*  Negative 207 (51.6) 194 (48.4) Hospital  Jimma 52 (42.6) 70 (57.4) <0.001  Gondar 24 (26.1) 68 (73.9)  Hawassa 65 (76.5) 20 (23.5)  Arba Minch 73 (57.9) 53 (42.1) *statistically significant Dexamethasone treatment was associated with an increase of the in-hospital mortality, COR = 3.18 (95 % CI 1.90–5.33); p < 0.001 and low GOS at discharge, COR = 2.65 (95 % CI 1.76–3.99); P < 0.001. However, there was no association with neurologic sequelae at discharge (Fig. 1).Fig. 1 Association between adjuvant dexamethasone treatment and discharge outcomes in patients treated as bacterial meningitis in Ethiopia, 2011–2015. COR – Crude odds ratio, GOS – Glasgow Outcome Score. * Denotes statistical significance As depicted on Table 2, dexamethasone was found to be one of the factors independently associated with poor outcome on multivariable analysis with the best-fit model. When further analysis was performed controlling for all potential confounders on multiple logistic regressions with forced entry, this association persisted. For instance, the odds of having low GOS at discharge was nearly 4 times, AOR = 3.94 (95 % CI 1.63–9.53; P = 0.002) and its association with in hospital death was also nearly as much, AOR = 3.60 (1.97–6.60; P < 0.001). Controlling for these individual variables also revealed similar finding as shown on Table 5.Table 5 Subgroup analysis for association of dexamethasone with discharge outcome in patients treated as bacterial meningitis at teaching hospitals in Ethiopia, 2011–2015 Variables Death Low GOS OR 95 % CI P OR 95 % CI P Age   < 50 years 3.08 1.74–5.48 <0.001* 2.83 1.79–4.48 <0.001*   > = 50 years 4.00 1.19–13.46 0.025* 2.68 1.0–7.21 0.051 Duration of illness   < =2 days 4.02 1.37–11.77 0.011* 2.05 0.94–4.50 0.072   > 2 days 2.93 1.62–5.30 <0.000* 2.91 1.79–4.72 <0.001* Level of conscious  14 and 15 2.21 0.94–5.21 0.070 1.82 0.996 0.051   < 14 3.69 1.84–7.40 <0.001* 388 20.3–7.44 <0.001* HIV status  Positive 3.25 0.46–22.93 0.237 18.00 1.63–198.51 0.018*  Negative 3.38 1.91–5.66 <0.001* 2.49 1.63–3.80 <0.001* Prior antibiotics  Yes 3.59 1.24–10.38 0.018* 5.97 2.47–14.44 <0.001*  No 3.04 1.68–5.50 <0.001* 2.07 1.30–3.29 0.002* Focal neurologic deficit  Yes 3.20 0.67–15.38 0.146 0.80 0.16–3.99 0.789  No 3.00 1.72–5.23 <0.001* 2.77 1.80–4.29 0.003* CSF appearance  Turbid 2.72 0.67–11.11 0.163 2.44 0.79–7.51 0.121  Clear 6.26 1.95–20.12 0.002* 4.28 2.14–8.58 <0.001* TB suspected  Yes 3.90 0.45–34.02 0.218 3.46 0.83–14.36 0.087  No 3.09 1.79–5.34 <0.001* 2.18 1.40–3.39 0.001* Diagnosis of meningitis  Confirmed – 1 1 – 1 1  Probable 2.31 0.68–7.86 0.180 3.22 1.24–8.36 0.016*  Suspected 3.38 1.73–6.59 <0.001* 2.09 1.16–3.75 0.014*  Non-cases 6.98 1.50–32.56 0.013* 3.17 1.41–7.15 0.005* *Statistically significant Dexamethasone and mortality – This association did not differ between older and younger patients, whether a patient had delayed presentation or not, and whether a patient took prior antimicrobial treatment or not. However, this association faded in HIV infected patients, those with neurologic deficit on presentation, those with abnormal CSF findings and TB suspected cases. Dexamethasone and low GOS – Age of the patient, presence or absence of HIV infection, and prior antibiotic treatment did not change the nature of this association. On the other hand, the association disappeared in TB suspected cases, those with confirmed ABM and neurologic deficit on presentation. In summary, adjunctive dexamethasone treatment was associated with poor outcome in most of the sub-groups. However, this was not the case in patients with a diagnosis of bacterial meningitis that was microbiologically proven or supported by CSF findings (those with turbid CSF, proven BM and probable cases of BM), and in those suspected to have TBM. In these cases, dexamethasone therapy was not associated with any positive or negative discharge outcomes (Table 5). The only one instance where dexamethasone was associated with positive outcome is decreased discharge neurologic sequelae in patients who had impaired consciousness at presentation, AOR = 0.42 (95 % CI 0.19–0.94; P = 0.033). Discussion Our study revealed that patients treated for clinically suspected ABM in teaching hospitals in Ethiopia were found to have a high mortality of 20 %. However, most of these patients did not receive a proper diagnostic workup and were treated only pragmatically [31]. Hence, the reported mortality may not reflect the true burden of the problem in the settings. Moreover, adjunctive dexamethasone treatment in patients who did not receive proper CSF analysis or in whom CSF findings were not compatible with acute bacterial meningitis was associated with poor discharge outcomes. The finding in our study may not reflect the real mortality associated with ABM in the setting. This is because the diagnosis of ABM in substantial fraction of the patients was unclear and it is likely that many of these patients suffered from diagnosis other than ABM [31]. Therefore, it is not possible to compare the data to other studies in patients with proven acute bacterial meningitis. It further explains why the mortality of 20 % is much lower than in meningitis studies from other low income countries like Malawi, where mortality reached 40 % [5, 6]. This is also reflected in the documented short-term neurologic sequela in survivors which was only 11.2 %, a rate that is lower than in most reports on the outcome of ABM [3, 5, 6, 15, 34, 35]. As expected, impaired consciousness at admission was associated with poor outcome. This may be due to the severity of the illness as well as the occurrence of complications like aspiration pneumonia which on itself was associated with a higher mortality. However, conventional poor prognostic indicators like age, causative bacteria, duration of illness, and HIV infection were not found to be associated with adverse outcome. The study was not able to detect an effect of these factors due to small number of confirmed cases and due to the fact that only the outcome at leaving the hospital was assessed. The study may also be underpowered to detect an association because of the small number of HIV cases. However, 17.9 % of patients did not receive HIV test which might have underestimated its real prevalence. Adjuvant corticosteroid in management of bacterial meningitis in low and middle income countries, and in settings with high HIV prevalence in particular, has never been proven beneficial [28–30]. To date, there is no recommendation of its use in such settings. However, physicians in settings with little evidence and diagnostic facilities like Ethiopia have continued its use based on recommendation for high income countries [36, 37]. The finding in our study, where dexamethasone was used in half of the patients, is a testimony of scepticism towards the current recommendation in low income countries. As it highlighted on Table 5, dexamethasone use was not associated with any positive or negative outcome in those patients with confirmed or probable cases of bacterial meningitis. The lack of evidence for its benefit in Ethiopia is consistent with previous findings from similar settings [28–30], although the number of patients with proven acute bacterial meningitis was too low to address this question. More important, however, this study demonstrated that dexamethasone treatment seems to be harmful in patients who were treated as ABM without any CSF findings that supported the diagnosis of ABM or when CSF was not analysed. One likely explanation for its association with unfavorable outcome is that, as it has repeatedly been highlighted in this paper, the majority of the patients likely suffered from other alternative diagnoses. Dexamethasone administration to patients with severe brain or systemic infections without simultaneously treating the underlying disease condition could have resulted in the poorer outcomes. As this is the first study that looked at the effect of dexamethasone in patients treated with only clinically presumptive diagnosis of ABM, it makes it valuable for the real situation in developing countries. However, our study has multiple limitations worth mentioning. This is a retrospective study which may be affected by poor documentation and loss of medical records. Moreover, patients who were given dexamethasone might have had severe disease from the outset and, hence, the poor outcome could be due to their underlying medical condition rather than the dexamethasone use per se. Last but not least, the study may not be representative for the country because it involved mainly teaching hospitals. Nevertheless, it must be assumed that the rate of treatment of cases with suspected ABM lacking any diagnostic workup is even higher throughout the country. Conclusion Adjuvant dexamethasone use in management of suspected but unproven cases of bacterial meningitis in teaching hospitals in Ethiopia was associated with an increased mortality and poor discharge GOS. These findings re-affirm the lack of evidences for its broad use for presumed meningitis in low income countries and show that there are potential deleterious effects in unconfirmed cases. Physicians practising under such circumstances should abide with the current recommendations and defer the use of adjuvant corticosteroid in clinically suspected cases of bacterial meningitis without CSF alterations that support the diagnosis. Additional file Additional file 1: Assessment of treatment strategies for bacterial meningitis in Ethiopia. The English version of the tool used for data collection. (PDF 836 kb) Abbreviations ABMAcute bacterial meningitis AORAdjusted odds ratio CIConfidence interval CORCrude odds ratio CSFCerebrospinal fluid GCSGlasgow coma scale GOSGlasgow outcome scale HIVHuman immunodeficiency virus LAMALeft against medical advice LOSLength of (hospital) stay OROdds ratio SDStandard deviation TBMTuberculous meningitis WHOWorld health organization Acknowledgements We would like to thank Else Kröner-Fresenius-Stiftung (EKFS) for funding the project. Our heartfelt thanks also go to all the four hospitals and their staffs for smooth facilitation of data collection. Funding The financial input of the Else-Kröner-Fresenius-Foundation (EKFS) is greatly appreciated. The organization did not have any direct role in designing of the study, data collection and analysis, interpretation of data and in writing the manuscript. Availability of data and materials All data and materials related this manuscript will be readily available for the public upon its acceptance for publication. The data can be requested from authors. Authors’ contributions EKG designed the study, developed the instruments, supervised data collection, analysed the data and wrote the manuscript. MK, MT and HWP took part in study design, instrument development, data analysis and writing of manuscript. AA, KL, AW and TS participated in instrument development, took part in data collection and reviewed the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable – This manuscript does not contain any individual person’s data. Ethics approval and consent to participate The study was conducted after ethical approval was obtained from Jimma University Ethical Review Board (Reference letter – RPGC/4026/2015) and permission was obtained from each of the four hospitals. Consent from participant was not needed as it was a retrospective study and this was waived by the ethics committee. Confidentiality of the data was assured through anonymity. ==== Refs References 1. Durand ML Calderwood SB Weber DJ Miller SI Southwick FS Caviness VS Jr Acute bacterial meningitis in adults. A review of 493 episodes N Engl J Med 1993 328 1 21 8 10.1056/NEJM199301073280104 8416268 2. Dzupova O Rozsypal H Prochazka B Benes J Acute bacterial meningitis in adults: predictors of outcome Scand J Infect Dis 2009 41 5 348 54 10.1080/00365540902849391 19306157 3. van de Beek D de Gans J Tunkel AR Wijdicks EF Community-acquired bacterial meningitis in adults N Engl J Med 2006 354 1 44 53 10.1056/NEJMra052116 16394301 4. Scarborough M Thwaites GE The diagnosis and management of acute bacterial meningitis in resource-poor settings Lancet Neurol 2008 7 7 637 48 10.1016/S1474-4422(08)70139-X 18565457 5. Wall EC Cartwright K Scarborough M Ajdukiewicz KM Goodson P Mwambene J High mortality amongst adolescents and adults with bacterial meningitis in sub-Saharan Africa: an analysis of 715 cases from Malawi PLoS One 2013 8 7 e69783 10.1371/journal.pone.0069783 23894538 6. Gordon SB Walsh AL Chaponda M Gordon MA Soko D Mbwvinji M Bacterial meningitis in Malawian adults: pneumococcal disease is common, severe, and seasonal Clin Infect Dis 2000 31 1 53 7 10.1086/313910 10913396 7. Radetsky M Duration of symptoms and outcome in bacterial meningitis: an analysis of causation and the implications of a delay in diagnosis Pediatr Infect Dis J 1992 11 9 694 8 10.1097/00006454-199209000-00002 1448306 8. Grimwood K Anderson VA Bond L Catroppa C Hore RL Keir EH Adverse outcomes of bacterial meningitis in school-age survivors Pediatrics 1995 95 5 646 56 7536915 9. Molyneux EM Tembo M Kayira K Bwanaisa L Mweneychanya J Njobvu A The effect of HIV infection on paediatric bacterial meningitis in Blantyre, Malawi Arch Dis Child 2003 88 12 1112 8 10.1136/adc.88.12.1112 14670782 10. Domingo P Suarez-Lozano I Torres F Pomar V Ribera E Galindo MJ Bacterial meningitis in HIV-1-infected patients in the era of highly active antiretroviral therapy J Acquir Immune Defic Syndr 2009 51 5 582 7 10.1097/QAI.0b013e3181adcb01 19512939 11. Madhi SA Madhi A Petersen K Khoosal M Klugman KP Impact of human immunodeficiency virus type 1 infection on the epidemiology and outcome of bacterial meningitis in South African children Int J Infect Dis 2001 5 3 119 25 10.1016/S1201-9712(01)90085-2 11724667 12. Aronin SI Peduzzi P Quagliarello VJ Community-acquired bacterial meningitis: risk stratification for adverse clinical outcome and effect of antibiotic timing Ann Intern Med 1998 129 11 862 9 10.7326/0003-4819-129-11_Part_1-199812010-00004 9867727 13. Klein M Pfister HW Leib SL Koedel U Therapy of community-acquired acute bacterial meningitis: the clock is running Expert Opin Pharmacother 2009 10 16 2609 23 10.1517/14656560903277210 19827989 14. Grandgirard D Gaumann R Coulibaly B Dangy JP Sie A Junghanss T The causative pathogen determines the inflammatory profile in cerebrospinal fluid and outcome in patients with bacterial meningitis Mediat Inflamm 2013 2013 312476 10.1155/2013/312476 15. van de Beek D de Gans J Spanjaard L Weisfelt M Reitsma JB Vermeulen M Clinical features and prognostic factors in adults with bacterial meningitis N Engl J Med 2004 351 18 1849 59 10.1056/NEJMoa040845 15509818 16. Lepur D Barsic B Community-acquired bacterial meningitis in adults: antibiotic timing in disease course and outcome Infection 2007 35 4 225 31 10.1007/s15010-007-6202-0 17646915 17. van de Beek D de Gans J McIntyre P Prasad K Steroids in adults with acute bacterial meningitis: a systematic review Lancet Infect Dis 2004 4 3 139 43 10.1016/S1473-3099(04)00937-5 14998499 18. van de Beek D de Gans J Adjunctive corticosteroids in adults with bacterial meningitis Curr Infect Dis Rep 2005 7 4 285 91 10.1007/s11908-005-0061-6 15963330 19. Assiri AM Alasmari FA Zimmerman VA Baddour LM Erwin PJ Tleyjeh IM Corticosteroid administration and outcome of adolescents and adults with acute bacterial meningitis: a meta-analysis Mayo Clin Proc 2009 84 5 403 9 10.1016/S0025-6196(11)60558-2 19411436 20. van de Beek D Farrar JJ de Gans J Mai NT Molyneux EM Peltola H Adjunctive dexamethasone in bacterial meningitis: a meta-analysis of individual patient data Lancet Neurol 2010 9 3 254 63 10.1016/S1474-4422(10)70023-5 20138011 21. Peterkovic V Trkulja V Kutlesa M Krajinovic V Lepur D Dexamethasone for adult community-acquired bacterial meningitis: 20 years of experience in daily practice J Neurol 2012 259 2 225 36 10.1007/s00415-011-6150-6 21706150 22. Bodilsen J Dalager-Pedersen M Schonheyder HC Nielsen H Dexamethasone treatment and prognostic factors in community-acquired bacterial meningitis: a Danish retrospective population-based cohort study Scand J Infect Dis 2014 46 6 418 25 10.3109/00365548.2014.887223 24645971 23. Pfausler B Schmutzhard E Controversies in neurology, Vienna, 2012: steroids in bacterial meningitis: no J Neural Transm 2013 120 2 343 6 10.1007/s00702-012-0939-z 23263539 24. Brouwer MC McIntyre P Prasad K van de Beek D Corticosteroids for acute bacterial meningitis Cochrane Database Syst Rev 2013 6 CD004405 23733364 25. Borchorst S Moller K The role of dexamethasone in the treatment of bacterial meningitis - a systematic review Acta Anaesthesiol Scand 2012 56 10 1210 21 10.1111/j.1399-6576.2012.02698.x 22524556 26. Spapen H Van Berlaer G Moens M Hubloue I Adjunctive steroid treatment in acute bacterial meningitis. “To do or not to do: that is the question” Acta Clin Belg 2011 66 1 42 5 10.1179/ACB.66.1.2062512 21485762 27. Peltola H Leib SL Performance of adjunctive therapy in bacterial meningitis depends on circumstances Pediatr Infect Dis J 2013 32 12 1381 2 10.1097/INF.0000000000000066 24569310 28. Molyneux EM Walsh AL Forsyth H Tembo M Mwenechanya J Kayira K Dexamethasone treatment in childhood bacterial meningitis in Malawi: a randomised controlled trial Lancet 2002 360 9328 211 8 10.1016/S0140-6736(02)09458-8 12133656 29. Scarborough M Gordon SB Whitty CJ French N Njalale Y Chitani A Corticosteroids for bacterial meningitis in adults in sub-Saharan Africa N Engl J Med 2007 357 24 2441 50 10.1056/NEJMoa065711 18077809 30. Nguyen TH Tran TH Thwaites G Ly VC Dinh XS Ho Dang TN Dexamethasone in Vietnamese adolescents and adults with bacterial meningitis N Engl J Med 2007 357 24 2431 40 10.1056/NEJMoa070852 18077808 31. Gudina EK Tesfaye M Adane A Lemma K Shibiru T Pfister HW Challenges of bacterial meningitis case management in Ethiopia Trop Med Int Health 2016 21 7 870 8 10.1111/tmi.12720 27145202 32. Ethiopian Central statistical agency Addis, Ethiopia2014 [cited 2016 20 April]. Available from: http://www.csa.gov.et/. Accessed 20 Apr 2016. 33. World Health Organization WHO–recommended standards for surveillance of selected vaccine-preventable diseases 2003 Geneva WHO 34. Ostergaard C Konradsen HB Samuelsson S Clinical presentation and prognostic factors of Streptococcus pneumoniae meningitis according to the focus of infection BMC Infect Dis 2005 5 93 10.1186/1471-2334-5-93 16253143 35. McMillan DA Lin CY Aronin SI Quagliarello VJ Community-acquired bacterial meningitis in adults: categorization of causes and timing of death Clin Infect Dis 2001 33 7 969 75 10.1086/322612 11528567 36. Drug Administration and Control Authority of Ethiopia Acute bacterial meningitis. Standard Treatment Guidelines for General Hospital 2010 2 Addis Ababa DACA 68 72 37. Ghana National Drug Policy (GNDP) Meningitis. Standard Treatment Guidelines 2010 Ghana Ministry of Health (GNDP) Ghana 356 62

PMC005xxxxxx/PMC5000409.txt

==== Front BMC Med Res MethodolBMC Med Res MethodolBMC Medical Research Methodology1471-2288BioMed Central London 20810.1186/s12874-016-0208-1Research ArticleCohort Multiple Randomised Controlled Trials (cmRCT) design: efficient but biased? A simulation study to evaluate the feasibility of the Cluster cmRCT design http://orcid.org/0000-0002-0849-3458Pate Alexander +44(0)161 306 0662alexander.pate@manchester.ac.uk 1Candlish Jane +44(0)161 306 0662jane.candlish@sheffield.ac.uk 1Sperrin Matthew +44(0)161 306 0662matthew.sperrin@manchester.ac.uk 1Van Staa Tjeerd Pieter +44(0)161 306 0662tjeerd.vanstaa@manchester.ac.uk 12on behalf of GetReal Work Package 2 1 Health eResearch Centre, Farr Institute for Health Informatics Research, University of Manchester, 1.003 Vaughan House, Manchester, M13 9PL, UK 2 Utrecht Institute of Pharmaceutical Sciences, Utrecht University, Utrecht, The Netherlands 26 8 2016 26 8 2016 2016 16 1 10925 3 2016 10 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background The Cohort Multiple Randomised Controlled Trial (cmRCT) is a newly proposed pragmatic trial design; recently several cmRCT have been initiated. This study tests the unresolved question of whether differential refusal in the intervention arm leads to bias or loss of statistical power and how to deal with this. Methods We conduct simulations evaluating a hypothetical cluster cmRCT in patients at risk of cardiovascular disease (CVD). To deal with refusal, we compare the analysis methods intention to treat (ITT), per protocol (PP) and two instrumental variable (IV) methods: two stage predictor substitution (2SPS) and two stage residual inclusion (2SRI) with respect to their bias and power. We vary the correlation between treatment refusal probability and the probability of experiencing the outcome to create different scenarios. Results We found ITT to be biased in all scenarios, PP the most biased when correlation is strong and 2SRI the least biased on average. Trials suffer a drop in power unless the refusal rate is factored into the power calculation. Conclusions The ITT effect in routine practice is likely to lie somewhere between the ITT and IV estimates from the trial which differ significantly depending on refusal rates. More research is needed on how refusal rates of experimental interventions correlate with refusal rates in routine practice to help answer the question of which analysis more relevant. We also recommend updating the required sample size during the trial as more information about the refusal rate is gained. Keywords Trials within CohortsCohort multiple randomised controlled trialClusterPragmaticInstrumental variablehttp://dx.doi.org/10.13039/501100004963Seventh Framework Programme115546Van Staa Tjeerd Pieter http://dx.doi.org/10.13039/501100000265Medical Research CouncilMR/K006665/1issue-copyright-statement© The Author(s) 2016 ==== Body Background Randomised controlled trials (RCTs) often fail to meet recruitment targets and are costly [1]. This problem can be even more prevalent in comparative effectiveness research where more patients are needed to detect smaller differences between treatments. Furthermore, the results from most randomised controlled trials may not be generalisable to routine practice [2], yet we use the results from these trials to inform clinical decision making [3]. There is a clear need for more pragmatic trials which are cost efficient, integrated with routine clinical care, have less stringent entry criteria and can address the clinical questions that current RCTs cannot [4–7]. The Cohort Multiple Randomised Controlled Trial (cmRCT) design can simplify the recruitment and conduct of trials compared with current RCTs. It was first proposed in 2010 [8], and is beginning to be used in practice with a total of 5 registered trials from 7 cohorts [9–16]. In this design, a large cohort is identified (e.g., patients at high risk of cardiovascular disease [CVD]) and followed using routinely collected data such as electronic health records [17]. The same cohort can be used for multiple interventions. Each intervention is offered to a randomly selected sample of patients eligible for that intervention, who are then compared with the rest of the eligible patients from the cohort that are still being treated as usual [8]. Randomisation can occur either at a patient or a cluster (site) level. The cluster design can offer dramatically improved accrual [1] and can further reduce costs through the implementation of the interventions in fewer places; cluster designs are the focus of this paper. The main advantages of cmRCTs are the low cost of recruiting the control group, the possibility to use it for multiple trials and the comparison of interventions to real life practice (the control group are not contacted for further consent). However, refusal to participate in the trial happens post randomisation so excluding these patients may result in selection bias. Alternatively an intention to treat (ITT) analysis could be used; however the refusal rates in a cmRCT may not reflect those in routine practice as the intervention may be viewed as experimental. In this case the ITT effect may lack interpretation outside the trial setting. Depending on the refusal rate it may be preferable to calculate the effect of accepting the treatment, this will be referred to as the treatment effect for the remainder of this paper. Refusal can cause a loss of statistical power and a bias in the estimation of the treatment effect, particularly if it is correlated with the outcome of interest. Instrumental variable analysis (IV) is a method to account for unmeasured confounding in epidemiological studies [18, 19] and can for non-compliance in RCTs [20, 21]. It is also applicable to the problem of treatment refusal in a cmRCT setting. The aim of this paper is (1) to estimate the extent of bias and loss of statistical power with various refusals scenarios, (2) to test the robustness of IV methods to correct for bias due to refusals and (3) devise strategies to account for loss in power. There is currently very little literature on the topic of cmRCTs, we provide practical recommendations to trial designers and decision makers on the conditions under which cluster cmRCT is a viable design for point of care trials and which statistical analysis methods to use. Methods A series of simulations are performed using Base SAS 9.4 Software in which cluster cmRCTs are conducted. In order to provide more realistic simulations they are based on an example of a cohort of patients at high risk of developing cardiovascular disease (CVD) and eligible for lipid lowering drugs according to the relevant criteria in the principal UK guidelines [22]. A novel intervention is tested against treatment as usual with a primary outcome of the time until a CVD event. This is an outcome that is of direct importance to a patient and may be identified with routinely collected data. Three patient characteristics are simulated: probability of refusing the intervention treatment, the risk of having a CVD event, and the time to death or censoring. Different scenarios are created by changing the average refusal probability of the population and changing the correlation between individuals’ risk of having an event and their probability of refusing treatment. The probability of a clinician refusing to offer the treatment to each patient is also simulated, and correlated to varying extents with patient risk. Once the patient characteristics have been generated, trial data is simulated through the same process of a cmRCT: treatment randomisation, refusal of treatment, application of intervention to those who accept and then the generation of times until an event. Weibull distributions are used to generate survival times. Each of the analysis methods explained in Analysis Methods are then applied to the simulated trial data to estimate the intervention effect. The exact simulation process is detailed in Simulation procedure. Analysis methods Four different methods for the analysis of a cluster cmRCT are tested. The methods are ITT, per protocol (PP) and two IV methods. ITT is the recommended method of analysis in pragmatic trials [5, 23, 24] analysing the groups based on the random treatment allocation. PP defines the treatment groups on the basis of the actual treatment received, with only those who follow the allocated treatment included in the analysis. The two IV methods tested are the two stage predictor substitution (2SPS) and two stage residual inclusion (2SRI) as outlined practically by Terza et al., [25]. They are both two stage modelling techniques and start by fitting a first stage model with treatment allocation as the explanatory variable and treatment received as the dependent variable (here treatment allocation acts as the IV). This model is then used to calculate the predicted values for treatment received and the residuals. In 2SPS, a second stage model is fitted to the outcome data using the predicted values for treatment received as the explanatory variables. In 2SRI, the second stage model is fitted to the outcome data using both the residuals and the actual treatment received as explanatory variables. The standard errors of parameter estimates in two stage modelling procedures are too small hence non-parametric bootstrapping [26] should be used to calculate them. IV estimators were the chosen method to estimate the causal effect as IV methods are believed to perform well in RCTs with non-compliance with assumptions more easily argued to hold. [27]. There is a wealth of literature on the theoretical properties of causal effect estimates and IVs [18–21, 28] which is not recited in this paper. Instead the performances of the four different analysis methods in a variety of scenarios are evaluated with respect to bias, standard error and statistical power. We define bias as the error in the estimation of the treatment effect as defined in section 1 (effect of accepting treatment). Simulation procedure Table 1 contains details of all variables used in the simulations. The cluster size chosen is J = 620 to match the average number of eligible patients per UK practice. This is calculated using published figures on GP practice size from the Health and Social Care Information Centre (HSCIC) [29] and statistics on the prevalence of CVD from the National Institute for Health and Care Excellence (NICE) [30]. The cluster size J is constant as it has been shown that variable cluster size has no effect on the results in terms of bias [31, 32]. The variances of the individual and cluster level random effects, σε2 and σu2, and the shape and scale of the Weibull distribution for time to CVD event, λc and γc, are chosen to match the mean 10-year CVD risk to published figures of 21.1 % (standard deviation 8.6 %) [22]. The mortality (censoring distribution) shape and scale, γm and λm, and the variances σε2 and σu2 give censoring of 5 % of all events and a correlation of 0.25 between Tikc and Tikm to represent informative censoring.Table 1 Description of all variables used in simulation Number of patients in cohort, control arm and intervention arm N, Ncon, Nint Number of clusters in trial K Size of each cluster J = 620 Treatment allocated to kth cluster Zk = 0/1 for control/intervention Treatment received by ith individual from kth cluster Xik = 0/1 for control/intervention Time until CVD event for ith individual from the kth cluster Tikc∼Weibullγc,λce−βXik+εik+Uk/γc Time until mortality (censoring distribution) for the ith individual from the kth cluster Tikm∼Weibullγm,λme−εik+Uk/γm Common baseline hazard function for time until CVD event hct=γctγc−1/λcγc,γc=1.2,λc=36 Common baseline hazard function for time until mortality hm(t)=γmtγm−1/λmγm,γm=1.2,λm=55 Individual hazard function for time until CVD event hikct=hcteεik+Uk+βXik Individual hazard function for time until mortality hikmt=hmteεik+Uk Individual level random effects ɛ ik ∼ N(0, σ ɛ 2) Cluster level random effects U k ∼ N(0, σ u 2) Intervention effect β = − 0.32 Ten year risk of a CVD event rik = P(Tikc < 10| Xik = 0, εik, Uk) Individual and average probability of patient refusing treatment pik,p=∑i,kpikN Individual and average probability of clinician refusing to offer treatment qik,q=∑i,kqikN Correlation between patient refusal probability and patient risk ρ p Correlation between clinician refusal probability and patient risk ρ q Censoring indicator C ik = I(T ik c ≥ min(T ik m, T max)) Trial follow up time T max = 3 Random variable observed for each patient Y ik = min(T ik c, T ik m, T max) For each scenario detailed previously, the following procedure was implemented:For j = 1,2,…,1000:Generate the random effects ɛik and Uk for each patient and cluster. i = 1,2,…,I. k = 1,2,…,K. For each patient, calculate unique 10 year risks (under the counterfactual scenario of receiving standard care) of a CVD event, rik. Assign patient and clinician refusal probabilities pik and qik Order patients by their risk, rik. Assign refusal probabilities sequentially in a linear fashion between the lower limit (LL) and upper limit (UL) such that such that ∑ pik/N = p and ∑ qik/N = q. Randomise treatment allocation Zk to control or intervention on a 4:1 basis. Zk = 0/1 if assigned to control/intervention. Generate the treatment received where Xik = 0/1 if control/intervention is received. If Zk = 0 then Xik = 0, if Zk = 1 then Xik = min {Bernoulli(1 − pik), Bernoulli(1 − qik)}. Apply intervention effect β and random effects to the hazard function, hikct=hcteβXik+εik+Uk. Generate survival times Tikc and Tikm. These survival distributions correspond to the respective hazard functions. Generate the censoring indicator Cik, The total observed trial data is then {Yik, Cik, Zik, Xik}, a set of censored survival data, treatment allocations and treatments received. Fit a Cox proportional hazards model to the data with respect to the four analysis methods ITT, PP, 2SRI and 2SPS, to produce an estimate β^j of the intervention effect β, which is the log of the hazard ratio, and record the p-value, pj. When j = 1000, calculate the mean β¯=∑jβ^j/1000, the percentage bias β¯−β/β and the statistical power ∑jI(pj < 0.05)/1000. Also, calculate a parametrically bootstrapped standard error of the individual estimate s.e.β^=s.d.β^j, the standard error of the mean s.e.β¯=s.d.β^j/1000, and a confidence interval for the percentage bias CI=[100*β¯−1.96*s.e.β¯−β/β,100*β¯+1.96*s.e.β¯−β/β. Different scenarios are created by varying the following variables. The intra-cluster coefficient (ICC) takes values 0.025 and 0.05, simulated by (σε2, σu2) = (0.6, 0.2) and (0.57, 0.27) respectively. Average patient and clinician refusal probabilities p and q take values 0.1, 0.2 and 0.3. The correlation between refusal probability and risk, ρp, takes values zero, low, medium and high, simulated by having lower limits and upper limits for individual refusal probabilities as LLUL∈pp2p34p3p35p30,2p. The correlation between clinician refusal and risk takes the same set of values. The reason for this structure is to give control over the correlation between individual risk and refusal probabilities. The treatment effect is fixed at β = − 0.32, which equates to on average a 25 % reduction in 10 year risk of CVD. 1000 independent sets of independent trial data are generated for each scenario. Sample sizes are calculated at a fixed ratio of 4:1 control to study intervention, the type 1 error is 0.05 and required power is 0.8. Sample sizes are calculated through simulation [33] as sample size formulas for informatively censored clustered survival data are not common. Trials characteristics (effect size, refusal rate, baseline risk) are assumed to be known. Trial data is simulated using the above process and analysed using ITT. For each combination of refusal rates, the smallest N (that is a multiple of J = 620) such that the proportion of p-values < 0.05 is 80 % is chosen as the required sample size in that scenario. There are then two recruitment methods which alter the required sample size. Recruitment method 1 calculates the sample size assuming no refusal. Recruitment method 2 factors in the refusal rate in the sample size calculation (assuming refusal to be non-informative and independent of individual risks). All simulation scenarios are run using both recruitment methods. The power realised varies from 0.8 as we use the smallest number of clusters that achieve at least a power of 0.8, in recruitment method 2 this changes depending on the refusal rate. The outcome of interest is the time until a CVD event so cox proportional hazards models are fitted to produce estimates for the intervention effect. To account for the clustering of the data, three types of Cox proportional hazard model are fitted: marginal, lognormal frailty, and gamma frailty models [34, 35]. The lognormal model is correctly specified because the generated random effects (frailties) are normally distributed (Table 1), whereas the gamma frailty model is miss specified. The output from the robust marginal model has a different interpretation to the frailty models in that the hazard ratio returned is between any two randomly selected patients from the population, as opposed to the hazard ratio of any two people randomly selected from the same cluster [35]. Clustering is not taken into account in the first stage of the IV model as the inclusion of residuals in the second stage model (2SRI) is expected to take account of variation in refusal rates between clusters. Results Figure 1 shows the magnitude of bias and loss of statistical power for the four analysis methods for varying average refusal rates and negative correlation between refusal and individual risk (for recruitment method 1). As expected, ITT underestimates the treatment effect. Refusal probabilities as small as 0.1 lead to bias between 9 and 16 %, refusal of 0.2 between 18 and 30 %, and refusal of 0.3 between 21 and 42 %, depending on the direction of the correlation. PP provides the most biased effect estimates when correlations are large with substantial reductions in statistical power. The two IV analyses provide similar results and substantially less biased effect estimates compared with ITT. The bias with both IV methods is below 6 % in all scenarios except when refusal is high with negative correlation, in which case 2SPS and 2SRI overestimate the effect of the study intervention by 13 and 17 % respectively. All methods have reduced statistical power with increasing refusals. ITT has the same statistical power as the 2SPS method in nearly all scenarios (i.e., the lines in Fig. 1 overlapping). For positive correlation (Fig. 2) 2SRI provides effect estimates with the lowest level of bias compared to the three other analysis methods, although it is associated with a statistical power between 0.8 and 0.56 rather than the 0.83 obtained from the sample size calculation. The trends in bias and power for ITT and 2SPS change direction with the correlation causing an increase in bias and a drop in power.Fig. 1 Percentage bias and power of the four analysis methods for varying levels of patient refusal and correlation between individual patient refusal probabilities and risk. Clinician refusal = 0, correlation is negative, recruitment method 1 is used and ICC =0.025. The black line in the power graph represents the expected power in the trial. A lognormal frailty model is fitted to the data Fig. 2 Percentage bias and power of the four analysis methods for varying levels of patient refusal and correlation between individual patient refusal probabilities and risk. Clinician refusal = 0, correlation is positive, recruitment method 1 is used and ICC =0.025. The black line in the power graph represents the expected power in the trial. A lognormal frailty model is fitted to the data Figure 3 shows the results for recruitment method 2 and negative correlation between refusal and individual risk. There is no difference from recruitment method 1 with respect to bias or trends in statistical power; however the overall power of the trial stays consistent as refusal probabilities are changed. The only visible effect of changing refusal probabilities is to strengthen the effect of changing correlation. Importantly, the power of ITT and IV methods tends not to drop below the desired level. With positive correlations and recruitment method 2 (Fig. 4) there is a similar pattern when comparing to recruitment method 1. 2SRI provides the least biased estimates with the statistical power ranging between 0.81 and 0.88, depending on refusal. 2SPS and ITT yield more biased estimates with increasing reductions in statistical power as the rate of refusal increased.Fig. 3 Percentage bias and power of the four analysis methods for varying levels of patient refusal and correlation between individual patient refusal probabilities and risk. Clinician refusal = 0, correlation is negative, recruitment method 2 is used and ICC =0.025. The black line in the power graph represents the expected power in the trial. A lognormal frailty model is fitted to the data Fig. 4 Percentage bias and power of the four analysis methods for varying levels of patient refusal and correlation between individual patient refusal probabilities and risk. Clinician refusal = 0, correlation is positive, recruitment method 2 is used and ICC =0.025. The black line in the power graph represents the expected power in the trial. A lognormal frailty model is fitted to the data Sensitivity analyses are conducted evaluating the effects of miss-specification of the statistical models. Figure 5 and Fig. 6 show the estimates and statistical power with a robust marginal and gamma frailty model fitted to the data respectively (for recruitment method 2 and negative correlations). The miss-specified models perform slightly worse than the correctly specified models in terms of percentage bias while the statistical power is slightly higher. If there was a greater variation between clusters, you would expect a lack of collapsibility to cause a difference between the marginal and frailty estimates [36]. This should be considered along with what output is desired (conditional/marginal) if running a cluster cmRCT. Results for positive correlations also did not vary greatly with model miss-specifications (data not shown).Fig. 5 Percentage bias and power of the four analysis methods for varying levels of patient refusal and correlation between individual patient refusal probabilities and risk. Clinician refusal = 0, correlation is negative, recruitment method 2 is used and ICC =0.025. The black line in the power graph represents the expected power in the trial. A robust marginal frailty model is fitted to the data Fig. 6 Percentage bias and power of the four analysis methods for varying levels of patient refusal and correlation between individual patient refusal probabilities and risk. Clinician refusal = 0, correlation is negative, recruitment method 2 is used and ICC =0.025. The black line in the power graph represents the expected power in the trial. A robust gamma frailty model is fitted to the data In Appendix A a look into the accuracy of the bias estimates are provided and results from scenarios including changes in the ICC and when clinician refusal > 0. Discussion This study has shown that refusal of a novel intervention in a cluster cmRCT design can lead to bias and reductions in power. The ITT estimates have a high bias, which increases with increasing refusal and is affected by correlations between refusal and the risk of outcome of interest. IV analyses using the 2SRI method substantially reduce the bias but yield small overestimates (generally < 6 %) of the treatment effect when refusal rates are high and correlation strong. The 2SPS estimates for IV are highly affected by the correlation structure and produce very biased estimates with positive correlation. Recruitment method 1 causes a loss in statistical power irrelevant of the analysis method used. On the other hand, cluster cmRCT are correctly powered or overpowered for recruitment method 2. This study found, as expected, that refusals can lead to a large underestimate of the treatment effect when ITT is used (also known as dilution bias) [37]. Despite this, the published protocols of cmRCT either propose ITT for the primary analysis [9, 11, 12] or methods have not been stated yet [10, 13–16]. This follows the recommendation to use ITT for the analysis of pragmatic trials [5, 24, 38]. The main argument for this is that treatment refusals do happen in actual clinical practice and ITT would thus evaluate the value of offering a treatment. However, as stated earlier, refusal rates may be different in trials (due to e.g. more complex consent and recruitment procedures) and the ITT effect will not reflect that in routine practice. In cmRCTs, randomisation precedes the recruitment and consent procedures, and thus may affect results more than traditional trials. The results presented here highlight the large biases associated with ITT when estimating the treatment effect. Our results indicate the importance of evaluating in a cmRCT whether refusal rates are higher than expected in actual clinical practice (is estimating the ITT effect valuable?) and whether this refusal may be related to the outcome as this may increase the bias in estimating the treatment effect. It has been reported that the risk of early discontinuation may be correlated with the risk of outcome of interest [39]. IV deals with refusal and non-compliance in RCTs to some extent [27, 40] and has been adopted in practice [41–43]. This study found that IV indeed minimised bias due to refusal. Of the published cmRCTs, only two trials RECTAL BOOST [9] and SPIN [11] propose to use IV, and it is as a secondary analysis or if refusal rates exceed a predefined limit. The two IV methods proposed are 2SPS and 2SRI. 2SPS is a simple extension of ITT and can be obtained by dividing the ITT estimate by the average treatment refusal. Our simulations show that 2SPS is generally more biased than 2SRI. We also show that 2SPS estimates are sensitive to the data structure, when the direction of correlation between refusal and risk changed so did the trend in effect estimates. In the scenarios where 2SPS is less biased than 2SRI (negative correlation), it is only by a small amount. This is in line with the findings by Terza et al. [25], who concluded that 2SRI is the more appropriate method for nonlinear models and we therefore recommend 2SRI should be used if an IV analysis is carried out. However the effect estimate of interest in a pragmatic trial (the ITT effect in routine practice) is likely to lie somewhere in between the ITT and IV estimates and contextual information is needed to assess this. For example, if refusal rates are particularly high (> 50 %) then the IV estimate is unlikely to represent the effect of this drug in routine practice. This is in line with the findings of van der Velden et al [44], who imply both ITT and IV analyses should be carried out to analyse the results of a cmRCT. Recruitment method 1 causes a drop in statistical power when using ITT or IV methods. This drop in power can be explained by the dilution bias for ITT as the apparent treatment effect is smaller. Although the IV estimate is generally unbiased, it suffers the same drop in power due to larger standard errors resulting from using a two stage modelling technique. If refusal rates are factored into the sample size calculation (recruitment method 2) the statistical power improves as expected. There is a paucity of literature on the methods for powering a RCT in case of refusals when using IV; most literature deals with ITT [45–47]. Our study highlights the need to take into account treatment refusal into the power calculation. Methods, such as simulations, will need to be developed to adjust IV analyses (with 2SRI) for non-informative as well informative refusals. A limitation of this study is that we have only considered a homogeneous treatment effect. The results of this study are thus not generalizable to a heterogeneous treatment effect. The results of [48] can be referred to when deciding on which method to use if a heterogeneous treatment effect is thought to be present. In the case of a heterogeneous treatment effect, under certain assumptions IV estimates the complier average causal effect (CACE) [28]. The results from IV analysis should be presented following procedures laid out by Swanson and Hernán [49]. A second limitation is that the simulations presented here are based on aggregate data which may not be relevant to certain populations. It is thus important that each cluster cmRCT assesses prior to the start of trial the likely scenarios for refusal, risk of outcomes and cluster sizes and then estimate the potential impact on statistical power and level of bias. This would then guide the feasibility of the cluster cmRCT and the required sample size. Conclusions In conclusion, cluster cmRCT can be an efficient design for conducting pragmatic trials, however there are still many questions to answer: What is the impact on bias and power when multiple trials are conducted within a cohort? What would be the influence of effect modification? What will happen if there is an overlap is secondary endpoints? The question addressed in this paper is how to deal with differential refusal in the intervention arm. If the refusal in a cmRCT is similar to that in routine clinical practice, then an ITT analysis will provide a valid estimate of the ITT effect in routine practice. If refusal differs from that in routine clinical practice, an IV analysis may provide a more accurate estimate. More research is needed on how refusal rates of experimental interventions correlate with refusal rates in routine practice to help answer the question of which analysis more relevant. For now, we recommend running both analyses and providing context about the intervention. Refusals can also adversely affect the statistical power of a cluster cmRCT and should be incorporated into the power calculation. An incorrect estimate of the refusal rate can lead to the recruitment of more patients than necessary or an underpowered trial. Therefore we recommend updating the required sample size during the trial as more information about the refusal rate is gained. Finally, it is important to note that the results, recommendations and discussion raised in this paper are very much applicable to the standard cmRCT design. Appendix A In scenarios where clinician refusal is bigger than 0 the trends seen in Figs. 1, 2, 3 and 4 carry over to the changes in clinician refusal and correlation. The main result to report is that 2SRI performs similarly in all scenarios. In fact the bias only reaches over 6 % in one scenario where clinician refusal and correlation are 0.1 and zero (Fig. 1). This shows the 2SRI method is only slightly biased even as refusal rates and correlation strengths are quite large. In the same scenarios when the ICC = 0.05 the required sample sizes for each recruitment method changed, however the biases and power in each scenario were no different. The standard error of individual IV estimates (s.e.β^), which have not been reported, are always larger than that of ITT and PP. In recruitment method 1 the standard error of the IV methods increases to around 0.15, almost half the magnitude of β. In recruitment method 2 the size of the standard errors did not change from around 0.11 (reflected by the consistent power). The 95 % confidence intervals around the percentage bias calculated is always between ∓ 2 – 3 % of the bias shown. Abbreviations 2SPSTwo stage predictor substitution 2SRITwo stage residual inclusion CACEComplier average causal effect cmRCTCohort multiple randomised controlled trial CVDCardiovascular disease HSCICHealth and social care information centre ITTIntention to treat IVInstrumental variable NICENational institute for health and care excellence PPPer protocol RCTRandomised controlled trial Acknowledgements The research leading to these results was conducted as part of Work Package 2 of the GetReal consortium. This paper only reflects the personal views of the stated authors. For further information please refer to http://www.imi-getreal.eu/. Funding The work leading to these results has received support from the Innovative Medicines Initiative Joint Undertaking under grant agreement no. 115546, resources of which comprise financial contribution from the European Union Seventh Framework Programme(FP7/2007-2013) and EFPIA companies’ in kind contribution. The study was also partly supported by the University of Manchester’s Health eResearch Centre (HeRC) funded by the Medical Research Council (MRC) Grant MR/K006665/1. Availability of data and materials Code to run the simulations are available upon request to the first author, alexander.pate@manchester.ac.uk. Authors’ contributions AP ran the simulations and drafted the manuscript; MS provided consultation when running the simulations and conducted critical revision of the manuscript for intellectual content; JC provided assistance in running the simulations, consultation when running the simulations and conducted critical revision of the manuscript for intellectual content; TVS conceptualised the ideas for the work, provided consultation when running the simulations and conducted critical revision of the manuscript for intellectual content. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Ethics and consent to publish Not applicable. ==== Refs References 1. Vickers AJ Clinical trials in crisis: Four simple methodologic fixes Clin Trials 2014 11 6 615 621 10.1177/1740774514553681 25278228 2. Zwarenstein M Oxman A Why are so few randomized trials useful, and what can we do about it? J Clin Epidemiol 2006 59 11 1125 1126 10.1016/j.jclinepi.2006.05.010 17027421 3. Zwarenstein M Treweek S What kind of randomised trials do patients and clinicians need? Evid Based Med 2009 14 4 101 103 10.1136/ebm.14.4.101 19648418 4. Luce BR Kramer JM Goodman SN Connor JT Tunis S Whicher D Schwartz JS Medicine and Public Issues Annals of Internal Medicine Rethinking Randomized Clinical Trials for Comparative Effectiveness Research: The Need for Transformational Change Ann Intern Med. 2009 151 206 209 10.7326/0003-4819-151-3-200908040-00126 19567619 5. Thorpe KE Zwarenstein M Oxman AD Treweek S Furberg CD Altman DG Tunis S Bergel E Harvey I Magid DJ Chalkidou K A pragmatic-explanatory continuum indicator summary (PRECIS): a tool to help trial designers J Clin Epidemiol 2009 62 5 464 475 10.1016/j.jclinepi.2008.12.011 19348971 6. van Staa T-P Goldacre B Gulliford M Cassell J Pirmohamed M Taweel A Dalaney B Smeeth L Pragmatic randomised trials using routine electronic health records: putting them to the test BMJ 2012 344 1 10.1136/bmj.e55 22315246 7. Purgato M Barbui C Stroup S Adams C Pragmatic design in randomized controlled trials Psychol Med 2014 45 02 225 230 10.1017/S0033291714001275 25065958 8. Relton C Torgerson D O’Cathain A Nicholl J Rethinking pragmatic randomised controlled trials: introducing the ‘cohort multiple randomised controlled trial’ design BMJ 2010 340 c1066 10.1136/bmj.c1066 20304934 9. Burbach JM Verkooijen HM Intven M Kleijnen J-PJ Bosman ME Raaymakers BW van Grevenstein WM Koopman M Seravalli E van Asselen B Reerink O RandomizEd controlled trial for pre-operAtive dose-escaLation BOOST in locally advanced rectal cancer (RECTAL BOOST study): study protocol for a randomized controlled trial Trials 2015 16 1 1 8 10.1186/s13063-015-0586-4 25971836 10. Mitchell N Hewitt C Adamson J Parrott S Torgerson D Ekers D Holmes J Lester H McMillan D Richards D Spilsbury K Godfrey C Gilbody S A randomised evaluation of CollAborative care and active surveillance for Screen-Positive EldeRs with sub-threshold depression (CASPER): study protocol for a randomized controlled trial Trials 2011 12 1 225 10.1186/1745-6215-12-225 21988800 11. Kwakkenbos L Jewett LR Baron M Bartlett SJ Furst D Gottesman K Khanna D Malcarne VL Mayes MD Mouthon L Poiraudeau S Sauve M Nielson WR Poole JL Assassi S Boutron I Ells C van den Ende CH Hudson M Impens A Körner A Leite C Costa Maia A Mendelson C Pope J Steele RJ Suarez-Almazor ME Ahmed S Coronado-Montoya S Delisle VC Gholizadeh S Jang Y Levis B Milette K Mills SD Razykov I Fox RS Thombs BD The Scleroderma Patient-centered Intervention Network (SPIN) Cohort: protocol for a cohort multiple randomised controlled trial (cmRCT) design to support trials of psychosocial and rehabilitation interventions in a rare disease context BMJ Open 2013 3 8 1 11 10.1136/bmjopen-2013-003563 12. Uher R Cumby J Mackenzie LE Morash-conway J Glover JM Aylott A Propper L Abidi S Bagnell A Pavlova B Hajek T Lovas D Pajer K Gardner W Levy A Alda M A familial risk enriched cohort as a platform for testing early interventions to prevent severe mental illness BMC Psychiatry 2014 14 1 344 10.1186/s12888-014-0344-2 25439055 13. Young-Afat HMVDA van Gils CH van den Bongard HJ van Vulpen M Introducing the new ‘cohort multiple Randomised Controlled Trial’ design for evaluation of interventions for breast cancer: The UMBRELLA study Eur J Cancer 2014 50 S210 S219 14. R. Uher, “Skills for Wellness - Full Text View - ClinicalTrials.gov,” 2013. [Online]. Available: https://clinicaltrials.gov/show/NCT01980147. [Accessed: 29-Oct-2015]. 15. Bower P. ISRCTN - ISRCTN12286422: CLASSIC Proactive Telephone Coaching and Tailored Support (PROTECTS). 2014. 16. M. van Vulpen, “Randomized Trial Comparing Conventional Radiotherapy With Stereotactic Radiotherapy in Patients With Spinal Metastases - VERTICAL Study - Full Text View - ClinicalTrials.gov,” 2015. [Online]. Available: https://www.clinicaltrials.gov/ct2/show/NCT02364115. [Accessed: 29-Oct-2015]. 17. van Staa TP Dyson L McCann G Padmanabhan S Belatri R Goldacre B Cassell J Pirmohamed M Torgerson D Ronaldson S Adamson J Taweel A Delaney B Mahmood S Baracaia S Round T Fox R Hunter T Gulliford M Smeeth L The opportunities and challenges of pragmatic point-of-care randomised trials using routinely collected electronic records: Evaluations of two exemplar trials Health Technol Assess (Rockv) 2014 18 43 1 146 18. Angrist JD Imbens GW Rubin DB Association S Jun N Identification of Causal Effects Using Instrumental Variables J Am Stat Assoc 1996 91 434 444 455 10.1080/01621459.1996.10476902 19. Greenland S An introduction to instrumental variables for epidemiologists Int J Epidemiol 2000 29 4 722 729 10.1093/ije/29.4.722 10922351 20. Frangakis BCE Rubin DB Addressing complications of intention-to-treat analysis in the combined presence of all-or-none treatment-noncompliance and subsequent missing outcomes Biometrika 1999 86 2 365 379 10.1093/biomet/86.2.365 21. Dunn G Maracy M Tomenson B Estimating treatment effects from randomized clinical trials with noncompliance and loss to follow-up: the role of instrumental variable methods Stat Methods Med Res 2005 14 4 369 395 10.1191/0962280205sm403oa 16178138 22. Wu J Zhu S Yao GL Mohammed MA Marshall T Patient Factors Influencing the Prescribing of Lipid Lowering Drugs for Primary Prevention of Cardiovascular Disease in UK General Practice: A National Retrospective Cohort Study PLoS One 2013 8 7 1 10 23. Roland M Torgerson DJ Understanding controlled trials: What are pragmatic trials? BMJ 1998 316 7127 285 10.1136/bmj.316.7127.285 9472515 24. Armijo-Olivo S Warren S Magee D Intention to treat analysis, compliance, drop-outs and how to deal with missing data in clinical research: a review Phys Ther Rev 2009 14 1 36 49 10.1179/174328809X405928 25. Terza JV Basu A Rathouz PJ Two-stage residual inclusion estimation: Addressing endogeneity in health econometric modeling J Health Econ 2008 27 3 531 543 10.1016/j.jhealeco.2007.09.009 18192044 26. Efron BYB Nonparametric estimates of standard error: The jackknife, the bootstrap and other methods Biometrika 1981 68 3 589 599 10.1093/biomet/68.3.589 27. Groenwold RHH Uddin MJ Roes KCB De Boer A Martin E Gatto NM Klungel OH Instrumental variable analysis in randomized trials with non- compliance and observational pharmacoepidemiologic studies OA Epidemiology 2014 2 1 9 28. Hernán MA Robins JM Instruments for causal inference: an epidemiologist’s dream? Epidemiology 2006 17 4 360 372 10.1097/01.ede.0000222409.00878.37 16755261 29. HSCIC, “Numbers of Patients Registered at a GP Practice - April 2015,” 2015. [Online]. Available: http://www.hscic.gov.uk/searchcatalogue?productid=17788&topics=1%2fPrimary+care+services%2fGeneral+practice&sort=Most+recent&size=10&page=2#top. [Accessed: 30-Oct-2015]. 30. Townsend N, Williams J, Bhatnagar P, Wickramasinghe K, Rayner M (2014). Cardiovascular disease statistics, 2014. British Heart Foundation: London. 31. Eldridge SM Ashby D Kerry S Sample size for cluster randomized trials: Effect of coefficient of variation of cluster size and analysis method Int J Epidemiol 2006 35 1292 1300 10.1093/ije/dyl129 16943232 32. Kerry SM Martin Bland J Unequal cluster sizes for trials in English and Welsh general practice: Implications for sample size calculations Stat Med 2001 20 377 390 10.1002/1097-0258(20010215)20:3<377::AID-SIM799>3.0.CO;2-N 11180308 33. Arnold BF Hogan DR Colford JMJ Hubbard AE Simulation methods to estimate design power: an overview for applied research BMC Med Res Methodol 2011 11 94 10.1186/1471-2288-11-94 21689447 34. O’Quigley J Stare J Proportional hazards models with frailties and random effects Stat Med 2002 21 21 3219 3233 10.1002/sim.1259 12375300 35. Glidden DV Vittinghoff E Modelling clustered survival data from multicentre clinical trials Stat Med 2004 23 3 369 388 10.1002/sim.1599 14748034 36. Greenland S Robins J Pearl J Confounding and collapsibility in causal inference Stat Sci 1999 14 29 46 10.1214/ss/1009211805 37. Adamson J Cockayne S Puffer S Torgerson DJ Review of randomised trials using the post-randomised consent (Zelen’s) design Contemp Clin Trials 2006 27 305 319 10.1016/j.cct.2005.11.003 16455306 38. Hollis S Campbell F What is meant by intention to treat analysis? Survey of published randomised controlled trials BMJ 1999 319 7211 670 674 10.1136/bmj.319.7211.670 10480822 39. Snapinn SM Jiang Q Iglewicz B Informative noncompliance in endpoint trials Curr Control Trials Cardiovasc Med 2004 5 1 5 10.1186/1468-6708-5-5 15233844 40. Ye C Beyene J Browne G Thabane L Estimating treatment effects in randomised controlled trials with non-compliance: a simulation study BMJ Open 2014 4 e005362 10.1136/bmjopen-2014-005362 24939814 41. Goldsmith LP Lewis SW Dunn G Bentall RP Psychological treatments for early psychosis can be beneficial or harmful, depending on the therapeutic alliance: an instrumental variable analysis Psychol Med 2015 45 11 2365 2373 10.1017/S003329171500032X 25805118 42. Jago R Edwards MJ Sebire SJ Tomkinson K Bird EL Banfield K May T Kesten JM Cooper AR Powell JE Blair PS Effect and cost of an after-school dance programme on the physical activity of 11–12 year old girls: The Bristol Girls Dance Project, a school-based cluster randomised controlled trial Int J Behav Nutr Phys Act 2015 12 1 128 10.1186/s12966-015-0289-y 26437720 43. Halpern SD French B Small DS Saulsgiver K Harhay MO Audrain-McGovern J Loewenstein G Brennan TA Asch DA Volpp KG Randomized Trial of Four Financial-Incentive Programs for Smoking Cessation N Engl J Med 2015 372 22 2108 2117 10.1056/NEJMoa1414293 25970009 44. van der Velden JM, Verkooijen HM, Young-Afat DA, Burbach JPM, van Vulpen M, Relton C, van Gils CH, May AM, Groenwold RHH. The cohort multiple randomized controlled trial design: a valid and efficient alternative to pragmatic trials?. Int J Epidemiol. 2016. doi:10.1093/ije/dyw050. 45. Lakatos E Sample size determination in clinical trials with time-dependent rates of losses and noncompliance Control Clin Trials 1986 7 3 189 199 10.1016/0197-2456(86)90047-4 3802834 46. Sato T Sample size calculations with compliance information Stat Med 2000 19 19 2689 2697 10.1002/1097-0258(20001015)19:19<2689::AID-SIM555>3.0.CO;2-0 10986542 47. Jiang Q Snapinn S Iglewicz B Calculation of sample size in survival trials: the impact of informative noncompliance Biometrics 2004 60 3 800 806 10.1111/j.0006-341X.2004.00231.x 15339304 48. Wan F Small D Bekelman JE Mitra N Bias in estimating the causal hazard ratio when using two-stage instrumental variable methods Stat Med 2015 34 14 2235 2265 10.1002/sim.6470 25800789 49. Swanson SA Hernán MA Commentary Epidemiology 2013 24 3 370 374 10.1097/EDE.0b013e31828d0590 23549180

PMC005xxxxxx/PMC5000410.txt

==== Front BMC CancerBMC CancerBMC Cancer1471-2407BioMed Central London 273310.1186/s12885-016-2733-zResearch ArticleThe insulin-like growth factor system is modulated by exercise in breast cancer survivors: a systematic review and meta-analysis Meneses-Echávez José Francisco menesesjose77@gmail.com 1Jiménez Emilio González emigoji@ugr.es 2Río-Valle Jacqueline Schmidt jschmidt@ugr.es 2Correa-Bautista Jorge Enrique jorge.correa@urosario.edu.co 1Izquierdo Mikel mikel.izquierdo@gmail.com 3Ramírez-Vélez Robinson robin640@hotmail.comrobinson.ramirez@urosario.edu.co 11 Centro de Estudios en Medición de la Actividad Física (CEMA), Escuela de Medicina y Ciencias de la Salud, Universidad del Rosario, Bogotá, DC Colombia 2 Departamento de Enfermería. Facultad de Ciencias de la Salud, Universidad de Granada, Granada, Spain 3 Department of Health Sciences Public, University of Navarra, Pamplona, Spain 25 8 2016 25 8 2016 2016 16 1 6825 5 2015 29 7 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Insulin-like growth factors (IGF´s) play a crucial role in controlling cancer cell proliferation, differentiation and apoptosis. Exercise has been postulated as an effective intervention in improving cancer-related outcomes and survival, although its effects on IGF´s are not well understood. This meta-analysis aimed to determine the effects of exercise in modulating IGF´s system in breast cancer survivors. Methods Databases of PuMed, EMBASE, Cochrane Central Register of Controlled Trials, EMBASE, ClinicalTrials.gov, SPORTDiscus, LILACS and Scopus were systematically searched up to November 2014. Effect estimates were calculated through a random-effects model of meta-analysis according to the DerSimonian and Laird method. Heterogeneity was evaluated with the I2 test. Risk of bias and methodological quality were evaluated using the PEDro score. Results Five randomized controlled trials (n = 235) were included. Most women were post-menopausal. High-quality and low risk of bias were found (mean PEDro score = 6.2 ± 1). Exercise resulted in significant improvements on IGF-I, IGF-II, IGFBP-I, IGFBP-3, Insulin and Insulin resistance (P < 0.05). Non-significant differences were found for Glucose. Aerobic exercise improved IGF-I, IGFBP-3 and Insulin. No evidence of publication bias was detected by Egger´s test (p = 0.12). Conclusions Exercise improved IGF´s in breast cancer survivors. These findings provide novel insight regarding the molecular effects of exercise on tumoral microenvironment, apoptosis and survival in breast cancer survivors. Electronic supplementary material The online version of this article (doi:10.1186/s12885-016-2733-z) contains supplementary material, which is available to authorized users. Keywords Breast cancerExerciseInsulin-Like Growth Factor Binding ProteinsTumor Microenvironmentissue-copyright-statement© The Author(s) 2016 ==== Body Background Insulin-like growth factors (IGFs) are mitogens involved in regulating cell proliferation, differentiation, and apoptosis [1]. The IGF system includes the single-chain polypeptides IGF-I and IGF-II and six binding proteins (IGFBP-I - IGFBP-6) [2]. The IGFBP proteases may also be considered as part of the IGF system because they indirectly regulate the action of IGFs [3]. The IGF family has been linked to several metabolic and disease states, including type 1 diabetes and cancer, especially of the lung, breast, and prostate [3–6]. Both IGF-I and IGF-II exert mitogenic and antiapoptotic actions and regulate tumor cell proliferation and differentiation [3], whilst IGFBP-3 regulates the mitogenic action of IGFs and inhibits their antiapoptotic effects in breast cancer cells due to IGF- inhibitory effects on breast cancer cell growth [7]. In addition, high levels of IGFBP-3 has been associated with low concentrations of estrogen receptor (ER) or progesterone receptor and large tumor size, suggesting a poor prognosis and decreased survival in cancer patients [8, 9]. Exercise has been proposed as an effective non-pharmacological intervention to promote psychological well-being during and following cancer treatment [10–12]. However, the role of exercise in the modulation of the IGF system remains poorly understood and experimental evidence has emerged. At the same time, other researchers have proposed that exercise can be used as a mechanism to decrease IGF levels and aid in cancer prevention [13, 14]. Numerous studies have reported higher levels of circulating IGF associated with physical activity, although many other studies have reported no difference or even a decrease in IGF levels. For example, in 2009, Irwin et al. [15] reported significant reductions in IGF-I and IGFBP-3 in postmenopausal women after a 6-month walking-based intervention compared to non-exercisers. However, Sprod et al. [16] found no significant changes in IGFBP-I and IGFPB-3 after a 12-week intervention of Tai Chi Chuan in twenty-one breast cancer survivors. A limited comprehensive summary has been published that systematically reviews all literature on this topic. In light of this lack of consensus in the literature, the aim of this meta-analysis was to determine the effects of exercise in modulating the IGF system in breast cancer survivors. Methods We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Statement to conduct this review [17]. No funding was received. The PubMed, EMBASE, Cochrane Central Register of Controlled Trials, EMBASE, ClinicalTrials.gov, SPORTDiscus, LILACS and Scopus databases were systematically searched between May and November 2014 by three blinded authors (JFME, JSRV and EGJ) without restrictions on language. The reviewers were blinded to both the name of the authors and the results of the studies. The following search terms were used: ´breast cancer´ and ´exercise´ or ´physical activity´ and ´insulin´ or ´glucose´ or ´growth factors´ or ´IGF´ or ´IGFBP´. The reference lists from retrieved articles were checked to identify additional titles. The authors also examined data from previous reviews published by Ballard-Barbash et al. [18] and Löf and colleagues [19]. Moreover, two authors (MI and RR-V) searched for other relevant trials listed in journals that specialized in oncology (e.g., BMC Cancer, Breast Cancer Research, Cancer, Cancer Epidemiology, Biomarkers & Prevention, Journal of Clinical Oncology, Journal of Oncology Practice and The Lancet Oncology). Aiming to provide stronger sensitivity to the search process, the authors contacted high-profile researchers in this area to ask for other possibly relevant trials, published or unpublished. Selection criteria Two authors (JSR-V and JEC-B) independently checked all of the retrieved trials against the eligibility criteria (Table 1). The title and abstract were examined, and full-text was obtained if ambiguity regarding the eligibility of the study was noted. A third author arbitrated the consensus for eligibility (EG-J). Attempts were made to contact authors of trial reports if clarification was necessary.Table 1 Inclusion criteria considered in the systematic review Design  • Randomized controlled trial Participants  • Women with breast cancer, without restriction to a particular stage of diagnosis or treatment Intervention  • Exercise training (i.e., aerobic, resistance training, stretching exercises and Tai Chi Chuan). Outcome measures  • Insulin-like growth factor -1 (IGF-1)  • Insulin-like growth factor -2 (IGF-2)  • Insulin-like growth factor-binding protein -1 (IGFBP-1)  • Insulin-like growth factor-binding protein -3 (IGFBP-3)  • Insulin  • Insulin Resistance  • Glucose Comparisons  • Exercise training versus conventional care A cancer survivor was defined as a person who is diagnosed with cancer and survives from the time of diagnosis through the balance of his or her life [20]. Exercise interventions were defined as a form of physical activity that is planned, structured and repetitive and aims to improve fitness, performance or health [21]. Hence, we included randomized controlled trials (RCTs) that compared exercise interventions (aerobic, resistance training and stretching exercises such as Tai Chi Chuan) with a control group (conventional care) in women with breast cancer and that measured the following biomarkers: insulin-like growth factors (IGF-I and IGF-II), insulin-like growth factor-binding protein (IGFBP-I and IGFBP-3), and insulin serum levels as well as insulin resistance and glucose. This set of biomarkers was selected because they play a vital role in the tumoral microenvironment and cancer prognosis [22, 23]. Finally, we excluded trials where exercise was combined with pharmacological interventions. Methodological quality assessment The methodological quality of the studies including their risk of bias was assessed using the Physiotherapy Evidence Database (PEDro) scale [24]. The PEDro scale scores the methodological quality of randomized trials and has a maximum possible score of 10. Scores were based on all information available from both the published version and from communication with the authors. A score of 5 of 10 was set as the minimum score for inclusion in the review. The score for each included study was determined by two trained authors (JFM-E and MI). Disagreements were solved by consensus or by a third reviewer (JEC-B). We calculated the inter-observer agreement using the Kappa (k) statistic [25]; the agreement rate between authors was k = 0.91 for methodological quality assessment. Data extraction and analysis Relevant data were extracted independently by two reviewers (JFM-E and RR-V) using a standard form and a third author (JEC-B) mediated in cases of disagreement. The reviewers extracted information about the methods (i.e., design, breast cancer staging, participants and interventions) and the outcome data for the experimental and control groups. High agreement was observed between reviewers (k = 0.89). Changes in the Insulin-Like Growth Factors were reported as differences between arithmetic means pre and post exercise interventions. Statistical heterogeneity was evaluated using the I2 statistic (I2 = [(Q - df) / Q] X 100 %, where Q is the chi-square statistic and df is its degrees of freedom), which was defined according to the following categories [26]: negligible heterogeneity, 0 % – 40 %; moderate heterogeneity, 30 % – 60 %; substantial heterogeneity, 50 % – 90 %; and considerable heterogeneity, 75 % – 100 %. Other possible sources of heterogeneity were evaluated via subgroup analysis and a cumulative meta-analysis model if necessary. We conducted a random-effects model of the meta-analysis when substantial heterogeneity (I2 > 50 %) was present. Continuous outcomes were reported as the Standardized Mean Difference (SMD) with the 95 % confidence interval (95 % CI), with statistical significance set at a P < 0.05. All analyses were weighted by the inverse variance. Publication bias was examined using Egger´s test (P < 0.05) and the funnel plot based on the number of studies included (i.e. if more than 10 trials were included). Based on data availability, we conducted subgroup analysis to explore the particular effects of the modes of exercise separately. All analyses were conducted by JFM-E using Stata (Version 12.0; Stata Corp, College Station, TX). Results Characteristics of the studies included A total of five randomized controlled trials (n = 235) were included [15, 16, 27–29]. Figure 1 presents the Additional file 1 flow diagram. All groups were similar at baseline with 113 women allocated to the experimental groups and 122 women allocated to the control groups. The average publication date was 2008 ± 3.5 years. An enzyme-linked immunosorbent assay (ELISA) was used by all studies included.Fig. 1 Flowdiagram for search strategy methods. Flowdiagram is performed according to Additional file 1 Statement Methodological quality and risk of bias assessment We found a high-quality and low risk of bias (mean PEDro score = 6.2 ± 1) across studies. No study performed blinding of participants or therapists and three trials (60 %) blinded their assessors for the analyses (Table 2).Table 2 PEDro Scale scores for the included trials (n = 5) Study Random allocation Concealed allocation Groups similar at baseline Participant blinding Therapist blinding Assessor Blinding <15 % dropouts Intention to treat analysis Between-group difference reported Point estimate and variability reported Total (0 to 10) Fairey et al. [26] (2003) Y N Y N N Y Y Y Y Y 7 Irwin et al. [14] (2009) Y N Y N N Y Y Y Y Y 7 Janelsins et al. [27] (2011) Y Y Y N N N N N Y Y 5 Schmitz et al. [28] (2005) Y N Y N N Y Y N Y Y 6 Sprod et al. [15] (2012) Y Y Y N N N N Y Y Y 6 Compliance rate 100 % 40 % 100 % 0 % 0 % 60 % 60 % 60 % 100 % 100 % 100 % N No, Y Yes, PEDro Physiotherapy Evidence Database Characteristics of the participants Most women were postmenopausal, with an average age of 48 ± 3.2 years (range 48-59 years), and were classified with tumor stages I-IIIA after anti-cancer treatment. Chemotherapy was the most common treatment (n = 176) followed by radiotherapy (n = 124), and 57 participants received hormonal therapy, with the majority of participants receiving tamoxifen. Finally, 133 women received mastectomy and 104 were treated through lumpectomy. Characteristics of the exercise interventions The interventions had a mean length of 22.2 ± 13.5 weeks with an average of 2.8 ± 0.5 sessions per week. The longest exercise intervention length was 12 months reported by Schmitz et al. [29]. The mean session duration was 73 ± 9.6 min. Exercise interventions included aerobic exercise (i.e., walking and stationary cycling) in 2 trials (40 %) [15, 26], resistance training (i.e., strength training) was implemented by Schmitz et al. [29] and Tai Chi Chuan exercises were implemented in two trials [16, 27]. The training intensity varied considerably among studies, ranging from 50 % to 90 % of the maximum heart rate. The adherence rate was 83.7 ± 8.7 %. No major adverse effects were reported. Finally, all studies reported pre-exercise screening before high intensity physical training. Table 3 summarizes the characteristics of the included studies.Table 3 Characteristics of the five randomized controlled trials included in the systematic review and meta-analysis Study ID Design Stage of Disease Participants Interventions Outcome measures Fairey et al. [26] (2003) RCT Breast Cancer Stage I –IIIB Characteristics of cancer treatment = Women who had completed surgery, radiotherapy, and/or chemotherapy. N = 53 Female = 53 Exp (n = 25) Age (yr) = 59 (5) Con (n = 28) Age (yr) = 58 (6) Exp = Aerobic exercise. Length = 15 weeks. Duration = Exercise began at 15 min for weeks 1–3, and then systematically increased by 5 min every 3 weeks thereafter to 35 min for weeks 13–15. Frequency = 3 ses/wk. Intensity = 70 %-75 %. Con = Conventional care. Godin Leisure-Time Exercise Questionnaire, fasting blood. Irwin et al. [14] (2009) RCT Breast Cancer Stage 0-IIIA Characteristics of cancer treatment = Women who had completed surgery, radiotherapy, and/or chemotherapy. N = 68 Female = 68 Exp (n = 36) Age (yr) = 56.4 (9.5) Con (n = 32) Age (yr) = 55.6 (7.7) Exp = A combined supervised training program at a local health club and a home aerobic training program. Length = 24 weeks. Duration = 129 min/wk. Frequency = Participants exercised three times per week and were instructed to exercise two days/ week on their own, either at the health club or home. Intensity = Moderate-intensity. Con = Conventional care. Duration = 45 min/wk. Ainsworth’s Compendium of Physical Activities and fasting blood. Janelsins et al. [27] (2011) RCT Breast Cancer Stage 0-IIIb Characteristics of cancer treatment = Surgery with axillary lymphadenectomy and both post-surgery radiotherapy and chemotherapy. Female = 19 Exp (n = 9) Age (yr) = 54.33 (10.64) Con (n = 10) Age (yr) = 52.70 (6.67) Exp = Tai Chi Chuan (TCC). Length = 12 weeks. Duration = 60 min/session Frequency = 3 session/week. Intensity = Moderate or vigorous. Con = Psychosocial therapy Bioelectrical impedance tests, Fasting blood. Schmitz et al. [28] (2005) RCT Breast Cancer Stage I-III Characteristics of cancer treatment = Radiation treatment, chemotherapy, axillary dissection, and hormonal therapy N = 85 Female = 85 Exp (n = 33) Age (yr) = 53.3 (8.7) Con (n = 36) Age (yr) = 52.8 (7.6) Exp = weight training Length = 12 month (26 weeks) Duration = 60 min each session Frequency = twice-weekly Intensity = moderate Con = conventional care 0-6 month weight training 7-12 month Body weight, height, body fat, lean mass, body fat %, and waist circumference, as well as fasting glucose, insulin, insulin resistance, insulin-like growth factor-I (IGF-I), IGF-II, and IGF-binding protein-1, IGFBP-2, and IGFBP-3. Sprod et al. [16] (2012) RCT Breast Cancer Stage 0–IIIb Characteristics of cancer treatment = Surgery (lymphadenectomy and mastectomy) post-surgery radiotherapy and chemotherapy. N = 19 Female = 19 Exp (n = 9) Age (yr) = 54.33 (3.55) Con (n = 10) Age (yr) = 52.70 (2.11) Exp = Tai chi chuan exercise. Length = 12 weeks. Duration = 60 min/ses. Frequency = 3 ses/wk. Intensity = low to moderate Con = Standard support therapy control (SST) Cytokine levels and fasting blood. RCT Randomized Controlled Trial, Exp Experimental Group, Con Control Group Data are presented as mean (SD) Effects of exercise on insulin-like growth factors (IGFs) and their binding proteins (IGFBP-I and IGFBP-3) Changes in the circulating levels of IGF-I after exercise training were evaluated in five studies [15, 16, 26–29]. The pooled SMD was -0.74 (95 % CI -1.14 to -0.34; I2 = 52.8 %), indicating a moderate reduction in IGF-I following exercise (Fig. 2).Fig. 2 Meta-analysis for the effect estimate of exercise on circulating levels of IGF-I. Standardized Mean Difference (SMD) was calculated for the Random effects model of meta-analysis Similar improvements were obtained for IGF-II (SMD = -0.96, 95 % CI -1.33 to -0.59; I2 = 91.4 %) [26, 28] (Fig. 3), which was measured in two studies [26, 28]. These estimates were obtained using a random-effects model. A meta-regression analysis to explore dose–response relationships was not conducted due to the limited number of studies included.Fig. 3 Meta-analysis for the effect estimate of exercise on circulating levels of IGF-II. Standardized Mean Difference (SMD) was calculated for the Random effects model of meta-analysis Based on data from four articles [15, 26–28], the pooled estimates revealed that exercise improved the serum levels of Insulin-like growth factor-binding protein-I (IGFBP-I) (SMD = 0.51, 95 % CI 0.20 to 0.82; I2 = 62 %) (Fig. 4).Fig. 4 Meta-analysis for the effect estimate of exercise on circulating levels of IGFBP-I. Standardized Mean Difference (SMD) was calculated for the Random effects model of meta-analysis All of the studies included [15, 16, 26–28] evaluated the serum concentrations of IGFBP-3 and demonstrated that exercise training significantly increased the serum levels of this biomarker in women with breast cancer (SMD = 0.54, 95 % CI 0.27 – 0.80; I2 = 84.2 %) (Fig. 5).Fig. 5 Meta-analysis for the effect estimate of exercise on circulating levels of IGFBP-3. Standardized Mean Difference (SMD) was calculated for the Random effects model of meta-analysis In addition, exercise interventions resulted in significant differences in the levels of insulin (SMD = 0.94, 95 % CI 0.70 – 1.19; I2 = 93.8 %) (Fig. 6) and insulin resistance (SMD = -0.35, 95 % CI -0.70 to -0.009; I2 = 0 %). Non-significant differences were obtained for glucose levels (SMD = -0.16, 95 % CI -0.43 to 0.10; I2 = 0 %).Fig. 6 Meta-analysis for the effect estimate of exercise on circulating levels of insulin. Standardized Mean Difference (SMD) was calculated for the Random effects model of meta-analysis Subgroup analysis by mode of exercise Regarding the subgroup analyses, aerobic exercise improved the serum concentrations of IGF-I, IGFBP-3 and insulin. Aerobic exercise analysis for IGFBP-I was not possible because only Fairey et al. [26] evaluated this marker. Tai-Chi training resulted in significant benefits for insulin levels. Tai-Chi also improved the serum concentrations of IGF-I, IGFBP-I and IGFBP-3, although these effects did not reach significance. Resistance training analysis was not conducted because only Schmitz et al. [29] evaluated this mode of exercise. Subgroup analysis for IGF-II was not possible because the two studies that measured this marker implemented different exercise modes [26, 28]. Figure 7 displays the subgroup analysis according to mode of exercise for IGFBP-3. Table 4 describes the effect estimates for the subgroup analyses undertaken in the meta-analysis.Fig. 7 Meta-analysis for subgroup analysis by mode of exercise. Standardized Mean Difference (SMD) was calculated for the Random effects model of meta-analysis Publication bias No evidence of publication bias was detected by Egger´s test (P = 0.12); a funnel plot was not built due to the limited number of studies included in the pooled analysis. Discussion The most remarkable finding from this meta-analysis was that exercise training improved the serum levels of IGF-I, IGF-II, IGFBP-I and IGFBP-3 in breast cancer survivors after successful anticancer treatment. Similar conclusions have been reported in previous experimental studies [15, 26, 28]. Moreover, it is important to highlight that this is the first meta-analysis that has summarized the effectiveness of exercise training in modulating the IGF system in breast cancer survivors because a previous systematic review regarding exercise and blood biomarkers in breast cancer survivors was published by Löf and colleagues [19], but the authors did not undertake data synthesis analysis. The mitogenic and antiapoptotic effects of IGF-1 are related to a poorer prognosis in breast cancer [30] and increased all-cause mortality [31]. The type of exercise did not appear to affect any putative association; however, it is probable that different exercise modalities cause different responses in IGF-1. Our pooled analysis demonstrated that exercise reduced IGF-I concentrations in women with breast cancer after successful treatment. These findings are consistent with those by some studies included in our meta-analysis, such as the trial published by Fairey et al. [26], in which a 15-week aerobic exercise intervention resulted in significant decreases in IGF-I levels (10.9 %) in fifty-three postmenopausal breast cancer survivors. Data from a Yale study [15] also confirm our findings; in this study, the authors found an 8.9 % significant reduction in IGF -I in an experimental group composed of 38 breast cancer survivors that completed 150 min/wk of moderate intensity aerobic exercise during 5 weeks compared to a control intervention (i.e., instructions for patients to maintain their current physical activity level). Another consideration in assessing studies using an exercise intervention is the timing of blood sampling in relation to exercise. Most studies that have demonstrated a post-exercise increase in IGF-1 found an immediate post-exercise spike followed by a gradual return to baseline or lower than baseline IGF -1 levels over the next 30 min to several hours [13]. It has been demonstrated that IGFBP-3 restricts IGF-1 availability and biological activity [32] and thus, low levels of IGFBP-3 have been associated with an increased risk of breast cancer [33] and a poorer prognosis and have been postulated as predictors of distant recurrence of breast carcinoma in postmenopausal women [1, 34]. We found that exercise training increased IGFBP-3 serum levels in breast cancer survivors, although high statistical heterogeneity was observed in the overall effect estimate (I2 = 84.2 %). We obtained similar results for the aerobic exercise subgroup analysis. These findings are consistent with those published by Fairey et al. [26] and Irwin et al. [15] from the Yale study described above. In addition, when adjusted by exercise mode in the subgroup analysis, we found that Tai Chi was an effective intervention in increasing IGFBP-3 serum levels in breast cancer survivors, although statistical significance was not reached. Similar results were published by Janelsins et al. [28] in a randomized controlled trial in 19 breast cancer survivors, where a 12-week exercise intervention of Tai Chi increased IGFBP-3 serum levels compared to non -exercise. Conversely, non-significant changes in IGFBP-3 were observed by Sprod et al. [16] in a more recent study with a similar intervention using Tai Chi. Interestingly, the authors reported an association between changes in IGFBP-3 and physical functioning, suggesting a link between changes in IGF binding proteins and some domains of quality of life in breast cancer survivors, although these associations warrant additional research. However, several studies that have reported a change in IGFBP-3 following an acute exercise challenge usually found a pattern similar to that found for total IGF-1 [13]. Regarding the secondary outcomes of this meta-analysis, our analyses showed that exercise produces significant increases in insulin and significant decreases in the insulin resistance of breast cancer survivors; reductions in the glucose levels did not reach statistical significance. Subgroup analysis by mode of exercise was limited for insulin and insulin resistance due to the number of studies included. Similar to our results, Sprod et al. [16] reported slight increases in insulin levels after a Tai Chi intervention. Nonetheless, other studies have reported mixed findings. Schmitz et al. [29] found no changes in insulin or glucose after weight training exercise in 85 breast cancer survivors; Ligibel et al. [35] detected significant reductions in insulin levels after a twice-weekly resistance training intervention for 16 weeks in breast cancer survivors. Lastly, Irwin et al. [15] stated that the lack of changes in insulin and glucose levels can be affected by weight status at baseline (i.e., obese breast cancer survivors have higher insulin levels than participants with normal or lower weight), suggesting that heavier participants can benefit more from exercise compared to leaner participants with respect to changes in glycemic control. In this sense, several biologically plausible mechanisms could explain the effects of exercise in modulating the IGF and IGFBP systems. It is widely known that exercise has the potential to reduce both hepatic and muscle insulin resistance and to increase glucose availability due to insulin signaling pathways, improvements in capillary density leading to a better delivery of muscle glucose, increases in glucose protein transporters and effects on mRNA [36]. These conditions decrease the insulin concentration due to lower concentrations of IGFs via insulin-mediated changes in IGFBP concentrations [14]. However, further research is needed to confirm these mechanisms, especially in breast cancer survivors during and after anticancer treatment regimens, and gain insight regarding the benefits that exercise and multidimensional behavioral change interventions can provide on cancer treatment related outcomes and survival, moving from preventive strategies toward patients facing cancer. Only one study examined the effects of resistance training alone, and this method was also beneficial [29]. The effects of resistance exercise have not been addressed by the American Cancer Society but have been examined recently in people undergoing cancer treatment [37]. However, the present meta-analysis indicates that further evidence regarding the effects of resistance training during and after anticancer treatment. Besides, to understand the possible mechanisms, more information is required regarding the effects of initial chemotherapy and radiation therapy on muscle satellite (progenitor) cells that proliferate in response to resistance exercise [10, 11]. Strengths and limitations To our knowledge, this is the first meta-analysis that evaluates the changes on insulin-like growth factors and their binding proteins after exercise training in breast cancer survivors. Our results provide novel insight regarding the role of exercise as a non-pharmacological and non-cytotoxic effective intervention in modulating the tumoral microenvironment as well as in the management of cancer treatment-related side effects (i.e., fatigue, depression and impairments of quality of life). In addition, there were numerous methodological limitations that impacted the generalizability of studies, including a lack of adjustment for confounding factors (e.g., plasma volume, participant age or body composition) and a lack of consideration of effect modification [13]. Furthermore, our findings have crucial implications on cancer recurrence and disease free survival rates. In addition, all studies included exhibited moderate to high methodological quality and low risk of bias, which is an important issue in terms of external validity. Nevertheless, some limitations with regard to our study exist that are important to state. The overall effects estimates were increased due to different modes of exercise across the studies included, although such differences were approached through subgroup analysis according to the mode of exercise. High statistical heterogeneity levels were detected for most of the effect estimates, which suggests some caution when interpreting our findings. This evidence of heterogeneity was counteracted by a random effects model of analysis and can be explained by differences in some characteristics of exercise such as intensity, duration, intervention length, follow up periods and adherence rates across studies. Furthermore, dose–response relationships were not explored due to the number of studies included, and further trials might provide specific details regarding training intensity, duration and length of exercise interventions in order to strength the consensus in this field. Finally, considering that all studies involved women who completed their therapeutic treatments, it is important that further studies include patients during the active treatment stages to elucidate the effects of exercise on IGFs in patients undergoing anti-cancer treatment. Conclusions Exercise training is an effective and safe intervention for the improvement of serum levels of the IGF system and its binding proteins (IGFBP-I and IGFBP3) as well as for insulin and glucose control in breast cancer survivors, suggesting a beneficial role of exercise for the tumoral microenvironment and breast cancer recurrence and disease free survival rates in women with breast malignancies. Important components for future research have been identified that should address many of the limitations found in the reviewed studies, which would advance this area of research by answering questions on exercise, IGFs, and health, an area that is growing in interest and importance. High-quality studies are necessary to determine an optimal exercise program and to assess the clinical relevance of the results of available research. Additional file: Additional file 1: PRISMA 2009 Checklist. (DOCX 34 kb) Abbreviations CIConfidence interval ELISAEnzyme-linked immunosorbent assay IGFInsulin-like growth factors IGFBPInsulin-like growth factor binding protein MDMean difference PEDroPhysiotherapy evidence database RCTRandomized controlled trials SDStandard deviation Acknowledgements The authors would like to acknowledge Universidad Santo Tomás, Bogotá for the financial support to the GICAEDS Group (Project: Práctica del autoexamen de seno y los conocimientos, factores de riesgo y estilos de vida relacionados al cáncer de mama en mujeres jóvenes de la USTA FODEIN, Number: 4110060001-008). Authors’ contributions JFM-E and RR-V participated in the study design, acquisition of the data, data analysis and interpretation and drafting the manuscript. MI, EG-J, JSR-V and JEC-B participated in the data analysis and interpretation and drafting the manuscript. All authors have read and approved the manuscript. Competing interests The authors declare that they have no competing interests. Ethics approval and consent to participate This systematic review and meta-analysis included experimental studies that followed the provisions stated in the Declaration of Helsinki and were approved by the Ethics Committee. Two investigators (JFM-E and RR-V) confirmed that the studies included had ethics committee approval and that the participants signed consent forms. ==== Refs References 1. Jones JI Clemmons DR Insulin-like growth factors and their binding proteins: biological actions Endocr Rev 1995 16 3 34 7758431 2. Yu H Rohan T Role of the insulin-like growth factor family in cancer development and progression J Natl Cancer Inst 2000 92 1472 1489 10.1093/jnci/92.18.1472 10995803 3. Hankinson SE Willett WC Colditz GA Hunter DJ Michaud DS Deroo B Circulating concentrations of insulin-like growth factor-1 and risk of breast cancer Lancet 1998 351 1393 1396 10.1016/S0140-6736(97)10384-1 9593409 4. Yu H Spitz MR Mistry J Gu J Hong WK Wu X Plasma levels of insulin-like growth factor-I and lung cancer risk: a case–control study J Natl Cancer Inst 1999 91 151 156 10.1093/jnci/91.2.151 9923856 5. Chan JM Stampfer MJ Giovannucci E Gann PH Ma J Wilkinson P Plasma insulin-like growth factor-I and prostate cancer risk: a prospective study Science 1998 279 563 566 10.1126/science.279.5350.563 9438850 6. Ma J Pollak MN Giovannucci E Chan JM Tao Y Hennekens CH Prospective study of colorectal cancer risk in men and plasma levels of insulin-like growth factor (IGF)-I and IGF-binding protein-3 J Natl Cancer Inst 1999 91 620 625 10.1093/jnci/91.7.620 10203281 7. Nickerson T Huynh H Pollak M Insulin-like growth factor binding protein-3 induces apoptosis in MCF7 breast cancer cells Biochem Biophys Res Commun 1997 237 690 693 10.1006/bbrc.1997.7089 9299428 8. Yu H Levesque MA Khosravi MJ Papanastasiou-Diamandi A Clark GM Diamandis EP Insulin-like growth factor-binding protein-3 and breast cancer survival Int J Cancer 1998 79 624 628 10.1002/(SICI)1097-0215(19981218)79:6<624::AID-IJC12>3.0.CO;2-9 9842972 9. Baxter RC IGF binding proteins in cancer: mechanistic and clinical insights Nat Rev Cancer 2014 14 5 329 341 10.1038/nrc3720 24722429 10. Meneses-Echávez JF González-Jiménez E Ramírez-Vélez R Effects of supervised exercise on cancer-related fatigue in breast cancer survivors: a systematic review and meta-analysis BMC Cancer 2015 15 77 10.1186/s12885-015-1069-4 25885168 11. Meneses-Echávez JF González-Jiménez E Ramírez-Vélez R Supervised exercise reduces cancer-related fatigue: a systematic review J Physiother 2015 61 1 3 9 10.1016/j.jphys.2014.08.019 25511250 12. Schmid D Leitzmann MF Association between physical activity and mortality among breast cancer and colorectal cancer survivors: a systematic review and meta-analysis Ann Oncol 2014 25 7 1293 1311 10.1093/annonc/mdu012 24644304 13. Orenstein MR Friedenreich CM Review of physical activity and the IGF family JPAH 2004 1 291 320 10.1123/jpah.1.4.291 14. Meneses-Echávez JF Ramírez-Vélez R Gonzalez-Jimenez E Sanchez-Perez MJ Molina-Montes E Physical activity and Insulin-like growth factors in breast cancer survivors: A metaanalysis Eur J Surg Oncol 2014 40 11 S65 10.1016/j.ejso.2014.08.141 15. Irwin ML Varma K Alvarez-Reeves M Cadmus L Wiley A Chung GG Randomized controlled trial of aerobic exercise on insulin and insulin-like growth factors in breast cancer survivors: the Yale exercise and survivorship study Cancer Epidemiol Biomarkers Prev 2009 18 306 313 10.1158/1055-9965.EPI-08-0531 19124513 16. Sprod LK Janelsins MC Palesh OG Carroll JK Heckler CE Peppone LJ Health-related quality of life and biomarkers in breast cancer survivors participating in tai chi chuan J Cancer Surviv 2012 6 2 146 154 10.1007/s11764-011-0205-7 22160628 17. Liberati A Altman DG Tetzlaff J Mulrow C Gøtzsche PC Ioannidis JP The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration BMJ 2009 339 b2700 10.1136/bmj.b2700 19622552 18. Ballard-Barbash R Friedenreich CM Courneya KS Siddiqi SM McTiernan A Alfano CM Physical activity, biomarkers, and disease outcomes in cancer survivors: a systematic review J Natl Cancer Inst 2012 104 11 815 840 10.1093/jnci/djs207 22570317 19. Löf M Bergström K Weiderpass E Physical activity and biomarkers in breast cancer survivors: a systematic review Maturitas 2012 73 2 134 142 10.1016/j.maturitas.2012.07.002 22840658 20. National Cancer Institute. Office of Cancer Survivorship, Research Definitions. 2014. http://cancercontrol.cancer.gov/ocs/statistics/definitions.html. [Accessed 21 May 2014]. 21. Caspersen CJ Powell KE Christenson GM Physical activity, exercise, and physical fitness: definitions and distinctions for health-related research Public Health Rep 1985 100 126 131 3920711 22. Kashyap MK. Role of insulin-like growth factor-binding proteins in the pathophysiology and tumorigenesis of gastroesophageal cancers. Tumour Biol. 2015;36(11):8247–57. 23. Moschos SJ Mantzoros CS The role of the IGF system in cancer: from basic to clinical studies and clinical applications Oncology 2002 63 4 317 332 10.1159/000066230 12417786 24. De Morton N The PEDro scale is a valid measure of the methodological quality of clinical trials: a demographic study Aust J Physiother 2009 55 2 129 133 10.1016/S0004-9514(09)70043-1 19463084 25. Viera AJ Garrett JM Understanding inter-observer agreement: the kappa statistic Fam Med 2005 37 5 360 363 15883903 26. Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0. The Cochrane Collaboration. 2011. Retrieved from: http://handbook.cochrane.org/, [Accessed 21 Feb 2015]. 27. Fairey AS Courneya KS Field CJ Bell GJ Jones LW Mackey JR Effects of exercise training on fasting insulin, insulin resistance, insulin-like growth factors, and insulin-like growth factor binding proteins in postmenopausal breast cancer survivors: a randomized controlled trial Cancer Epidemiol Biomarkers Prev 2003 12 8 721 727 12917202 28. Janelsins MC Davis PG Wideman L Katula JA Sprod LK Peppone LJ Effects of Tai Chi Chuan on insulin and cytokine levels in a randomized controlled pilot study on breast cancer survivors Clin Breast Cancer 2011 11 3 161 170 10.1016/j.clbc.2011.03.013 21665136 29. Schmitz KH Ahmed RL Hannan PJ Yee D Safety and efficacy of weight training in recent breast cancer survivors to alter body composition, insulin, and insulin-like growth factor axis proteins Cancer Epidemiol Biomarkers Prev 2005 14 7 1672 1680 10.1158/1055-9965.EPI-04-0736 16030100 30. Hartog H Boezen HM de Jong MM Schaapveld M Wesseling J van der Graaf WTA Prognostic value of insulin-like growth factor 1 and insulin-like growth factor binding protein 3 blood levels in breast cancer Breast 2013 22 1155 1160 10.1016/j.breast.2013.07.038 23968866 31. Duggan C Wang C-Y Neuhouser ML Xiao L Smith AW Reding KW Associations of insulin-like growth factor and insulin-like growth factor binding protein-3 with mortality in women with breast cancer Int J Cancer 2013 132 1191 1200 10.1002/ijc.27753 22847383 32. Christopoulos PF Msaouel P Koutsilieris M The role of the insulin-like growth factor-1 system in breast cancer Mol Cancer 2015 14 1 43 10.1186/s12943-015-0291-7 25743390 33. Lee AV Lee D Manni A Role of the IGF system in breast cancer proliferation and progression Endocrinology of Breast Cancer 1999 Totowa, NJ Humana Press 187 203 34. Goodwin PJ Ennis M Pritchard KI Trudeau ME Koo J Hartwick W Hoffman B Hood N Insulin-like growth factor binding proteins 1 and 3 and breast cancer outcomes Breast Cancer Res Treat 2002 74 65 76 10.1023/A:1016075709022 12150454 35. Ligibel JA Campbell N Partridge A Chen W Salinardi T Chen H Impact of a mixed strength and endurance exercise intervention on insulin levels in breast cancer survivors J Clin Oncol 2008 26 907 912 10.1200/JCO.2007.12.7357 18281663 36. Ivy JL Zderic TW Fogt DL Prevention and treatment on non-insulin-dependent diabetes mellitus Exerc Sport Sci Rev 1999 27 40 1 35 10791012 37. American Cancer Society Cancer facts & figures 2012 Atlanta American Cancer Society

PMC005xxxxxx/PMC5000411.txt

==== Front Zoological LettZoological LettZoological Letters2056-306XBioMed Central London 5210.1186/s40851-016-0052-5Research ArticleFunctional morphology of giant mole crab larvae: a possible case of defensive enrollment Rudolf Nicole R. nicole.rudolf.nr@googlemail.com Haug Carolin carolin.haug@palaeo-evo-devo.info Haug Joachim T. joachim.haug@palaeo-evo-devo.info Ludwig-Maximilians-Universität München, Fakultät für Biologie, Biozentrum, Großhaderner Str. 2, 82152 Planegg-Martinsried, Germany 26 8 2016 26 8 2016 2016 2 1 1718 2 2016 10 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Mole crabs (Hippidae) are morphologically distinct animals within Meiura, the “short-tailed” crustaceans. More precisely, Hippidae is an ingroup of Anomala, the group which includes squat lobsters, hermit crabs, and numerous “false” crabs. Within Meiura, Anomala is the sister group to Brachyura, which includes all true crabs. Most meiuran crustaceans develop through two specific larval phases. The first, pelagic one is the zoea phase, which is followed by the transitory megalopa phase (only one stage). Zoea larvae are rather small, usually having a total size of only a few millimeters. Zoea larvae of some hippidan species grow significantly larger, up to 15 mm in size, making them the largest known zoea larvae of all anomalan, and probably all meiuran, crustaceans. It has been suggested that such giant larvae may be adapted to a specific defensive strategy; i.e., enrollment. However, to date such giant larvae represent a rarity. Methods Eight specimens of large-sized hippidan larvae from museum collections were photographed with a Canon Rebel T3i digital camera under cross-polarized light. Additionally, one of the specimens was documented with a Keyence BZ-9000 fluorescence microscope. The specimen was subsequently dissected to document all appendages in detail. UV light (377 nm) was used for illumination, consistent with the specimen’s autofluorescence capacities. For high-resolution images, composite imaging was applied. Results All specimens differ in important aspects from all other known hippidan zoea larvae, and thus probably represent either previously unreported larvae or stages of known species, or larvae of unknown species. The sixth pleon segment articulates off the telson, a condition not previously reported in hippidan zoea larvae, but only for the next larva phase (megalopa). The larvae described here thus most likely represent the ultimate pelagic larval stages, or rare cases of ‘early megalopae’. The morphological features indicate that giant hippidan larvae perform defensive enrollment. Conclusions Our investigation indicates a larger morphological diversity of hippidan larvae than was known previously. Moreover, their assumed functional morphology, similar to the condition in certain stomatopod larvae, indicates a not yet directly observable behavior by these larvae, namely defensive enrollment. In a wider context, we are only just beginning to understand the ecological roles of many crustacean larvae. Electronic supplementary material The online version of this article (doi:10.1186/s40851-016-0052-5) contains supplementary material, which is available to authorized users. Keywords Giant larvaZoeaHippidaeDefensive behaviorMuseum materialhttp://dx.doi.org/10.13039/501100001659Deutsche ForschungsgemeinschaftHa 6300/3-1Haug Joachim T. Bavarian Equal Opportunities Sponsorship of the LMUEU SynthesysDK-TAF-2591FR-TAF-5181FR-TAF-5175Haug Carolin Haug Joachim T. issue-copyright-statement© The Author(s) 2016 ==== Body Background Within the diverse group of Eucrustacea, Hippidae is a rather small ingroup with a distinct adult morphology; its representatives are known as sand or mole crabs [1]. Hippidae is an ingroup of Anomala (often also termed Anomura), the group uniting hermit crabs, false crabs, and squat lobsters. Anomala and Brachyura (true crabs) together form Meiura. Within Hippidae, three species groups are generally differentiated: Emerita Scopoli, 1777, Hippa Fabricius, 1787 and Mastigochirus Miers, 1878 [2]. As with other meiurans, representatives of Hippidae develop through two distinct larval phases: a zoea phase with 3–6 pelagic zoea stages, followed by a critical metamorphic molt into a single megalopa stage representing a still-swimming transitory form [3]. The juvenile and adult stages have a benthic mode of life in intertidal and upper subtidal sandy marine environments [4]. Representatives of Hippidae are special among Meiura in that some of their zoea larvae may achieve impressive sizes. These can reach shield lengths of over 6 mm and, together with the long and slender pleon, may be more than 15 mm long when outstretched [4], whereas most meiuran megalopae are significantly smaller. In fact, these probably represent the largest zoea larvae of all anomalan crustaceans, and possibly all meiurans. Martin and Ormsby ([4], their Fig. 1b) depicted one such super-sized specimen positioned with a strongly anteriorly flexed pleon. They furthermore pointed out how well the “opercular-like” telson (term from [4]) fits the ventral shape of the shield. While not further discussed in this original work, the function of this tight fit seems most likely to be a specific defensive strategy; more precisely, these larvae appear able to perform defensive enrollment.Fig. 1 a–b Commonly known species that exhibit defensive enrollment. A 1–A 3 Autofluorescence images of Chiton spec. (Polyplacophora). A 1 Ventral view. A 2 Lateral view. A 3 Dorsal view. B 1–B 4 Composite images under cross-polarized light of a mantis shrimp larva (Stomatopoda, Erichthus-type, see [7]). B 1 Ventral view. B 2 Lateral view. B 3 Posterior view. B 4 Frontal view Enrollment is a defensive mechanism which apparently evolved several times independently within Metazoa, often combined with morphological specializations, such as hard plates or large sclerotized spines (e.g., [5–7]). In enrollment, the body is strongly curved ventrally, forming a nearly perfect ball, and the anterior and posterior end lie adjacent to each other. As a result, sclerotized or hardened dorsal structures protect the softer, ventral side of the body and all appendages. Within vertebrates, armadillos (Xenarthra, Mammalia) are able to bend their body to such an extent that they form a ball (e.g. [8]); their “armor” of dorsal overlapping plates composed of bone with a covering of keratin [9], provides protection in this position. Among mollusks, polyplacophorans (chitons) roll up their bodies to the ventral side when detached from the substrate. In this posture, their dorsal shell plates protect the broad and fleshy foot ([10, 11]; see also Fig. 1a). In arthropods, enrollment of the body as a protective mechanism against predators and other threats is widespread and primarily known from terrestrial arthropods, e.g., pill bugs and pill millipedes. However, some extinct marine species, e.g., trilobites, also performed enrollment [12–15]. Here again, as described for the other groups, the body is strongly curved ventrally and the tergites (dorsal sclerotisations of the segments) protect the softer ventral side of the body and all the appendages. More recently it has been reported that certain larval representatives of mantis shrimps (Stomatopoda, Eucrustacea) are also able to tightly enroll their bodies. The pleon is bent forward, constituting a sclerotized protection for the entire body with no major gaps ([7]; see also Fig. 1b). While the description in Martin and Ormsby [4] indicated the possibility of morphological adaptations for enrollment in hippidan larvae, this appears not to have been investigated further. In the present report, we present new specimens of giant hippidan larvae and provide a detailed description of the general morphology using modern imaging techniques. We discuss morphological details which support the interpretation that these larvae can indeed perform defensive enrollment. We furthermore document an unexpected morphological diversity among giant hippidan larvae. Methods Materials Eight hippidan larval specimens were the basis for the present investigation. Six of the specimens came from the zoological collections of the Natural History Museum of Denmark, Copenhagen (ZMUC), registered under the numbers ZMUC-CRU-8679 to 8684. These specimens were collected during the Dana expeditions (1921–22 and 1928–30; Schmidt 1926, 1931; Broch 1936). One specimen came from the crustacean collections of the Senckenberg Naturmuseum Frankfurt (Mu_267), and one from the Muséum national d’Histoire naturelle Paris (MNHN-IU-2014-5468). All are currently stored in 70 % ethanol, probably after previous fixation in formalin. For ventral and dorsal documentation (always within the storage liquid), some specimens were carefully outstretched and fixed with a cover slip. For large specimens, posterior and anterior ends were each fixed with separate cover slips. In other orientations, specimens were either propped against glass or metal objects, or placed into depressions. Specimens in unusual positions were not altered, but kept in this specific position. A single specimen (ZMUC-CRU-8679) was dissected directly in 70 % ethanol using a dissection microscope. Documentation All eight specimens were photographed using a Canon Rebel T3i camera with a MP-E 65 mm macro lens. Light was provided by a Canon Macro Twin Flash MT 24EX or a MeiKe FC 100 LED ring light. Light sources were equipped with polarization filters. A cross-polarized filter was placed in front of the lens. Cross-polarized light reduces reflections and enhances colour contrast (e.g., [16] and references therein). Additionally, one of the eight specimens (ZMUC-CRU-8679) was documented in 70 % ethanol using a Keyence BZ-9000 fluorescence microscope with either a 2×, 4× or 10× objective (resulting in approximately 20×, 40×, and 100× magnification, respectively; in a few cases the zoom function of the camera was also employed) depending on the different sizes of the body parts. UV light (377 nm) was used for illumination, using the autofluorescence capacities of the specimens (e.g. [17]). For high-resolution images, composite imaging was applied [18, 19]. Image processing Image stacks were fused with the computer software CombineZP into sharp images. Adobe Photoshop CS3 was used to merge different sharp image details resulting in large panorama images. Finally, images were edited in Adobe Photoshop CS6 (optimization of the histogram and sharpness, manual removing of dirt particles etc. e.g. [7]). Drawings For better comparison, the different telson shapes of the specimens ZMUC-CRU-8679, 8682, and 8683 were drawn in Adobe Illustrator CS 3. Presentation The description is provided as a descriptive matrix (Additional file 1) [20]. This allows a more direct comparison of corresponding structures, which may facilitate future detailed descriptions of other larvae. Terminology Most terms applied are standard crustacean terms (e.g. [21, 22]). However, we have sought to keep terminology neutral to the extent possible, in the interests of allowing comparisons across a wider (arthropod) range. Special terminology of malacostracan or decapod-type is provided in brackets. Results In the following, we describe one of largest specimens (specimen A) in detail. Furthermore, we provide a morphological description of comparable features of the additional seven specimens (specimen B–H). As the latter ones were not dissected, only features that were available in the intact specimens are described. Specimen A (ZMUC-CRU-8679): Habitus (Fig. 2). Small arthropod larva with a globose shield, bearing a long, anteriorly directed, rostral spine (slightly shorter than shield length) and lateral spines (similar length as rostral spine).Fig. 2 a–d Autofluorescence images of a hippidan larva (ZMUC-CRU-8679) and a spider crab larva (Maja sp.). a Ventral view, fully enrolled. b Ventral view, fully outstretched. c Posterior view. d Dorsal view. Abbreviations: ant = antenna; atl = antennula; ce = compound eye; lb = labrum; mxp = maxilliped; pl = pleon; pls = postero-lateral spine; rst = rostral spine; te = telson; tp = thoracopod; vg = ventral gape Body (Fig. 2) differentiated into cephalothorax, pleon and non-somitic telson. Body with 20 segments, ocular segment plus 19 appendage-bearing (post-ocular) segments. Ocular segment incorporated into the cephalothorax, dorsal area contributes to the shield. Post-ocular segment 1-13 (Fig. 2) incorporated into the cephalothorax, dorsal area contributes to the shield. Post-ocular segment 14-19 (Fig. 2) are separate pleon segments, each dorsally forming a tergite. Cephalothorax (Fig. 2) shield more or less spherical, without setae; large, shield-like, cuticular structure formed by dorsal region of cephalothoracic segments. Anterior rim of the shield drawn out into prominent rostral spine. Posterior rim of the shield slightly convex, with a confined gape, as wide as the posterior gape of the shield. Rostrum unpaired anterior extension of shield, elongated, without spines; anterior region slightly bent upwards. Shield length about 8.5 mm (measured with rostral spine) and 5.1 mm without rostral spine, maximum shield width (measured without spines), about 5.2 mm (about 60 % of shield length with rostral spine). Rostral spine about 40 % of the shield length with rostral spine. Post-ocular segment 14 (Fig. 2) anterior-posterior dimension about 25 % of the shield length (without rostral spine); total width of the segment 25 % of the maximum shield width, as wide as the posterior gape of the shield; tergo-pleura not developed; anterior region of post ocular segment 15 slightly convex. Post-ocular segment 15 (Fig. 2) anterior-posterior dimension about 5 % of the shield length (without rostral spine). Total width of the segment 25 % of the maximum shield width, as wide as the posterior gape of the shield. Tergo-pleura not developed. Post-ocular segment 15 armed with one cone-shaped spine in the middle of anterior rim of the segment. Post-ocular segment 16 (Fig. 2) anterior-posterior dimension about 20 % of the shield length (without rostral spine). Total width of the segment about 25 % of the maximum shield width. Tergo-pleura not developed. Post-ocular segment 17 (Fig. 2) anterior-posterior dimension about 20 % of the shield length (without rostral spine). Total width of the segment about 25 % of the maximum shield width. Tergo-pleura not developed. Post-ocular segment 18 (Fig. 2) anterior-posterior dimension about 15 % of the shield length. Total width of the segment 50 % of the maximum shield width, measured on posterior rim of the segment. Axial region 25 % of the maximum shield width (without rostral spine). Tergo-pleura about 40 % of the axial region, on each side. Post-ocular segment 19 (Fig. 2) anterior-posterior dimension about 10 % of the shield length (without rostral spine). Total width of the segment about 50 % of the maximum shield width. No clear differentiation between axial region and tergo-pleura. Telson (Fig. 2) in dorsal view more or less rectangular. About 45 % of shield length (without rostral spine) and about 30 % wider than long. Anterior rim slightly concave, posterior rim convex. The lateral rim on each side slightly convex, telson width suddenly increased after about 20 % from anterior to posterior rim. Telson shape in lateral view distally tapering. Tip of telson more or less triangular-shaped from dorsal view, with a flattened tip. Forty-seven simple setae on tip of telson. Further lateral setae shorter than distal ones. The 20th setae counted from each terminal rim are the longest ones, the most central one is about 50 % shorter than the longest. Telson armed with two spines on distal rim as protrusion of lateral rim on each side. Lateral eyes (Fig. 2) compound eyes, with numerous ommatidia covered by cornea; stalked. Hypostome-labrum complex (Fig. 2) with more or less triangular-shaped labrum in ventral view, anteriorly surrounded by hypostome. Appendage 1 (Antennula) (Fig. 3) differentiated into peduncle and one flagellum. Antennula with aesthetascs. Peduncle more or less tube-shaped and curved to outer lateral rim of the shield. With spine-like protrusion on distal part of the inner lateral rim. Not yet divided into elements, future subdivision into three elements visible. Width of broadest part about 50 % of maximum length. Flagellum 1 not yet developed. Flagellum 2 tube-shaped with a rounded tip. About 35 % shorter than peduncle, with a slightly curved inner lateral rim with numerous setae (aesthetascs) arranged in six tiers.Fig. 3 Autofluorescence images of compound eyes, labrum, antennula, antenna, mandible, maxillula, and maxilla of the hippidan specimen (ZMUC-CRU-8679). Abbreviations: ba = basipod; cx = coxa; ed = endit; en = endopod; ex = exopod; fl = flagellum; ge = gnathic edge; pd = peduncle. Arrow: excretory opening Appendage of post-ocular segment 2 (Antenna) (Fig. 3) differentiated into coxa, basipod (peduncle), endopod and a paddle-shaped exopod; bears opening of antennal gland on basipod. Peduncle not yet divided into elements, with one spine on distal rim of basipod, where endopod arises from it. Endopod pointed and curved, not yet divided into elements, without setae. Exopod paddle-shaped, with 17 plumose setae on the rounded tip and the outer lateral rim. Appendage of post-ocular segment 3 (Mandible) (Fig. 3) differentiated into coxa with endite and mandibular palp. Coxa elongate in medio-lateral axis, medially ending in a row of about 12 teeth. Row consisting of more or less lobate teeth, short triangular and longer elongate teeth with a pointed tip. Mandibular palp not yet developed, but future palp visible. Sternal protrusion of mandibular segment (paragnaths) u-shaped with two lateral elongate paddle-shaped setae bearing protrusions on distal rim. About 35 % wider than maximum length and about as large as hypostome-labrum complex. Appendage of post-ocular segment 4 (Maxillula) (Fig. 3) differentiated into coxa with coxal endite and basipod with basipodal endite and endopod. Coxal endite more or less triangular-shaped from proximal to distal, with a rounded tip, with four elongated plumose setae at the tip. Basipodal endite paddle-shaped, elongate, with four spines at the tip, armed with tiny spines; about 30 % longer than coxal endite. Endopod pointed extension on basipod, not subdivided; one elongate, plumose seta, and one smaller seta on the tip. Appendage of post-ocular segment 5 (Maxilla) (Fig. 3) differentiated into coxa and basipod, both drawn out into two pronounced lobate endites each and exopod. Coxa with two lobate endites with four setae on each lobe. Distal lobe smaller, than proximal one. Basipod with two lobate endites with four setae on each lobe. Distal lobe larger, than proximal one. Endopod not yet developed. Exopod of appendage 5 largest element, bilobed with a distal and proximal lobe; with numerous plumose setae around the rim. Appendage of post-ocular segment 6 (Maxilliped 1) (Fig. 4) with coxa and basipod, from which endopod and exopod arise. Coxa and basipod with endites. Coxa more or less tube-shaped; with endite. Coxal endite small, more or less triangular from proximal to distal, without setae. Basipod more or less rectangular, with heart-shaped protrusion of anterior part; bears endopod and exopod; about 73 % longer than coxa, about twice as long as wide. Basipodal endite very prominent, slightly curved; with nine plumose setae, and four spines (armed with tiny spines) on inner lateral rim. Endopod with five elements (ischium, merus, carpus, propodus, dactylus); about as long as maximum length of basipod. Endopod element 1 more or less tube-shaped; about 45 % longer than wide, with four plumose setae on distal rim. Endopod element 2 more or less tube-shaped; about as long as preceding element; about 45 % longer than wide, with two plumose setae on distal rim. Endopod element 3 more or less tube-shaped, about 30 % shorter than preceding element; about 40 % longer than wide, with two plumose setae on distal rim. Endopod element 4 more or less tube-shaped; about 20 % shorter than preceding element, about 30 % longer than wide, with three plumose setae on distal rim. Endopod element 5 pointed, about 16 % shorter than preceding element; about 60 % longer than wide, with five plumose setae on distal rim. Exopod tube-shaped, tapering; not yet subdivided into elements; bent backwards. 85 % longer than wide, with about 12 plumose setae on the tip; laterally with one lobate protrusion slightly beyond the inner proximal rim and one lobate protrusion slightly below the tip.Fig. 4 Autofluorescence images of maxillipeds 1–3; thoracopods 4, 5, 6; thorax region; pleopods (orange); uropods (arrow = details of the exopod with setulae-bearing setae). Abbreviations: ba = basipod; cx = coxa; en = endopod; gi = gills; mxp = maxilliped; tp = thoracopod Appendage of post-ocular segment 7 (Maxilliped 2) (Fig. 4) with coxa and basipod, from which endopod and exopod arise. Coxa and basipod with endites. Coxa more or less tube-shaped, without setae; coxal endite small, more or less triangular from proximal to distal, without setae. Basipod more or less rectangular-shaped, with heart-shaped protrusion of anterior part; bears endopod and exopod, about 70 % longer than coxa, about twice as long as wide. Basipodal endite very prominent, slightly curved, with 3 plumose setae, on inner lateral rim. Endopod with four elements, about as long as maximum length of basipod. Endopod element 1 more or less tube-shaped, about 40 % longer than wide, with three plumose setae on inner distal rim. Endopod element 2 more or less tube-shaped; about 45 % longer than wide; about 10 % longer than preceding element; with two plumose setae on inner distal rim. Endopod element 3 more or less tube-shaped; about 50 % longer than maximum width, and about the same length than preceding element; with two plumose setae on inner distal rim. Endopod element 4 tapering with a rounded tip; about 60 % longer than wide; about 25 % shorter than preceding element; with three plumose setae on tip. Exopod of appendage 7 tube-shaped, tapering; not yet subdivided into elements; bent backwards, about 10 % shorter than endopod and about 85 % longer than wide; with about 12 plumose setae on the tip; setae bearing tip bent to the inner lateral side; laterally with one lobate protrusion slightly beyond the inner proximal rim and one lobate protrusion slightly below the tip. Appendage of post-ocular segment 8 (Maxilliped 3) (Fig. 4) with coxa and basipod, from which endopod and exopod arise; without endites. Coxa more or less rectangular-shaped, without setae; coxal endite not developed. Basipod more or less tube-shaped; bears endopod and exopod. Endopod about 55 % longer than coxa; about twice as long as wide. Endopod with four elements, separation indicated by faint lines, about 15 % shorter than maximum length of basipod and about 70 % longer than wide. Endopod element 1 more or less tube-shaped, about 20 % longer than maximum width; without setae. Endopod element 2 more or less rectangular shaped; about 30 % wider than maximum length; about 40 % shorter than preceding element; without setae. Endopod element 3 more or less tube-shaped; about 50 % longer than maximum width, and about the same length than preceding element; with two plumose setae on inner distal rim. Endopod element 4 tapering, about as long as maximum width; about 15 % longer than preceding element, with two simple setae on tip. Exopod tube-shaped, tapering. Not yet subdivided into elements; slightly bent backwards; about 15 % longer than endopod and about 80 % longer than wide; with about 18 plumose setae on the tip. Setae bearing tip bent to the inner lateral side. Laterally with one lobate protrusion slightly below the tip. Appendage of post-ocular segment 9 (Thoracopod 4) (Fig. 4) with coxa and basipod (difficult to differentiate in this developmental stage) and endopod; without setae. Endopod of appendage 9 consists of five visible elements (difficult to identify at this early developmental stage) separated by faint lines; distal part of this appendage is modified to a prominent chela. Endopod elements 1–2 probably corresponding to ischium and merus, not yet separated; more or less tube-shaped; about 25 % longer than maximum width. Endopod element 3 (carpus) more or less tube-shaped, curved to the inner side; about 15 % longer than maximum width; about 12 % shorter than preceding element. Endopod element 4 (propodus) about 45 % longer than maximum width, with outgrowth, which represents the complement of the following element; outgrowth about 35 % shorter than mainpart of the element. Endopod element 5 (dactylus) movable against outgrowth of propodus; tapering, slightly curved; about 50 % of maximum length of preceding element (propodus). Chela is formed by the articulation of element 5 (dactylus) against an outgrowth of element 4 (propodus). Appendage of post-ocular segment 10 (Thoracopod 5) (Fig. 4) with coxa and basipod (difficult to differentiate in this early developmental stage) and endopod. Without setae. Coxa difficult to differentiate in this early developmental stage; without setae. Basipod difficult to differentiate in this early developmental stage; without setae. Endopod consists of five elements (difficult to identify in this early developmental stage); distal part without chela. Endopod element 1 (ischium) more or less triangular-shaped from proximal to distal; about 25 % longer than maximum width; without setae. Endopod element 2 (merus) more or less tube-shaped, slightly curved to inner side; about 40 % longer than maximum length and 40 % longer than preceding element; without setae. Endopod element 3 (carpus) tube-shaped; about 30 % longer than maximum width and about 8 % shorter than preceding element; without setae. Endopod element 4 (propodus) tube-shaped, about as long as maximum width and about 30 % shorter than preceding element; without setae. Endopod element 5 (dactylus) tapering, with a slightly rounded tip; about 80 % longer than maximum width, and about 60 % longer than maximum width; without setae. Appendage of post-ocular segment 11 (Thoracopod 6) (Fig. 4) with coxa and basipod (difficult to differentiate in this early developmental stage) and endopod; without setae. Appendage of post-ocular segment 12 (Thoracopod 7) (Fig. 4) with coxa and basipod and endopod (difficult to differentiate in this early developmental stage, separation indicated by faint lines); without setae. Appendage of post-ocular segment 13 (Thoracopod 8) (Fig. 4) with coxa and basipod and endopod (difficult to differentiate in this early developmental stage, separation indicated by faint lines); without setae. Appendage of post-ocular segment 14 (Pleopod 1) (Fig. 4) not found, not documented. Appendage of post-ocular segment 15 (Pleopod 2) (Fig. 4) differentiated into basipod and endopod. Separation indicated by a faint line. Basipod elongate tube-shaped, about 50 % longer than maximum width. Endopod of appendage 15 tube-shaped, with a rounded tip; about 40 % longer than maximum width, and about 20 % shorter than basipod. Appendage of post-ocular segment 16 (Pleopod 3) (Fig. 4) differentiated into basipod and endopod. Separation indicated by a faint line. Basipod elongate tube-shaped, about 35 % longer than maximum width. Endopod of appendage 16 tube-shaped, with a rounded tip; about 25 % longer than maximum width, and about 35 % shorter than basipod. Appendage of post-ocular segment 17 (Pleopod 4) (Fig. 4) differentiated into basipod and endopod. Separation indicated by a faint line. Basipod elongate tube-shaped, about 35 % longer than maximum width. Endopod of appendage 17 tube-shaped, with a rounded tip; about 40 % longer than maximum width, and about 10 % shorter than basipod. Appendage of post-ocular segment 18 (Pleopod 5) (Fig. 4) differentiated into basipod and endopod. Separation indicated by a faint line. Basipod elongate tube-shaped, about 47 % longer than maximum width. Endopod of appendage 18 tube-shaped, with a rounded tip; about 40 % longer than maximum width, and about 35 % shorter than basipod. Appendage of post-ocular segment 19 (Uropod) (Fig. 4) differentiated into basipod and endopod and exopod. Basipod tube-shaped, 55 % longer than maximum width. Endopod tube-shaped with a rounded tip, about 40 % longer than maximum width. Without setae. Exopod paddle-shaped, with about 16 plumose setae on the distal rim, and the inner lateral rim of appendage. With one spine representing the delineation of setae and extention of the outer rim of exopod. Morphological description of comparable features of specimens B–H Specimen B (ZMUC-CRU-8681) (Fig. 5):Fig. 5 a–h Composite images under cross-polarized light of hippidan specimens. A 1–A 2 Dorsal and ventral view (entirely enrolled) of specimen ZMUC-CRU-8679; morphotype 1. A 3 Detail of telson of morphotype 1. B 1–B 2 latero-ventral and lateral view of specimen ZMUC-CRU-8681 (entirely enrolled); morphotype 3. C 1–C 2 Dorsal and ventral view of specimen ZMUC-CRU-8680 (entirely enrolled); morphotype 3. D 1–D 3 Dorsal, ventral and lateral view of specimen ZMUC-CRU-8683 (enrolled); morphotype 2. D 4 Detail of telson of morphotype 2. E 1–E 3 Frontal, postero-lateral, and ventral view of specimen MNHN-IU-2014-5468 (entirely enrolled); morphotype 4. F 1 Ventral view of specimen Mu_267 (entirely enrolled); morphotype 1. G 1 Ventral view of specimen ZMUC-CRU-8684 (entirely outstretched); morphotype 2. H 1–H 2 Ventral and dorsal view of specimen ZMUC-CRU-8682 (entirely outstretched); morphotype 3. H 3 Detail of telson of morphotype 3 Shield. Shield length about 8.5 mm (measured with rostral spine). Maximum width (measured without spines) about 3.5 mm (about 40 % of shield length). Rostral spine. Anterior part slightly bent upwards. About 45 % of the shield length. Telson. Anterior and posterior rim slightly convex. The lateral rim on each side slightly convex, width slowly increasing from anterior to posterior. Rim of telson more or less curly brace-shaped, with a rounded tip. Telson armed with two spines on distal rim as protrusion of lateral rim on each side. Specimen C (ZMUC-CRU-8680) (Fig. 5): Shield. Shield length about 7.5 mm (measured with rostral spine). Maximum width (measured without spines, about 3.4 mm (45 % of shield length). Rostral spine. Anterior part slightly bent upwards. About 35 % of the shield length. Telson. Anterior rim slightly convex, posterior rim slightly concave. The lateral rim on each side slightly convex, width slowly increasing from anterior to posterior. Rim of telson more or less curly brace-shaped, with a rounded tip. Telson armed with two spines on distal rim as protrusion of lateral rim on each side. Specimen D (ZMUC-CRU-8683) (Fig. 5): Shield. Shield length about 5 mm (measured with rostral spine). Maximum width (measured without spines, about 2.3 mm (45 % of shield length). Rostral spine. Anterior part slightly bent upwards. About 35 % of the shield length. Telson. Anterior rim slightly convex, posterior rim slightly concave. The lateral rim on each side slightly convex, width slowly increasing from anterior to posterior. Rim of telson more or less triangular-shaped from dorsal view, with a flattened tip. Telson armed with two spines on distal rim as protrusion of lateral rim on each side. Specimen E (MNHN-IU-2014-5468) (Fig. 5): Shield. Shield length about 7.4 mm (measured with rostral spine) and about 3 mm (measured without rostral spine). Maximum width (measured without spines, about 2.5 mm (30 % of shield length). Rostral spine. Anterior part strongly bent downwards. About 120 % of the shield length. Telson. Anterior rim slightly convex, posterior rim slightly concave. Lateral rim difficult to recognize; apparently slightly convex, probably with lobate structure. Telson width suddenly increasing after about 45 % from anterior to posterior rim. Rim of telson more or less triangular-shaped from dorsal view, with a slightly flattened tip. Telson probably armed with two spines on distal rim as protrusion of lateral rim on each side. Specimen F (Mu_267) (Fig. 5): Shield. Shield length about 8.3 mm (measured with rostral spine). Maximum width (measured without spines), about 4 mm (50 % of shield length). Rostral spine. No bending visible. Not documented from lateral, or ventro-lateral view. About 50 % of the shield length Telson. Anterior rim slightly convex, posterior rim slightly concave. Lateral rim slightly convex, with lobate structure. Telson width suddenly increased after about 40 % from anterior to posterior rim. Rim of telson more or less triangular-shaped from dorsal view, with a flattened tip. Telson armed with two spines on distal rim as protrusion of lateral rim on each side. Specimen G (ZMUC-CRU-8684) (Fig. 5): Shield. Shield length about 5.2 mm (measured with rostral spine). Maximum width (measured without spines), about 2.5 mm (50 % of shield length). Rostral spine. No bending visible. Not documented from lateral, or ventro-lateral view. About 40 % of the shield length. Telson. Anterior rim slightly convex, posterior rim slightly concave. The lateral rim on each side slightly convex, width slowly increasing from anterior to posterior. Rim of telson more or less triangular-shaped in dorsal view, with a flattened tip. Telson armed with two spines on distal rim as protrusion of lateral rim on each side. Specimen H (ZMUC-CRU-8682) (Fig. 5): Shield. Shield length about 9 mm (measured with rostral spine). Maximum width (measured without spines, about 3.6 mm (40 % of shield length). Rostral spine. Anterior part slightly bent upwards. About 40 % of the shield length. Telson. Anterior rim slightly convex, posterior rim slightly concave. The lateral rim on each side slightly convex, width slowly increasing from anterior to posterior. Rim of telson more or less curly brace-shaped, with a rounded tip. Telson armed with two spines on distal rim as protrusion of lateral rim on each side. Discussion Identification of the specimens The specimens described here show some inter-individual differences, but are sufficiently similar to be discussed together. The overall morphology immediately identifies them as reptantian zoea larvae; the embryonic-like posterior thoracopods identify them as an ingroup of Meiura (cf. e.g. [3, 23]). Most meiuran zoeae possess a forked telson (e.g. [3, 23–25]), while the posterior rim of the telson of the here described specimens has a roughly convex shape in dorsal view (with additional lobate protrusions in some specimens). This shape is known from zoea larvae of hippidan species (Figs. 5, 6). Hippidan larvae, similar to the present specimens, have been described with a more or less spherical shield with a rostral spine and two large spines on each postero-lateral margin, but no postero-dorsal spine. Finally, representatives of Hippidae achieve significantly larger sizes as zoea larvae [3, 4] compared to other meiurans. This is also true for the specimens described here (Figs. 2, 5).Fig. 6 Drawings of the different telson-shapes in dorsal view of the investigated species. a Morphotype 1. b Morphotype 2. c Morphotype 3 After metamorphosis to the megalopa stage in hippidans, antennula and antenna include a long, setose flagellum, the mandible is divided into two parts, thoracopods are divided into elements and largely resemble the setae-bearing juvenile and adult ones; the pleopods as well as the exopods of the uropods bear setae [4]. The new specimens are therefore interpreted as zoea larvae of species of Hippidae. A late zoea stage is indicated by a differentiation of the non-setae-bearing uropods into an endopod and exopod, by more than three tiers of aesthetascs on the antennula, comparably large maxillipeds, and the presence of primordial thoracopods and pleopods. The details of the appendages of specimen A also strongly resemble known late zoea features (besides their size) [26]. Although the other specimens have not been documented in detail, all are interpreted as representing rather late or final zoea stages. This is supported in particular by a notable difference to already known zoea larvae of Hippidae: in all specimens the sixth pleomere is already set off from the telson (Figs. 2, 4, 5, 6). Hippidan species for which the larval sequence is known are considered to have the sixth pleomere conjoined to the telson in all zoea stages (forming a pleotelson), becoming finally articulated in the megalopa stage [27]. Hence, in the specimens described here, we probably have ultimate zoea larvae, which could be also interpreted as early megalopae [28], retaining some (in fact most) zoea characters, but already having some megalopa characters. Hippidans are able to vary the number of larval stages (e.g. [29]); these specimens could represent such a case of a prolonged pelagic phase, most likely due to the lack of a settling trigger, which would induce transformation to the megalopa. Larvae of mole crabs – what is known so far Hippoidea includes the groups Blepharipodidae, Albuneidae and Hippidae, the latter two representing sister groups (e.g. [30]). Larval representatives of Albuneidae and Hippidae feature a more or less spherical shield equipped with one prominent lateral spine on each side (absent in larvae of blepharipodids) and one elongate rostral spine [25, 31, 32]. Despite obvious general similarities due to their close relationships, larvae of Albuneidae and Hippidae differ in many aspects [4, 26, 31, 33, 34], assuming the here described specimens are hippidans. A specimen observed by Gurney [25] was referred to as Albunea sp. It strongly resembles one of the specimens described herein (Fig. 5h), and is probably also a hippidan larva. Within Hippidae, there are currently only three species groups recognised: Emerita, Hippa and Mastigochirus (e.g. [4, 26, 33]). Species of Hippa and Emerita have very similar zoea larvae, with only a few differences. For comparison we refer to the last zoea stage before metamorphosis as the megalopa, since the detailed described larva (specimen A, ZMUC-CRU-8679) is most likely a very late stage zoea. In larvae of Hippa species the rostral spine is slightly curved downwards, but upwards in larval representatives of Emerita. Also, the former bear fewer aesthetascs on the antennula, a shorter flagellum on the antenna, fewer setae on the exopod of the maxilla, and fewer setae on the telson. Additionally, larval forms of Emerita (as far as known) do not achieve the impressive size of larval forms of Hippa. The specimens in the 6th zoeal stage of Hippa can reach a shield length of up to 6 mm, whereas zoea larvae in the same stage of Emerita only reach a shield length of about 2 mm (see [4, 26, 33]). Grouping of material described here Based on gross morphological aspects, the specimens described here can be sorted into four more or less distinct groups or morphotypes. This will allow an easier comparison with existing descriptions instead of treating each specimen separately. Morphotype 1 includes specimens A and F. Both share a more spherical shield, a straight rostral spine, a telson with lobate extensions laterally and a flattened tip. Morphotype 2 includes specimen D and probably G. Both possess a more elongate shield, a straight rostral spine and a telson lacking a lateral protrusion. Morphotype 3 includes specimens B, C and H. They all possess a shield comparable to morphotype 2, yet the shape of the posterior margin of the telson is not simple and convex (as in morphotype 2), but with a curled, brace-shaped distal margin with a small and rounded lobate tip. Morphotype 4 is represented by specimen E, which features a more or less spherical shield with a curved rostral spine. The telson of the specimen features lateral protrusions and a flattened tip as in morphotype 1. Comparison to the new material Specimen A (morphotype 1) is quite large, almost reaching the size of a specimen described by Martin and Ormsby [4], which was interpreted as representing the larva of a species of Hippa. The rostral spine of the latter specimen is curved downwards, whereas in specimen A the rostral spine is curved slightly upwards. The antennula of specimen A bears six, rather than the typical five, tiers of setae. It is divided into strongly curved peduncle and flagellum, instead of being not subdivided into a peduncle and flagellum (Fig. 3). The antenna also differs in many aspects. The basipod bears one spine on its distal rim, and there is no flagellum developed as described by Martin and Ormsby [4]. A marked difference constitutes the pointed and strongly curved endopod without setae and the presence of a well-developed paddle-shaped exopod with numerous plumose setae, instead of an endopod with spines and no exopod on the described species of Martin and Ormsby [4] (Fig. 3). The maxilla has two coxal and two basipodal endites and no endopod, instead of no endites and a setae-bearing endopod. The bilobed exopod and the number of setae of the specimen described here largely resemble the description of Martin and Ormsby [4] (Fig. 3). The number of telson setae is about twice the number described by Martin and Ormsby [4]. Additionally, the telson is equipped with two lobate structures on each lateral rim (Fig. 4). Concerning the large size, one might assume that the observed larva is a representative of Hippa. However, the considerable morphological differences between the described structures of the specimen of Martin and Ormsby [4] and specimen A (and morphotype 1 in general) does not support its interpretation as a representative of Hippa. Specimen A also differs in many aspects from larvae of species of Emerita described by Knight [26]. Concerning the larger number of aesthetascs on the antennula, and the higher number of setae on the exopod of the maxilla and the telson, specimen A resembles larvae of an Emerita species. Additionally, the rostral spine of specimen A is bent upwards, as presented in the drawings of Knight [26]. In specimen A, there is no flagellum on the antennula, the antenna bears a paddle-shaped exopod, the telson differs morphologically concerning the lobate structure on the lateral rim, and most strikingly, larvae of Emerita are significantly smaller; they achieve a mean size of only about 2 mm shield length [26]. Specimen E (morphotype 4), due to the downwards curvature of the rostral spine and the size of about 3 mm (Fig. 5), matches earlier descriptions of larvae of Hippa (e.g. [4]). Hence, specimen E is likely a larval representative of Hippa. The additionally documented specimens, although not investigated in detail, do not show many similarities with the known larvae of Hippa or Emerita either. They also differ among each other (see also below). While based on the size we can hypothesize that the specimens represent late, but different larval stages, not all specimens can be attributed to a single developmental sequence. Since all specimens described here, except specimen E, differ greatly from earlier described species (Fig. 5), we are unable to determine whether they are larvae of a species of Emerita, Hippa or Mastigochirus. These few specimens indicate that there is still an unknown morphological diversity within larval hippidans. The morphological differences are probably not caused by ontogenetic factors. As discussed above, the larvae most likely either represent ultimate zoea stages or even early megalopa stages. Therefore, they probably do not represent subsequent stages of a single species. The unexpected diversity may indicate that these larvae are individuals of species for which larvae are wholly unknown, differing more significantly than expected from other larval sequences. Yet, it is also possible that they represent special cases of developmental plasticity, which means that they are morphological variations of already known larvae caused by specific environmental conditions. In any case, the morphological diversity of hippidan larvae appears to be higher than that of hippidan adults. Enrollment The specimens described here have not been observed when alive. Still, we can make inferences about their original behavior based on their functional morphology, as recently suggested by Haug and Haug [7]. The basic idea is to employ approaches from paleontology; e.g., recording the preserved position of specimens, identifying specialized morphological features, and comparing these to those of animals with similar features in which behavior can be directly observed. Using these approaches, while any conclusions remain a matter of conjecture, they nonetheless represent an important tool for understanding. So far it has been impossible to breed giant hippidan larvae or to observe them directly in the field. Hence, the approach discussed here is currently the only possibility. Ideally the prediction made here can be tested in the field (see [7] for more details and a comparable example, also corroborated by field observation). All specimens were originally preserved in an enrolled position, indicating the possibility of the animals to achieve this position. Further morphological adaptations are given for each morphotype separately. Morphotype 1 (Figs. 2, 5, 6): Specimen A (ZMUC-CRU-8679), and specimen F (Mu_267). In these specimens the shield appears more or less spherical and the posterior gape of the shield has the same width as the pleonal tergites. This combines a spherical shield with maximum mobility of the pleon. The pleon can be stretched out or flexed forward, without any limitations. The large and lid-like telson features a lobate structure on each lateral margin, which ends up in a spine on the distal rim, and the rim of the telson is more or less triangular-shaped in dorsal view and has a slightly flattened tip. The width of the telson and the length of the pleon appear not to be entirely adapted to the ventral gape of the shield. The telson is broader, and due to a comparatively short pleon, the telson does not reach the anterior rim of the shield, even if the pleon is fully flipped forward. Therefore, antennula and antenna as well as the compound eyes and the distal parts of the maxillipeds are not fully concealed. Morphotype 2 and morphotype 3 (Fig. 5, 6): Specimen D (ZMUC-CRU-8683) and probably specimen G (ZMUC-CRU-8684); respectively specimen B (ZMUC-CRU-8681), specimen C (ZMUC-CRU-8680) and specimen H (ZMUC-CRU-8682). The two morphotypes differ from each other only in the shape of the posterior rim of the telson, but otherwise are quite similar and therefore treated together. Similar to morphotype 1, both morphotypes feature a posterior gape which has the same width as the pleon. In both morphotypes, the width of the telson is not adapted to the ventral gape of the shield. The telson is wider than the ventral gape and has a different shape. Yet, when flipped forward, the ventral gape is entirely covered across its width. The large telson has slightly convex lateral margins, which end up in a spine on each side, and the telson width increases slightly from anterior to posterior. Additionally, due to a longer pleon the larvae are able to flex the telson far anteriorly so that the ventral gape of the shield is entirely closed and almost all parts of the appendages are covered by the telson, which perfectly protects the entire body. Only the compound eyes and the distal parts of antennula and antenna remain exposed. Morphotype 4 (Fig. 5, 6): Specimen E (MNHN-IU-2014-5468) mainly resembles morphotype 1 in relation to the more or less spherical shield shape, the ventral gape and the shape of the telson. As in morphotypes 2 and 3, the pleon is flexed far anteriorly. Additionally, the telson is positioned inside the shield and reaches the anterior rim (Fig. 5). All appendages, except the eyes and the distal tip of the antennulae are protected by the shield, pleon and telson. The fully enrolled specimen has the appearance of a compact ball, armed with spines. Here the position of the telson is most important. The telson appears to perfectly fit into the shield, with “rail-like” protrusions of the shield keeping it in place. In this position, the telson is tightly locked in place, and the dorsal area of the pleon perfectly closes the posterior gape. With this, the lateral rim of the telson and the ventral rim of the shield apparently form coaptative structures that tightly enclose the enrolled body. The animal probably achieves this position by 1) flexing its pleon forward, 2) pressing it towards its ventral side right in front of the mouthparts and maxillipeds, and 3) sliding it back. During the last step, the lid-like telson is pulled inside the ventrally curved rims of the shield. As a result, the animal secures the enrolled position and achieves full protection of the ventral appendages. Interestingly, this mechanism appears to be arranged “the-other-way-round” as compared to the coaptative structures in stomatopod larvae where the telson is pushed forward in order to lock it [7]. So do giant hippidan larvae perform defensive enrollment? Based on our observations we can state that:all specimens are preserved in an enrolled position, indicating that the animals can achieve this position; the shield is large and drawn out ventrally for some distance, unlike in many other decapod zoea larvae, and it is able to house most of the appendages; the width of the shield and the width of the pleon are perfectly adapted to one another; this is not a widespread phenomenon (see Discussion in [7]) and can hence be interpreted as an adaptation for performing enrollment; a large lid-like telson covers most of the ventral gape if flipped forward; also this morphology is rather unusual for meiurans (where the pleon is usually forked); thus it is also possible that this morphology represents a further adaptation for enrollment; at least for morphotype 4 (specimen E) we have indications that there are coaptative structures; such structures are strong indicators of defensive enrollment, as these are developed in other groups for which enrollment seems now corroborated, such as trilobites ([15], their figs. 1E–F; [35, 36]) or stomatopod larvae ([7]). Hence, for morphotype 4 there should be little doubt about whether it was able to perform enrollment. For the other three, the stronger argument of the coaptative structures cannot be used. Yet, as in the discussion about different morphotypes of stomatopod larvae [7], the presence of only some of the adaptations cannot be used as an argument to exclude this behavior. Also among other animals which are known to perform defensive enrollment coaptative structures appear to be absent (e.g. polyplacophorans; stonefly larvae of Pteronarcys dorsata [37]). With this, we consider it likely that all specimens were able to perform defensive enrollment, but to differing degrees of specialization. The ventral “softer” part of the body is in all cases concealed by the spine-bearing shield and the sclerotized pleonal tergites. Yet, in morphotype 1 (specimens A, F) there is a larger unconcealed region remaining up to the anterior margin of the ventral gape of the shield (Fig. 5; similar to a stomatopod larva morphotype, [7]). Morphotypes 2 (specimens D, G) and 3 (specimens B, C, H) achieve the same defensive effectiveness since they flex the telson further anteriorly, and the ventral gape of the shield is almost entirely closed (Fig. 5). Morphotype 4 (specimen E) achieves the most effective degree of defense since there is no unconcealed region left up to the anterior margin of the ventral gape, and due to the position of the telson inside the shield (Fig. 5). Comparison to Brachyura and Stomatopoda Martin and Ormsby [4] have stated that hippidan larvae appear very similar to brachyuran larvae. This seems to be largely attributable to the shield morphology. Brachyuran zoeae also feature a spherical shield; yet here the rostral spine is directed more ventrally (e.g. [38]) instead of being anteriorly directed as in most hippidan larvae (Fig. 5). Additionally, brachyuran zoea larvae have no uropods [39, 40], whereas hippidan larvae feature uropods from their early zoeal stages onwards [26] (Figs. 2, 4, 5). Brachyuran zoea larvae have two lateral spines and the rostral spine (as hippidan zoeae), but additionally a long postero-dorsal spine, which is absent in hippidan larvae (Figs. 2, 5). Also the telson differs morphologically. There is a pronounced furca with a medial cleft in brachyurans (at least in early zoea stages; [23, 34]. Hippidan larvae never have a forked telson (Figs. 2, 3, 4 and 5), but the posterior rim is convex. Most strikingly, brachyuran larvae do not achieve the giant size of hippidan larvae; brachyuran larvae with small body lengths seem to be very common (e. g. [41, 42]). Discrete lengths are rarely stated for brachyuran larvae. A shield length (without the rostral spine) of 0.88 mm in an advanced zoeal stage has been reported [34]. The largest late stage zoea in our investigation (specimen F, Mu_267) reaches a shield length of about 5.5 mm (Fig. 5); 6 mm has been reported [4]. There have been no reports to date of defensive enrollment in brachyuran zoea larvae, nor do we see morphological adaptations for it, but that may change if such features are searched for directly. Currently, only a roughly spherical shield with three spines seems to be similar between brachyuran and hippidan zoeae (Fig. 2). Stomatopods (mantis shrimps) are also discussed here in light of their similarities to hippidan zoea larvae, since some mantis shrimp larvae can also tightly enroll their bodies [7] (Fig. 1). Interestingly, the specimens described here were found between mantis shrimp larvae in two museum collections (see Methods for details; also in other collections, pers. obs.) in roughly pre-sorted samples. This shows nicely how similar mantis shrimp and hippidan larvae appear at first sight. Also here especially the shield appears similar, even more similar than to brachyuran larvae as stomatopod larvae, like hippidan zoea larvae, lack the pronounced postero-dorsal spine of brachyuran zoeae (Figs. 1, 5). A future, more intensive functional comparison of hippidan and stomatopod larvae could reveal the evolutionary mechanisms leading to the morphological adaptation coupled to defensive enrollment. Conclusions and Prospects Our investigation indicates a broader morphological diversity of hippidan larvae than has been described previously. The functional morphological aspects of these larvae suggest a behavior by these larvae that has not been directly observable to date. It thus appears that we are just starting to understand the ecological roles played by many crustacean larvae. Hence, we expect to continue to uncover hidden morphological diversity among these larvae, and will seek to reconstruct their functional morphology and evolutionary history. Additional file Additional file 1: Descriptive matrix of specimen A (ZMUC-CRU-8679). (XLS 39 kb) Additional file 2: Persistent identifiers for specimens in database MorphDBase. (XLS 7.50 kb) Abbreviations ANTAntenna ATLAntennula BABasipod CECompound eye CXCoxa EDEndit ENEndopod EXExopod FLFlagellum GEGnathic edge GIGills LBLabrum MXPMaxilliped PDPeduncle PLPleon PLSPostero-lateral spine RSTRostral spine TETelson TPThoracopod VGVentral gape Acknowledgements We are grateful to numerous persons and institutions that made this investigation possible. Jason Dunlop, Berlin, kindly made linguistic improvements. Enrico Schwabe, ZSM Munich, Jørgen Olesen, Tom Schiøtte, Danny Eibye-Jacobsen, and Jens T. Høeg, ZMUC Copenhagen, the curating people from the crustacean collection at the Senckenberg museum Frankfurt, and Laure Corbari, MNHN Paris, provided access to the collections and assisted with the technical equipment. All authors would like to thank J. Matthias Starck, LMU Munich, for his support. This study was also supported by people providing free or extremely low-cost software, such as CombineZM/ZP, Microsoft Image Composite Editor, and OpenOffice. Funding This contribution is part of JTH’s research project kindly funded by the German Research Foundation (DFG) under Ha 6300/3-1. CH is grateful for support via the Bavarian Equal Opportunities Sponsorship of the LMU. The research visits at ZMUC Copenhagen and MNHN Paris of CH and JTH have been made possible by grants from the European Commission’s (FP 6) Integrated Infrastructure Initiative programme SYNTHESYS (DK-TAF-2591, FR-TAF-5175, FR-TAF-5181). Availability of data and materials High-resolution images of the investigated specimens supporting the conclusions of this article are available in the MorphDBase repoitory: https://www.morphdbase.de. The persistent identifiers are provided in Additional file 2. Authors’ contributions JTH and CH designed the study. JTH and CH documented specimens ZMUC-CRU-8680 to 8684, MNHN-IU-2014-5468, and Mu_267. NRR performed the preparation and documentation of specimen ZMUC-CRU-8679 and drafted the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate Not applicable. ==== Refs References 1. Haye PA Tam YK Kornfield I Molecular phylogenetics of mole crabs (Hippidae: Emerita ) J Crustac Biol 2002 22 903 915 10.1163/20021975-99990302 2. Haig J The status of Remipes testudinarius Latreille, and designation of a neotype for Hippa adactyla J.C. Fabricius (Decapoda, Hippidae) Crustaceana 1970 19 288 296 10.1163/156854070X00374 3. Harvey A Boyko CB McLaughlin P Martins JW Martin JW Olesen J Høeg JT Anomura Atlas of Crustacean Larvae 2014 Baltimore Johns Hopkins Univ Press 284 295 4. Martin J Ormsby B A large brachyuran-like larva of the Hippidae (Crustacea: Decapoda: Anomura) from the Banda Sea, Indonesia: the largest known zoea P Biol Soc Wash 1991 104 561 568 5. Ward P Shell sculpture as a defensive adaptation in ammonoids Paleobiol 1981 7 96 100 10.1017/S009483730000381X 6. Schwartz SS Predator-induced alterations in Daphnia morphology J Plankton Res 1991 13 1151 1161 10.1093/plankt/13.6.1151 7. Haug C Haug JT Defensive enrolment in mantis shrimp larvae (Malacostraca: Stomatopoda) Contrib Zool 2014 83 185 194 8. Gaudin TJ Biewener A The functional morphology of xenarthrous vertebrae in the armadillo Dasypus novemcinctus (Mammalia, Xenarthra) J Morphol 1992 214 63 81 10.1002/jmor.1052140105 1433308 9. Chen IH Kiang JH Correa V Lopez MI Chen PY McKittrick J Meyers MA Armadillo armor : Mechanical testing and micro-structural evaluation J Mech Behav Biomed Mater 2011 4 713 722 10.1016/j.jmbbm.2010.12.013 21565719 10. Squires RL Goedert JL An extant species of Leptochiton (Mollusca: Polyplacophora) in Eocene and Oligocene cold-seep limestones, Olympic Peninsula, Washington Veliger 1995 38 47 53 11. Schwabe E, Wehrtmann IS. Chitons. In: Wehrtmann IS, Cortés J, editors. Marine Biodiversity of Costa Rica, Central America. Springer + Business Media B.V.; 2009. p. 323–31. 12. Speyer SE Brett CE Trilobite taphonomy and Middle Devonian taphofacies Palaios 1986 1 312 327 10.2307/3514694 13. Bruton DL Haas W Functional morphology of Phacopinae (Trilobita) and the mechanics of enrolment Palaeontogr Abt A 1997 245 1 43 14. Lerosey-Aubril RL Angiolini L Permian trilobites from Antalya Province, Turkey, and enrollment in Late Palaeozoic trilobites Turkish J Earth Sci 2009 18 427 448 15. Feist R Lerosey-Aubril R Johnson R Coaptative devices, enrollment, and life habits in Paralejurus , a particular case in scutelluid trilobites Palaeobiol Palaeoenviron 2010 90 125 137 10.1007/s12549-010-0025-1 16. Haug C, Mayer G, Kutschera V, Waloszek D, Maas A, Haug JT. Imaging and documenting gammarideans. Int J Zool. 2011;380829. 17. Haug JT Haug C Kutschera V Mayer G Maas A Liebau S Castellani C Wolfram U Clarkson ENK Waloszek D Autofluorescence imaging, an excellent tool for comparative morphology J Microsc 2011 244 259 272 10.1111/j.1365-2818.2011.03534.x 21883208 18. Haug JT Haug C Ehrlich M First fossil stomatopod larva (Arthropoda: Crustacea) and a new way of documenting Solnhofen fossils (Upper Jurassic, Southern Germany) Palaeodiv 2008 1 103 109 19. Haug C Haug JT Waloszek D Maas A Frattigiani R Liebau S New methods to document fossils from lithographic limestones of southern Germany and Lebanon Palaeontol Electron 2009 12 6T 20. Haug JT Briggs DEG Haug C Morphology and function in the Cambrian Burgess Shale megacheiran arthropod Leanchoilia superlata and the application of a descriptive matrix BMC Evol Biol 2012 12 162 10.1186/1471-2148-12-162 22935076 21. Walossek D The upper Cambrian Rehbachiella and the phylogeny of Branchiopoda and Crustacea Lethaia 1993 26 1 318 10.1111/j.1502-3931.1993.tb01502.x 22. Garm A Watling L The crustacean integument: setae, setules, and other ornamentation Funct Morphol Divers 2013 1 167 198 10.1093/acprof:osobl/9780195398038.003.0006 23. Martin JW Martin JW Olesen J Høeg JT Brachyura Atlas of Crustacean Larvae 2014 Baltimore Johns Hopkins Univ Press 295 310 24. Aikawa H On larval forms of some Brachyura Rec Oceanogr Works Japan 1929 2 17 55 25. Gurney R Larvae of decapod Crustacea 1942 London Ray Society 26. Knight M The larval development of the sand crab Emerita rathbunae Schmitt (Decapoda, Hippidae) Pacific Sci 1967 21 58 76 27. Siddiqui FA Ghory FS Complete larval development of Emerita holthuisi Sankolli, 1965 (Crustacea: Decapoda: Hippidae) reared in the laboratory Turkish J Zool 2006 30 121 135 28. Haug JT Haug C An unusual fossil larva, the ontogeny of achelatan lobsters, and the evolution of metamorphosis Bull Geosci 2013 88 195 206 29. Diaz H Costlow JD Seasonal developmental sequences of Emerita talpoida (Say) (Anomura, Hippidae) J Coast Res 1987 3 169 177 30. Bracken-Grissom HD Cannon ME Cabezas P Feldmann RM Schweitzer CE Ahyong ST Felder DL Lemaitre R Crandall KA A comprehensive and integrative reconstruction of evolutionary history for Anomura (Crustacea: Decapoda) BMC Evol Biol 2013 13 128 10.1186/1471-2148-13-128 23786343 31. Stuck KC Truesdale FM Larval and early postlarval development of Lepidopa benedicti Schmitt, 1935 (Anomura: Albuneidae) reared in the laboratory J Crustac Biol 1986 6 89 110 10.2307/1547933 32. Konishi K Larval development of the spiny sand crab Lophomastix japonica (Duruflé, 1889)(Crustacea, Anomura, Albuneidae) under laboratory conditions Publ Seto Mar Biol Lab 1987 32 123 139 33. Hanson AJ The larval development of the sand crab Hippa cubensis (De Saussure) in the laboratory (Decapoda, Anomura) Crustaceana 1969 16 143 157 10.1163/156854069X00402 34. Thatje S Fuentes V First record of anomuran and brachyuran larvae (Crustacea: Decapoda) from Antarctic waters Polar Biol 2003 26 279 282 35. Esteve J Zamora S Gozalo R Liñán E Sphaeroidal enrolment in middle Cambrian solenopleuropsine trilobites Lethaia 2010 43 478 493 10.1111/j.1502-3931.2009.00205.x 36. Esteve J Hughes NC Zamora S Thoracic structure and enrolment style in Middle Cambrian Eccaparadoxides pradoanus presages caudalization of the derived trilobite trunk Palaeontology 2013 56 589 601 10.1111/pala.12004 37. Moore KA, Williams DD. Novel strategies in the complex defense repertoire of a stonefly (Pteronarcys dorsata) nymph. Oikos. 1990;57:49–56. 38. Williamson DI Abele LG Larval morphology and diversity The Biology of Crustacea 1982 New York Academic 43 110 39. Rice AL Crab zoeal morphology and its bearing on the classification of the Brachyura Trans Zool Soc London 1980 35 271 372 10.1111/j.1096-3642.1980.tb00060.x 40. Martin JW Crabs of the family Homolodromiidae, III. First record of the larvae J Crustac Biol 1991 11 156 161 10.2307/1548553 41. Wear RG Life-history studies on New Zealand Brachyura: 4. Zoea larvae hatched from crabs of the family Grapsidae New Zeal J Mar Freshw Res 1970 4 3 35 10.1080/00288330.1970.9515324 42. Lago RP Larval development of Spiroplax spiralis (Barnard, 1950) (Brachyura: Hexapodidae) in the laboratory; the systematic position of the family on the basis of larval morphology J Crustac Biol 1988 8 576 593 10.2307/1548694

PMC005xxxxxx/PMC5000412.txt

==== Front BMC Fam PractBMC Fam PractBMC Family Practice1471-2296BioMed Central London 51910.1186/s12875-016-0519-7Research ArticlePatient experiences with family medicine: a longitudinal study after the Dutch health care reforms in 2006 van den Hombergh Pieter p.hombergh@gmail.com 1van Doorn-Klomberg Arna Arna.vanDoornKlomberg@han.nl 2Campbell Stephen stephen.campbell@manchester.ac.uk 34Wensing Michel michel.wensing@radboudumc.nl 15Braspenning Jozé 0031 24 361 5305Joze.Braspenning@radboudumc.nl 11 Scientific Institute for Quality in Healthcare (IQhealth care), Radboud University Medical Center, PO Box 9101, Nijmegen, 6500 HB The Netherlands 2 Master Physician Assistant, HAN University of Applied Sciences, PO Box 9029, Nijmegen, 6500 JK The Netherlands 3 Centre for Primary Care, Institute of Population Health, University of Manchester, Williamson Building, Oxford Road, Manchester, M13 9PL UK 4 Centre for Research and Action in Public Health (CeRAPH), University of Canberra, Building 22, Floor B, University Drive, Bruce, ACT 2617 Australia 5 Department of General Practice and Health Services Research, University Hospital Heidelberg, Marsilius-Arkaden, Turm West, INF 130.3, Heidelberg, 69120 Germany 25 8 2016 25 8 2016 2016 17 1 1181 8 2015 16 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background In 2006 The Dutch Health Care system changed to a market oriented system. The GP remuneration changed from ± 2/3 capitation patients and 1/3 private patients before 2006 to a mixed payment scheme. From 2006 onward every patient was insured and the GP received partly capitation, partly fees for consultations and for specific services. This change coincided with many other organisational changes in General Practice care. Our research question was if during the years after 2006 patient experiences of Dutch family practice had changed. We also wanted to explore the influence of patient and practice characteristics on patient experiences. Data on patient experiences were available from 2007 to 2012. Method In a series of annual cross sectional patient surveys the performance of GPs and practices was measured. Patient sampling took place as a part of the Dutch accreditation program in 1657 practices involving 2966 GPs. Patients’ experiences, gender, age, health status, and number of annual consultations were documented as well as the type and location of practices. Linear regression analysis was used to examine time trends in patient experiences and the impact of patient and practice characteristics. Results 78,985 patients assessed the performance of 2966 GPs, and 45,773 patients assessed the organisation of 1657 practices. The number of patients with positive experiences increased significantly between 2007 and 2012; respectively 4.8 % for GPs (beta 0.20 and p < 0.0001) and 6.6 % for practices (beta 0.10, p < 0.004). Higher age, having no chronic illness, more frequent consultations and attending single-handed practices, predicted better patient experiences. Conclusions In our evaluation of patient experiences with general practice care from 2007 to 2012 we found an increase of 4.8 % for GPs and 6.6 % for practices respectively. This improvement is significant. While no direct causation can be made, possible explanations may be found in the various reforms in Dutch family practice since 2006. More insight is needed into key determinants of this improvement before policymakers and care providers can attribute the improvement to these reforms. Keywords Health care reformPayment systemPatient experienceQuality of carePractice performanceInvestment in Family MedicinePrimary careissue-copyright-statement© The Author(s) 2016 ==== Body Background In 2006 The Dutch Health Care system changed to a market oriented system (Table 1). The GP remuneration changed from a 2/3 capitation and 1/3 private patients before 2006 to a mixed payment scheme. From 2006 onward every patient was insured and the GP received partly capitation, partly fee for consultations and for specific services. Between 2005 and 2007 costs for GP care rose 11,2 % yearly before slowing down to 2.7 % yearly till 2012 [1]. This switch coincided with other organisational changes in general practice care, which were all meant to improve quality of care e.g. an increase in the number of nurse practitioners, incentivizing diagnostic and therapeutic activities, rewarding adherence to guidelines on availability and accessibility. and strengthening vocational training (Table 2). Changes like incentivizing care and increasing staff for chronic disease management improved clinical care in some settings, but whether these changes improved patient experience is unclear [2, 3]. Clinical care and patient experiences are distinct aspects of quality. Information on the patients’ overall evaluation of the quality of family practice care following the changes in The Netherlands was lacking [4]. The data of patient experiences using Europep [5] of a large sample of practices entering the Dutch Practice accreditation scheme each year, offered an opportunity to monitor patient experiences. The study also explored the impact of patient and practice characteristics.Table 1 Changes in primary health care in The Netherlands after 2006 A market-oriented health care system was introduced in 2006 together with a new system of basic health insurance replacing the previous distinction between private and public insurance. Adult citizens are obliged to choose a health insurance, for which they pay around € 1100,- per year (with only slight differences between insurers) plus additional taxation guaranteeing basic health care coverage and free access to the GP, but excluding € 350,- co-payment for specialist care (children’s healthcare is free). It amounts to an average family spending € 11.000,- or around 23 % of its income on healthcare. Insurers got purchasing power and the government withdrew from healthcare, but set strict regulations for insurers and providers. Regulated competition between healthcare providers and between health insurers was introduced for specialist care, but General Practice was exempted from this competition. General Practice income has increased since 2006 and GPs invested in premises, staff and infrastructure, including ICT and communication equipment. Their personal income increased as well. Along with the change to market-oriented financing the total budget for GP-care rose from € 1.922 in 2006 to 2.372 million in 2010, an increase of 14 %. In 2011 all insurers invested another 10 %. Before 2006 the macro budget for GP care had been constant. Until 2006 GPs received capitation payments for their public patients (2/3), and fees per consultation for their private patients (1/3). From January 2006, GPs are being paid according to a uniform, mixed payment scheme. GPs receive a partial capitation for each patient per year plus fees per consultation for basic day care. They receive ancillary payments (mainly on a fee-for-service basis) for additional or special therapeutic and diagnostic services and for the care for chronic diseases. They are compensated on an hourly basis for care during out-of-office-hours (evening, night and weekend care). The income of self-employed Dutch GPs was ± € 96.000 in 2005 and of GPs employed by other GPs € ± 73.000. Table 2 Changes in the practice organisation and in the training of GPs - GPs started to be organized in large care groups contracting chronic care in disease management programs. From 2006 onwards the availability of a nurse practitioner for chronic disease management rose from a few percent to over 90 % (treating chronic diseases: Diabetes Mellitus, Chronic Obstructive Pulmonary Disease, Cardiovascular Disease, Mental Care), accounting for part of the rise in practice income. Between 2007 and 2012 practice nurses’ time rose from 5.5 to 11.0 h per 1000 patients per week. - Broadening the diagnostic and therapeutic scope of the practice followed the selective incentives for extra services (± € 50 per service for minor surgery, spirometry, EKG, Cyriax injection, etc.). - Primary care practices became larger with more GPs working in one center. The number of single handed practices dropped between 2006 and 2012 from 46 % to 39 %. The number of GPs rose slightly from 8612 in 2006 to 8879 in 2012 (3 %) and patients per fte GP decreased slightly. Self-reported GP-time rose from 21 to 28 h per 1000 patients per week. - In 2008, the Dutch Association of Family Medicine (LHV) accepted new guidelines on availability and accessibility. Insurers offered € 4,- for each patient when the guidelines were met. Practices should minimally be open 6 h a day, 5 days a week and address emergency calls by a medically trained person within 30 s. The GP had to visit the emergency patients within 15 min. It was incentivized but also checked by the Dutch Inspection of Health Care and subsequent failure to meet the standard was financially penalized (in practices of > 2500 patients over € 10.000, - could be missed). Only 3 practices finally did not meet the target. (personal communication L. Rijkers, LHV) - A 5 year extensive project to renew the FM-training including training the trainers was completed in 2006 with a focus on assessing and improving consultation skills. The vocational training program of GPs involved ± 1600 trainers and ± 3000 trainees. Nearly half of the GP-population thus got extra education in communication and in treating according to clinical guidelines in a new curriculum of 8 days every year. The Europep questionnaire has proved to be sufficiently responsive to detect changes in countries with various health policy decisions [6]. The study of Petek et al. also used Europep and compared patients with cardiovascular disease in eight European countries in 2009 with a subgroup of patients with self-defined chronic illness from their study in 1998 [7]. It showed no overall trends for the eight countries combined, but some changes in specific countries. Allan at al. looked at the effect of Continuous Quality Improvement (CQI) on patient satisfaction using patient questionnaire data collected in a Patient Participation Program of the RACGP in Australia with a 10 year follow-up (1993–2003) [8]. They found no significant change in satisfaction but the scores showed little variation (often close to 100 % from the start). We did not find other long term studies showing measurable improvement in patient experience following organisational interventions. We therefore hypothesized that we would not find significant changes in patient experience in the years 2007-2012 in our study. Method Study design and setting The data were collected as part of the Dutch accreditation program (NPA) between 2007 and 2012. The study focused on the patient survey that was part of the data collection preceding the Practice visit. Participating in the NPA was voluntary, yet strongly supported by the Dutch College of General Practitioners, incentivized by the insurers and becoming a future reregistration obligation for GP-trainers. The incentives stimulated hundreds of practices to enter the NPA-program each year [9]. The central outcome measures were the scores on the Dutch Europep questionnaire, which measures patient experiences with the GP- and the practice organisation. Patients (>18 years) who visited the practice were asked by the practice assistant to complete the questionnaire in the waiting room before or after the consultation, and to drop it in a sealed box warranting anonymity. The practice staff was asked to make sure that close to 30 questionnaires per GP were returned. The results of the questionnaire were used only for internal feedback for the GP and the team. Participants Participants were 30 invited patients per GP entering the accreditation program. Measures The Europep questionnaire has 23 items; 17 items on the GP performance and 6 items on the practice management [5]. The items of Europep are a concise reflection of what patients view as important aspects of general practice care across European countries. All items use a Likert scale (1 = poor and 5 = excellent), and a sixth option “do not know/ not applicable”. To warrant anonymity no patient characteristics were asked for except gender and age. From 2009 to 2012 the questionnaires had additional questions on percentage with chronic illness and consultation rate to allow more in depth analysis. Europep was validated in several studies and has proved to show relevant variation on all items [10]. Practice and respondents characteristics are in Tables 3 & 4.Table 3 Practice characteristics compared to the national average Practice characteristics Study population, percentage National Average b 2007 2008 2009 2010 2011 2012 Total Practice type  Single handed 26.5 % 23.2 % 29.9 % 31.7 % 27.2 % 12.5 % 24.3 % 39.5 %  Duo or group 73.5 % 76.8 % 70.1 % 68.3 % 72.8 % 87.5 % 75.7 % 60.5 % Urbanization degree - High urbanisation d 45.5 % 42.5 % 43.8 % 42.3 % 44.6 % - 44.2 % 47.7 % - Moderate urbanisation d 44.0 % 37.2 % 43.2 % 46.3 % 42.7 % - 43.1 % 40.7 % - Rural d 10.5 % 20.3 % 13.0 % 11.4 % 12.7 % - 12.7 % 11.6 % Mean N of pats/ practice 4228 4882 4699 4545 4714 5171 4767 4055 a Training practice 52.2 % 42.5 % 48.4 % 69.7 % 66.5 % 53.9 % 57.5 % 33 % c Number of practices 323 323 265 237 230 279 1657 Number of GPs 323 323 435 602 540 743 2966 a Total Dutch population divided by the total number of primary care practices b NIVEL 2010 and Dutch national Compass, 2011 c Capacity committee d urbanization, high: > 1.500 addresses/km2, moderate: 500–1500 addresses/km2, rural: < 500 addresses/km2 Table 4 Characteristics of the respondents compared to the national average Characteristics of the respondents Study population, percentage or average (SD) National average a 2007 2008 2009 2010 2011 2012 Total n - - 8506 17,661 15,695 15,079 56.941 Age - - 51 (16) 51 (17) 52 (17) 52 (17) 52 (17) 39 Percentage women - - 65.5 % 64.8 % 64.5 % 64.4 % 64.6 % 50.5 % Percentage w. chronic illness - - 24.1 % 23.8 % 24.6 % 24.7 % 24.4 % 31.8 % Consultation rate - - 4.3 (4.1) 4.4 (4.3) 4.3 (4.0) 4.3 (4.4) 4.4 (4.3) 4.2 b a NIVEL 2010 and Dutch national Compass, 2011 b CBS (Dutch Central Statistical Office) in 2012 Analysis We assessed the scores per item by calculating the percentage of people with a rating of 4 or 5 on the 5-point Likert scale [11]. The response to the category “do not know/not applicable” was excluded. Two separate means were calculated for the items regarding GP- and practice performance. In a linear regression model using SPSS we evaluated the trend in patient experiences over time (from 2009 to 2012; 2007 and 2008 did not have sufficient data for a linear regression). Associations between time (year) and the total scores on GP- and practice performance were explored. We also analyzed the trend for each of the individual items in separate models. In addition, we used models in which we corrected for patient age, gender, whether they self-reported chronic illness, the consultation frequency of the patients and whether the practice was a single handed practice or not. Results Study population In total 2966 GPs in 1657 practices were included. We excluded 8.0 % of the questionnaires using three exclusion criteria; respondent age (3.8 % was below 18), number of questionnaires per practice (1.0 % had less than 10 questionnaires) and repeated measurements in the same practice (3.2 %). This resulted in 78,985 questionnaires on the performance of 2966 GPs and 45,773 questionnaires on the performance of their 1657 practices. Out of 30 questionnaires on average 27 were completed per GP and 28 per practice. The study practices were reasonably representative for Dutch general practices (Table 3). Training practices were overrepresented (around 50 % with a top of 69 % in 2010). General trend in patient experiences GP-score Overall, more than 80 % of the patients rated the performance of their GP positively on aspects of time, empathy, listening, examining, informing, treatment and advice (Table 5). The positive trend from 82.1 % to 86.9 % (Fig. 1) over 2007–2012 is significant (Beta 0.20 and p < 0.001). Table 5 shows that GPs improved their care from 2009 up to 2012 regarding time for the patient, empathy, shared decision making, communication, thoroughness, patient centeredness and providing information. An analysis which corrected for differences between the cross-sectional samples regarding patient and practice characteristics confirms the positive trend between 2009 and 2012 (Beta 0,10 and p < 0.001).Table 5 Trend of the various aspects of the Europep questionnaire from 2009 to 2012, corrected for patient age, gender, chronic illness, consultfrequency and practice type What is your opinion of the GP and/or general practice over the last 12 months with respect to: Aspect/item Score in 2009 % bèta p Composite GP score, n = 2713 GPs for each item 84.8 .096 .0000 1 making you feel you had time during consultations? 87.6 .085 .0001 2 interest in your personal situation? 83.3 .060 .0078 3 making it easy for you to tell him or her about your problems? 87.5 .080 .0004 4 involving you in decisions about your medical care? 84.9 .081 .0003 5 listening to you 91.4 .067 .0029 6 keeping your records and data confidential? 93.3 .078 .0005 7 quick relief of your symptoms 76.5 .069 .0020 8 helping you to feel well so that you can perform your normal daily activities? 81.0 .107 .0001 9 thoroughness 85.9 .105 .0001 10 physical examination of you? 87.6 .078 .0005 11 offering you services for preventing diseases? 79.3 .054 .0127 12 explaining the purpose of tests & treatments (screening, health checks? 87.0 .082 .0003 13 telling you what you wanted to know about your symptoms and/or illness? 87.7 .081 .0003 14 helping you deal with emotional problems related to your health status? 80.0 .058 .0098 15 helping you understand the importance of following his or her advice? 83.5 .061 .0063 16 knowing what he or she had done or told you during contacts? 81.0 .076 .0007 17 preparing you for what to expect from specialist or hospital care? 75.6 .142 .0001 Composite practice score, n = 1527 practices for each item 67.8 .097 .0039 18 the helpfulness of the staff (other than doctor) to you? 82.7 .106 .0025 19 getting an appointment to suit you? 74.6 .079 .0206 20 getting through to the practice on the telephone? 61.0 .081 .0220 21 being able to talk to the general practitioner on the telephone? 58.6 .041 .2403 22 Waiting time in the waiting room? 48.3 .062 .0670 23 Providing quick services for urgent health problems? 81.8 .114 .0010 Fig. 1 General trend in patient experience with primary care between 2007 and 2012: The grey background (2007 and 2008) are crude scores without correction for patient characteristics. The white background are scores corrected for patient characteristics (see Table 4) Practice score Patients rated the practice organisation slightly less positively than they rated their GP. The rating was 64.9 % in 2007 and 71.5 % in 2012 (rise of 6.6 %) (Fig. 1). This positive trend is significant (beta 0.19, p < 0.001) and was confirmed after correcting for differences between the cross-sectional samples in the period 2009–2012 (Beta 0.10, p < 0.004). Almost all items contribute to this positive trend (Table 5). The best scoring items are on “helpfulness of the staff”, “getting a suitable appointment” and “getting through to the practice on the phone”. Factors associated with patient experiences Table 6 shows which factors are associated with GP performance and practice management. The year of the visit is an important explanatory factor. The trend in the period 2009–2012 in improved patient experience was confirmed after correcting for background variables.Table 6 Linear regression models for both GP performance and Practice performance (2009–2012) Variable GP Performance Practice Performance bèta p bèta p Visitation year .096 .0001 .097 .004 Mean age of the GP .204 .0001 .250 .0001 %female patients .018 .443 −.056 .098 % Patients with chronic illness −.066 .007 −.117 .001 Mean practice consult frequency .181 .0001 .103 .003 Single handed vs other practice .035 .111 .269 .0001 R2 = .078 R2 = .170 N = 78,985 questionnaires of 2966 GPs; N = 45,773 questionnaires of 1657 practices Patient and practice factors which related to patient experiences are consistent across the year cohorts. The GP performance and practice management are rated significantly higher by older compared to younger patients and by patients with a higher frequency of consultations. Patients with a chronic disease rated the GP performance and the practice management less positively than other patients. Patients’ gender and practice urbanisation do not affect the ratings. The practice management was rated higher in single handed practices than in other types of practices, but practice type (group practice, health center) had no effect on GP performance. Discussion Patient experience with Dutch general practice care changed positively in the period 2007–2012 with 4.8 % on GP personal performance and 6.6 % on practice organisation (Fig. 1). The positive change could be demonstrated for each item. The increase followed a period of profound changes in the healthcare system among which ‘investments in primary care’ and ‘the introduction of incentives in the healthcare system’. This positive trend in Dutch patient experiences is reinforced by the yearly results of the European Health Consumers Index (EHCI), that reported the Netherlands to be in the top 3 since 2007 and a first position in 2014 with a 40 points advantage over number 2 in 2014 [12]. Changes in primary care since 2006 that could have affected patient experiences Various bodies of research provide potential determinants for positive change in patient experiences. A Cochrane meta-analysis showed that investment in practice nurses for the treatment of chronic diseases had a positive effect on patient experiences [2]. In our practices the available time of practice nurses rose from 4.8 to 5.7 h per 1000 patients between 2009 and 2011. More time per patient proved to be associated with better patient experience in a previous study that used the same data base of practice visits [13]. A change in reimbursement for the care of chronic patients proved to be effective in the UK for clinical care [14]. But in the evaluation of the pay-for-performance (P4P) program conducted from 2003 to 2007, patients did not experience changes in quality of care for communication, nursing care, coordination, and overall satisfaction [2, 15]. Some aspects of access improved but patients reported seeing their usual physician less often and gave lower satisfaction ratings for continuity of care [3]. Patients highly value the accessibility and availability of general practice [11]. Starting in 2008 Dutch insurers incentivised service with 4 € extra capitation per patient when targets on accessibility and availability were met (Table 1). This could have helped the positive change in patient experiences on access, availability and continuity. In the UK these aspects of care did not change or worsened after P4P was introduced [15]. A positive effect on patient experience may be attributed to the change to ‘fee per consultation’ (9 € ) and ‘home visit’ (14 €) meant to compensate for the proportionate lowering of the capitation (114 € to 57 €). Although GP care was exempted from market forces and competition, the new blended system of capitation and fee for service aligning incentives more closely to professional values may have influenced patient experiences positively [14, 16]. From 2006 onward the scope of diagnostic and therapeutic procedures widened and more procedures were done in the practice also because it was financially attractive. In previous research a wider scope was associated with better medical performance in videotaped patient contacts [17]. Improvement in patient experiences may also be attributed to an extensive project to renew vocational training, which was completed in 2006. The project focused on assessing and improving trainers’ skills in giving feedback, coaching of and assessing the CANMEDS (Canadian Medical Education Directives for Specialists concerning 7 competencies: Medical Expert, Communicator, Collaborator, Manager, Health Advocate, Scholar, and Professional). Dutch GP-trainers score significantly better on quality of care and organisation including patient experiences than non-training GPs [18]. In the UK investment in family medicine training—both in GP-trainers & trainees—improved the score on the GPPS in the P4P program in the UK. GP training practice status (29 % of practices) was a significant predictor of positive GPPS responses to all questions in the ‘doctor care’ (n = 6) and ‘overall satisfaction’ (n = 2) domains but not to any of the ‘nurse care’ or ‘out-of-hours’ domain questions [19]. Doctors in GP training practices appeared to offer more patient-centered care with patients reporting more positively on attributes of doctors such as ‘listening’ or ‘care and concern’. Extra patient time, introducing practice nurses, enhancing accessibility and availability, changing some payments in fee for services, improving consultation skills during vocational training all could have attributed to positive patient experiences. A negative effect could be expected from the increase in practice size between 2007 and 2012. Larger practices with more GPs have less positive experiences with care [6] and many patients prefer smaller practices [20, 21]. Strengths & limitations The data were collected within the Dutch accreditation program with a voluntary participation of practices and with maybe better practices entering first. This selection of the better practices entering first is at odds with the increase of improved patient experience. However, year after year practices who felt ready for the practice visit entered the program. This yearly mix of training and non-training practices is a selection of voluntary practices, but we had expected a negative change in patient experience in the later years, because less ambitious practices entered later in the program. We consider this a strength of this study. Another strength is the large and representative sample of practices that could be analysed and the high numbers of new patients each year that completed the questionnaires. Selection bias due to selecting patients would very likely to have been constant over the years. Unfortunately we could only do a linear regression analysis for the years 2009–2012, because we lacked sufficient data in 2007 & 2008. A point of discussion was that feminisation of the profession could have contributed to the improvement. However, the role of gender was doubtful in other studies [22–24]. Comparison with other studies Previous studies of patient experience did show no change or mixed results after organisational change. All but one of these studies concerned P4P-studies, whereas our study included all patients who had an appointment with their doctor. Our findings that a higher age, a ‘higher frequency of consulting the GP’, ‘having no chronic illness’ and ‘a short waiting time for the consultation’ were associated with more positive patient experiences on GP performance resonate with similar analyses in previous studies [8, 24–26]. Implications The positive change in patient experiences with family practice cannot be related to the interventions in General Practice care. Patient experience was found to be correlated to better clinical quality in hospitals in a review by Price et al. Research indicates that better patient care experiences are associated with higher levels of adherence to recommended prevention and treatment processes, better clinical outcomes, better patient safety within hospitals, and less health care utilisation [27]. In primary care the correlation between patient experience measured with national General Practice Patient Survey (GPPS) and the national pay-for-performance scheme (QOF) was weak. The 2 domains of quality of care remain predominantly distinct [19]. Longitudinal data collection on patient experiences should span longer periods with a standardized and validated instrument such as Europep, to allow comparison over the years. Such data could enable GPs and policymakers to make better choices on practice organisation, e.g. optimal list size, being a training practice, optimal staff, etcetera [28]. Conclusion In our evaluation following the trend in patient experiences from 2007 to 2012 we found an increase of 4.8 % for GPs and 6.6 % for practices respectively. This is considerable given the often reported limited range for improvement in patient experience surveys. Most previous studies of patient experiences over time showed no or mixed results. Though an attribution of the reforms to the improvement of patient experiences is impossible on the basis of this research, it is important to study the changes in Dutch Family Medicine preceding the improved patient experiences . The literature yields as possible contributors to improved patient experience: 1. The introduction of a practice nurse for chronic diseases, 2. incentivizing accessibility and availability, 3. change to a mixed capitation and fee for service payment + incentivizing additional diagnostic and therapeutic services and 4. improvement of the vocational FM-training. Policymakers and professionals could benefit from monitoring patient experiences. Abbreviations CQIContinuous Quality Improvement EHCIEuropean Health Consumers Index FMFamily Medicine GPGeneral Practitioner LHVDutch Association of Family Medicine NHGDutch College of General Practitioners NPADutch Accreditation Program Acknowledgements We thank the Dutch College of General Practitioners (NHG) and NHG Practice Accreditation for collecting and delivering the large amount of data for our study. Funding No funding, IQhealthcare partly employed the authors except Stephen Campbell. Availability of data and materials The raw data belong to the Dutch Practice Accreditation (NPA). Data can be obtained from the NPA on request. The accreditation organization is a member of the ISQUA (The International Society for Quality in Health Care). Authors’ contributions PvdH designed data collection tools of the VIP-method, drafted and revised the paper. AvD helped with the statistical analysis, cleaned and analyzed the data and revised the paper. SC collaborated for many years on implementation research in IQhealthcare, put the paper in international perspective and revised several versions of the paper. MW designed data collection tools of the Europep patient questionnaire, commented on the design of the study and revised all versions of the paper. JB initiated the collaborative project, supervised the project from its early start to the end, coached the participants and revised all versions. All authors accepted the revised paper. All authors read and approved the final manuscript. Competing interest The lead author* affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and registered) have been explained. Consent for publication Not applicable. Ethics approval and consent to participate The ethics committee of the Radboud university medical center provided a waiver for the study. License “The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, a worldwide license to the Publishers and its licensees in perpetuity, in all forms, formats and media (whether known now or created in the future), to 1) publish, reproduce, distribute, display and store the Contribution, 2) translate the Contribution into other languages, create adaptations, reprints, include within collections and create summaries, extracts and/or, abstracts of the Contribution, 3) create any other derivative work(s) based on the Contribution, 4) to exploit all subsidiary rights in the Contribution, 5) the inclusion of electronic links from the Contribution to third party material where-ever it may be located; and, 6) license any third party to do any or all of the above.” ==== Refs References 1. Hendriks M Spreeuwenberg P Rademakers J Delnoij DM Dutch healthcare reform: did it result in performance improvement of health plans? A comparison of consumer experiences over time BMC Health Serv Res 2009 9 167 10.1186/1472-6963-9-167 19761580 2. Martinez-Gonzalez NA Djalali S Tandjung R Huber-Geismann F Markun S Wensing M Rosemann T Substitution of physicians by nurses in primary care: a systematic review and meta-analysis BMC Health Serv Res 2014 14 214 10.1186/1472-6963-14-214 24884763 3. Campbell SM Reeves D Kontopantelis E Sibbald B Roland M Effects of pay for performance on the quality of primary care in England N Engl J Med 2009 361 4 368 378 10.1056/NEJMsa0807651 19625717 4. Rademakers J Delnoij D de Boer D Structure, process or outcome: which contributes most to patients’ overall assessment of healthcare quality? BMJ Qual Saf 2011 20 4 326 331 10.1136/bmjqs.2010.042358 21339310 5. Wensing M EUROPEP 2006: Revised Europep Instrument and User Manual 2006 Nijmegen Centre for Quality of Care Research, Radboud university 6. Wensing M Hermsen J Grol R Szecsenyi J Patient evaluations of accessibility and co-ordination in general practice in Europe Health Expect 2008 11 4 384 390 10.1111/j.1369-7625.2008.00507.x 19076666 7. Petek D Kunzi B Kersnik J Szecsenyi J Wensing M Patients’ evaluations of European general practice--revisited after 11 years Int J Qual Health Care 2011 23 6 621 628 10.1093/intqhc/mzr052 21831966 8. Allan J Schattner P Stocks N Ramsay E Does patient satisfaction of general practice change over a decade? BMC Fam Pract 2009 10 13 10.1186/1471-2296-10-13 19200400 9. van den Hombergh P Grol R van den Hoogen HJ van den Bosch WJ Assessment of management in general practice: validation of a practice visit method BrJGenPract 1998 48 436 1743 1750 10. Grol R Wensing M Mainz J Jung HP Ferreira P Hearnshaw H Hjortdahl P Olesen F Reis S Ribacke M Patients in Europe evaluate general practice care: an international comparison Br J Gen Pract 2000 50 460 882 887 11141874 11. Wensing M Mainz J Ferreira P Hearnshaw H Hjortdahl P Olesen F Reis S Ribacke M Szecsenyi J Grol R General practice care and patients’ priorities in Europe: an international comparison Health Policy (New York) 1998 45 3 175 186 10.1016/S0168-8510(98)00040-2 12. Björnberg: Euro Health Consumer Index 2014. Health Consumer Powerhouse Ltd.,2015. 13. van den Hombergh P Kunzi B Elwyn G van Doremalen J Akkermans R Grol R Wensing M High workload and job stress are associated with lower practice performance in general practice: an observational study in 239 general practices in the Netherlands BMC Health Serv Res 2009 9 118 10.1186/1472-6963-9-118 19604386 14. Roland M Campbell S Successes and failures of pay for performance in the United Kingdom N Engl J Med 2014 370 20 1944 1949 10.1056/NEJMhpr1316051 24827040 15. Campbell SM Kontopantelis E Reeves D Valderas JM Gaehl E Small N Roland MO Changes in patient experiences of primary care during health service reforms in England between 2003 and 2007 Ann Fam Med 2010 8 6 499 506 10.1370/afm.1145 21060119 16. Roland M Incentives must be closely aligned to professional values BMJ 2012 345 e5982 10.1136/bmj.e5982 22963938 17. Ram P Grol R van den Hombergh P Rethans JJ van der Vleuten C Aretz K Structure and process: the relationship between practice management and actual clinical performance in general practice Fam Pract 1998 15 4 354 362 10.1093/fampra/15.4.354 9792351 18. van den Hombergh P Schalk-Soekar S Kramer A Bottema B Campbell S Braspenning J Are family practice trainers and their host practices any better? Comparing practice trainers and non-trainers and their practices BMC Fam Pract 2013 14 23 10.1186/1471-2296-14-23 23433175 19. Llanwarne NR Abel GA Elliott MN Paddison CA Lyratzopoulos G Campbell JL Roland M Relationship between clinical quality and patient experience: analysis of data from the english quality and outcomes framework and the National GP Patient Survey Ann Fam Med 2013 11 5 467 472 10.1370/afm.1514 24019279 20. Põlluste K Kalda R Lember M Primary health care system in transition: the patient’s experience Int J Qual Health Care 2000 12 6 503 509 10.1093/intqhc/12.6.503 11202604 21. Kontopantelis E Roland M Reeves D Patient experience of access to primary care: identification of predictors in a national patient survey BMC Fam Pract 2010 11 61 10.1186/1471-2296-11-61 20799981 22. Rahmqvist M Patient satisfaction in relation to age, health status and other background factors: a model for comparisons of care units Int J Qual Health Care 2001 13 5 385 390 10.1093/intqhc/13.5.385 11669566 23. Pedersen LB Kjaer T Kragstrup J Gyrd-Hansen D Do general practitioners know patients’ preferences? An empirical study on the agency relationship at an aggregate level using a discrete choice experiment Value Health 2012 15 3 514 523 10.1016/j.jval.2012.01.002 22583462 24. Lyratzopoulos G Elliott M Barbiere JM Henderson A Staetsky L Paddison C Campbell J Roland M Understanding ethnic and other socio-demographic differences in patient experience of primary care: evidence from the English General Practice Patient Survey BMJ Qual Saf 2012 21 1 21 29 10.1136/bmjqs-2011-000088 21900695 25. Campbell JL Ramsay J Green J Age, gender, socioeconomic, and ethnic differences in patients’ assessments of primary health care Qual Health Care 2001 10 2 90 5 10.1136/qhc.10.2.90 11389317 26. Paddison C Elliott M Parker R Staetsky L Lyratzopoulos G Campbell JL Roland M Should measures of patient experience in primary care be adjusted for case mix? Evidence from the English General Practice Patient Survey BMJ Qual Saf 2012 21 8 634 640 10.1136/bmjqs-2011-000737 22626735 27. Anhang Price R Elliott MN Zaslavsky AM Hays RD Lehrman WG Rybowski L Edgman-Levitan S Cleary PD Examining the role of patient experience surveys in measuring health care quality Med Care Res Rev 2014 71 5 522 554 10.1177/1077558714541480 25027409 28. van den Hombergh P Campbell S Is ‘practice size’ the key to quality of care? Br J Gen Pract 2013 63 614 459 460 10.3399/bjgp13X671498 23998813

PMC005xxxxxx/PMC5000413.txt

==== Front BMC Infect DisBMC Infect. DisBMC Infectious Diseases1471-2334BioMed Central London 179310.1186/s12879-016-1793-7Research ArticleChanges in the prevalence and biofilm formation of Haemophilus influenzae and Haemophilus parainfluenzae from the respiratory microbiota of patients with sarcoidosis Kosikowska Urszula +48814487106urszula.kosikowska@umlub.pl 1Rybojad Paweł pawel.rybojad@umlub.pl 2Stępień–Pyśniak Dagmara dagmara.stepien@up.lublin.pl 3Żbikowska Anna anna_zbikowska@sggw.pl 4Malm Anna anna.malm@umlub.pl 11 Department of Pharmaceutical Microbiology with Laboratory for Microbiological Diagnostics, Medical University of Lublin, Chodzki Str. 1, 20-093 Lublin, Poland 2 Department of Thoracic Surgery, Medical University of Lublin, Lublin, Poland 3 Sub-Department of Veterinary Prevention and Avian Diseases, Institute of Biological Bases of Animal Diseases, Faculty of Veterinary Medicine, University of Life Sciences in Lublin, Lublin, Poland 4 Department of Food Technology, Faculty of Food Sciences, Warsaw University of Life Sciences (WULS-SGGW), Warsaw, Poland 26 8 2016 26 8 2016 2016 16 1 44917 12 2015 21 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Healthy condition and chronic diseases may be associated with microbiota composition and its properties. The prevalence of respiratory haemophili with respect to their phenotypes including the ability to biofilm formation in patients with sarcoidosis was assayed. Methods Nasopharynx and sputum specimens were taken in 31 patients with sarcoidosis (average age 42.6 ± 13), and nasopharynx specimens were taken in 37 healthy people (average age 44.6 ± 11.6). Haemophili were identified by API-NH microtest and by the matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) system. Biofilm was visualised by crystal violet staining and confocal scanning laser microscopy (CSLM). The statistical analysis was performed with Statgraphics Plus for Windows. Results In total, 30/31 patients with sarcoidosis and 31/37 healthy people were colonized by Haemophilus influenzae (6/30 vs. 1/31) and Haemophilus parainfluenzae (28/30 vs. 31/31) in the nasopharynx. The overall number of nasopharyngeal haemophili isolates was 59 in patients with sarcoidosis and 67 in healthy volunteers (H. influenzae 6/59 vs. 1/67, P = 0.05; H. parainfluenzae 47/59 vs. 65/67, P = 0.0032). Moreover, the decreased number of H. parainfluenzae biofilm-producing isolates was shown in nasopharyngeal samples in patients with sarcoidosis as compared to healthy people (19/31 vs. 57/65, P = 0.006), especially with respect to isolates classified as strong and very strong biofilm-producers (8/31 vs. 39/65, P = 0.002). Conclusions The obtained data suggest that the qualitative and quantitative changes within the respiratory microbiota concerning the overall prevalence of H. influenzae together with the decreased number of H. parainfluenzae strains and the decreased rate of H. parainfluenzae biofilm-producing isolates as compared to healthy people may be associated with sarcoidosis. Keywords SarcoidosisRespiratory microbiotaHaemophilus parainfluenzaeHaemophilus influenzaeBiofilmissue-copyright-statement© The Author(s) 2016 ==== Body Background Sarcoidosis is a chronic and enigmatic multisystem disease involving the lung, heart and the lymphatic system [1, 2]. The etiology of sarcoidosis is not likely to be due to any infections, but rather to an exaggerated and aberrant immune response of genetically susceptible individuals to unidentified antigens, including microorganisms, or several organic and inorganic substances [1–4]. Understanding the role of microbiota composition is a new frontier of human biology and the contemporary direction in the investigation of physiological or pathological phenomena of health or diseases [5–7]. Qualitative and quantitative shifts or perturbation in the microbiota can lead to the development of diseases. Microbiota monitoring and modification may be useful for determination of health, thus providing new means of protection and/or of intervention, and data interpretation [8–10]. The microbiota components predominantly colonizing the respiratory mucosa without causing any disease symptoms can occasionally cause respiratory infections. Besides possible positive or negative interactions between commensals and pathobionts as well as other potential pathogens, microbiota can play an important role for the human host organism in preventing of respiratory and invasive infections [11]. Additionally, microbiota disturbance can contribute to acquisition and carriage of pathogens, can predispose to viral co-infection, especially in people with an immature or damaged immune system. The human-restricted respiratory tract microbiota representatives are Haemophilus influenzae with significant pathogenicity and opportunistic commensal H. parainfluenzae [12, 13]. They may be etiologic agents of invasive or opportunistic diseases [14–16]. H. influenzae, both the encapsulated (mainly serotype b – Hib) and non-encapsulated (nontypeable H. influenzae - NTHi) strains have also been associated as potential pathogens with chronic or recurrent and invasive diseases (e.g. bacteremia or sepsis, otitis media, chronic bronchitis, and community-acquired pneumonia) often reported in children and rarely in adults. H. parainfluenzae, as an opportunistic bacteria, less often reported as an etiologic agent of infectious diseases, may cause systemic or other respiratory infections (e.g. epiglottitis, meningitis, bacteremia or sepsis, bronchitis, chronic obstructive pulmonary disease, and infective endocarditis). The human microbiota is a reservoir of opportunistic and potential pathogens (pathobionts), including haemophili, living mainly in a diverse community of biofilm [17–19]. Biofilm as a structure of microbial community enveloped in a polymeric matrix and adhered to both natural and synthetic surfaces may be regarded as a phenotypic adaptation and protective or pathogenic factor in many infections, depending on the condition [11, 20–22]. It was found to be a form of microbial life important both in colonization and in chronic and recurrent or acute diseases such as otitis media and pneumonia caused by NTHi species [17]. Biofilm is estimated to be involved in about 65 % of human infections with bacterial etiology [23]. Adhesive properties, as well as biofilm formation by microoorganisms together with its intrinsic antimicrobial resistance, exopolysaccharide production and quorum sensing are factors allowing for adaptation to host organism [20]. Both H. influenzae and H. parainfluenzae have been found to be a biofilm-forming bacteria. The objectives of the present study were: the analysis of the correlations of diagnostic results in patients with sarcoidosis based on simple regression, haemophili isolation in nasopharyngeal and sputum specimens, antimicrobial resistance determination in H. influenzae and H. parainfluenzae clinical isolates, biofilm production by clinical isolates of these species together with the analysis of its structure. Methods Patients A group of 31 adult patients (average age 42.6 ± 13) with a suspicion of sarcoidosis who were diagnosed in 2011 at the Chair and Department of Thoracic Surgery (Medical University of Lublin, Poland), participated in the study. The selection criterion was sarcoidosis, which was diagnosed with clinical findings suggesting an incidence of this disease. Patients were directed for diagnosis because of radiological findings such as: lymphadenectomy or tumour of mediastinum, or the presence of small nodules and infiltrations in the lung parenchyma, sclerosis, thickening or fibrosis discovered in CT scans. Multivariable demographic, clinical, radiographic and histological data were collected on the basis of the patients’ questionnaires and information protocol. All patients were diagnosed by means of bronchoscopy, mediastinoscopy or/and lung biopsy. Before the procedure blood samples were collected for standard blood tests (basic metabolic panel and complete blood count). The obtained tissue samples were evaluated by the same pathomorphologist. The histopathological findings were usually described as tuberculosis like granulation which could be considered as sarcoidosis in accordance with clinical changes. A control group of 37 healthy volunteers (average age 44.6 ± 11.6) who agreed to participate in the survey was also included. They did not suffer from respiratory infections and had not received an antimicrobial therapy for at least three months prior to the examination or had not been admitted to hospital for at least two years. Written informed consent for participation was obtained from people who agreed to take part in the study and filled out the survey. The Ethics Committee of the Medical University of Lublin approved study protocol (KE-0254/75/2011). Microbiological processing of haemophili isolates A total of 31 nasopharyngeal swabs and 31 sputum specimens were taken from patients with sarcoidosis on the day of hospitalization or a day after. Additionally, 37 nasopharyngeal specimens were collected from healthy people. After incubation (48 h, 35 ± 2 °C, 5 % CO2) the colonies with morphological differences were identified independently on selective HAEM-medium (Haemophilus-chocolate-agar, bioMérieux, France). The growth of bacteria in the form of individual colonies or from abundant to very abundant number of morphologically different colonies on Chocolate agar was observed. Initially biochemical identification and biotyping of 192 Gram-negative isolates (125 – from patients with sarcoidosis and 67 – from healthy people) was carried out using the API-NH microtest (bioMérieux). The phenotypes of haemophili isolates were differentiated based on various observable properties in the growth morphology (e.g. the shape and size of the colony, smooth or rough surface, texture, colony elevation), on a set of biochemical reactions (according to API NH results) and antimicrobial susceptibility results. API-NH is a standardized system for the identification of Neisseria, Haemophilus (and related genera) and Moraxella (B.) catarrhalis, which uses microtests and a specially adapted database. Next, for the species differentiation the matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS, Bruker Daltonik, Germany) system was used according to the procedure described earlier [24]. Software used for data acquisition was MALDI-Biotyper 3.0 (Bruker Daltonik) software. The species which were not identified as H. influenzae or H. parainfluenzae were classified as other Haemophilus spp. Antibiotic sensitivities were determined by the disc diffusion method using Haemophilus-Test-Medium (HTM, Oxoid) according to [25]. Direct colony suspensions standardized to 0.5 McFarland standard were prepared using colonies from an overnight HAEM incubation (35 °C, 5 % CO2). H. influenzae ATCC10211 was used to verify the growth of bacteria on HTM medium. Different discs with antimicrobial agents (BD BBL™, Becton Dickinson and Company, USA) namely ampicillin 10 μg, amoxicillin/clavulanate 20/10 μg, ampicillin/sulbactam 10/10 μg, cefuroxime 30 μg, cefotaxime 30 μg, ceftazidime 30 μg, imipenem 10 μg, aztreoname 30 μg, azithromycin 15 μg, tetracycline 30 μg, trimethoprim/sulfamethoxazole 1.25/23.75 μg, ciprofloxacin 5 μg were used. Resistance (R) zone diameter interpretative criteria were ≤18 mm for ampicillin, ≤19 mm for amoxicillin-clavulanate and ampicillin-sulbactam, ≤16 mm for cefuroxime, ≤25 mm for cefotaxime and ceftazidime, ≤15 mm for imipenem, ≤ 25 mm for aztreoname, ≤11 mm for azithromycin, ≤25 mm for tetracycline, ≤10 mm for trimethoprim/sulfamethoxazole, ≤20 mm for ciprofloxacin, according to [25]. Multidrug-resistant (MDR) isolates were defined as having resistance to at least three different classes of antimicrobials. Beta-lactamase activity was screened by Pen-microtest (part of API-NH strip) and confirmed with a nitrocefin-based test (Cefinase™-Discs, BD BBL™) recommended for testing haemophili [25]. Then, E-test strips (bioMérieux) for a determination of susceptibility on the basis of the minimal inhibitory concentration of ampicillin (MICAm) were used. The MIC criterion, defined as the lowest concentration of the antimicrobial agent that prevents visible growth of the microorganism, for ampicillin-resitant haemophili was ≥ 4 μg/ml [25]. Biofilm detection Biofilm formation was examined during the stationary culture in vitro in 24-well polystyrene microplates (24 F-Well Microplates, Thermo Scientific™ Nunc™, Denmark) using a 0.1 % crystal violet (CV) stain as previously described [26]. The Tripticasein Soy Broth (TSB, Biocorp, Poland) supplemented with Haemophilus Test-Medium Supplement (HTMS, Oxoid) designated as TSB + HTMS was used. Overnight cultures were diluted in TSB + HTMS-medium and standardized at 570 nm with an initial optical density of OD570 ~ 0.08 ± 0.02 (~0.5 McFarland-standard) using a microplate reader ELx800 (BioTek Inc., USA). Next, 500 μl of the microbial suspension was inoculated for each well and incubated (35 °C, 24 h, 5 % CO2). The growth of haemophili was assessed by measuring the OD570. Nonadherent cells were removed by rinsing the wells with sterile water. The biofilm was detected with OD570 according to the method based on staining with 500 μl 0.1 % CV. Each isolate was tested in triplicate in three series. TSB + HTMS without bacteria was incubated under the same conditions and served as blank control. Haemophili were classified as biofilm-producers as described elsewhere [26]. The experiments were performed in triplicate and the results were averaged. For the purposes of a detailed analysis of the obtained results the classification of biofilm producers was introduced on the basis of criteria proposed by Stepanović et al. [27] and modified by Kosikowska et al. [26]. They defined the cut-off an optic density OD (ODc) for the microtiter-plate test as three standard deviations above the mean OD of the negative control. The bacteria were classified as follows: OD ≤ ODc - non-producers (category 0); ODc < OD ≤ 2 x ODc - weak producers (category 1); 2 x ODc < OD ≤ 4 x ODc – moderate producers (category 2); 4 x ODc < OD ≤ 8 x ODc – strong producers (category 3), 8 x ODc < OD – very strong producers (category 4). ODc was 0.1 ± 0.02 in our experiments. Additionally, H. influenzae ATCC10211, H. parainfluenzae ATCC7901, H. parainfluenzae ATCC33392 and H. parainfluenzae ATCC51505 reference strains were used as biofilm producers. All tests were carried out three times and the results were averaged. To visualize H. influenzae and H. parainfluenzae biofilm formation in a 24 h culture in 24-well polystyrene microplates, the inverted microscope Axiovert 200 M equipped with LSM5 Pascal Head (Carl Zeiss, Germany, with magnification 200×) was used. To obtain images of the biofilm, cultures were stained with Bacterial Live/Dead®BacLight™-L7012-kit (Invitrogen, USA) accordingly to the manufacturer’s procedure. Biofilm was stained with Live/Dead® L7012 kit with two components, Syto9 and propidium iodide (PI), which are nucleic acid stains. The Syto9 stain penetrates through the intact and damaged cell wall, intercalates into DNA, and emits green fluorescence (detection of both live and dead bacteria). The PI stain diffuses only across the dead and damaged cell wall, intercalates into DNA and emits red fluorescence (detection of dead bacteria). The use of a combination of these two dyes is that PI displaces Syto9 (it had a higher affinity) reducing its fluorescence. Thus, the bacteria with intact cell membranes stain fluorescent green, whereas with damaged cell membranes stain fluorescent red. Biofilm formation by 7 H. influenzae and 18 H. parainfluenzae randomly selected clinical isolates taken in patients with sarcoidosis was screened using confocal microscopy. The thickness of biofilm (μm) and measurement area covered by biofilm (%), content of live cells (%) and biofilm area covered by live cells (%) were detected. Overnight cultures were standardized in TSB + HTMS-medium. Then 500 μl/well of cultures was added (for each strain in quadruplicate). After incubation (35 °C, 24 h, 5 % CO2) the content of the wells was removed and each well was washed three-four times with 0.85 % NaCl. Then, 500 μl/well of 0.85 % NaCl was added and then stained with 1.5 μl of Live/Dead-kit solution. Microphotographs were taken in the green and/or red channel in a Confocal Laser Scanning Microscope LSM5-PASCAL (CLSM; Carl Zeiss, Germany) after incubation for 15 min in the dark at room temperature. This experiment was repeated twice. The planimetric measurement of the biofilm was performed based on the microphotographs taken at 50× or 200× magnification in a two-dimensional scan (2D). Biofilm parameters were calculated using ImageJ-1.43e software (Wayne Rasband, National Institutes of Health, USA). A three-dimensional (3D) image of biofilm was reconstructed using the CLSM (200× magnification). Statistical analysis The statistical analysis was performed with Statgraphics Plus for Windows, Version 4.1 (Statistical Graphics Corp. 1999, Statpoint Technologies, Inc. Warrenton, Virginia, USA). To assess the relationship between the variables a simple regression analysis was used. The aim of the statistical analysis was preliminary determination of the relationship between individual factors and the dependent variables. The relationship between age/gender and different clinical manifestations and laboratory findings were studied. The hypotheses were raised: H0 - there is no association between age/gender and different clinical manifestations and laboratory findings, and H1 - there is an association between age/gender and different clinical manifestations and laboratory findings. Quantitative variables were presented as mean (± standard deviation, SD) and median values. In some studies Fisher’s test was evaluated. The level of P < 0.05 was usually considered as statistically significant. Results Characteristics of patients The main characteristics of patients with sarcoidosis are summarized in Table 1. Demographic and clinical data were obtained from the patients’ files and based on a questionnaire conducted for the presented variables. “Fatigue” was self reported as being unable to go any further. Susceptibility to “fatigue” was associated with a short walk on a flat surface, or with stair walk to the first or second floors. Symptoms such as shortness of breath, wheezing, cough, expectoration, hypertension, coronary artery disease, allergy, asthma, and recurrent infections (≥3/year) were also considered. Almost 26 % of the patients with sarcoidosis showed recurrent infections in childhood and had other symptoms such as anaemia, celiac disease, or endocrinological disorders. Some of them had two parallel disorders, for example hypertension and coronary artery disease (about 13 %), and/or allergy (about 10 %).Table 1 Baseline clinical characteristics, laboratory findings, and predisposing factors in the group of patients with sarcoidosis Variable No. of cases/No. of patients or range (%) or average ± SD Gender  Female 13/31 41.9  Male 18/31 58.1 Age (years) 24–64 42.6 ± 13.0 Tobacco smoke  Active 9/31 29.0  Secondary 20/31 64.5 Area of residence  Rural 15/31 48.4  Urban 16/31 51.6 Healthy condition  Shortness of breath, wheezing 12/31 38.7  Cough 16/31 51.6  Expectoration 4/31 12.9  Fatigue: 22/31 71.0   a short walk (n = 1) 1/22 4.5   1st floor (n = 4) 4/22 18.2   2nd floor (n = 17) 17/22 77.3  Hypertension, coronary artery disease 4/31 12.9  Allergy 3/31 9.7  Asthma 1/31 3.2  Recurrent infection (≥3/year) 1/31 3.2  Recurrent infections in childhood 8/31 25.8  Other (anaemia, celiac disease, thyroid disorders) 8/31 25.8  Tuberculosis 2/31 6.5 Laboratory findings  FVC (ml) 2740–6200 4208.1 ± 883.2  FEV1 (ml) 1600–5050 3231 ± 810.3  FEV1/FVC% 58–95 76.5 ± 8.7  pH 7.35–7.49 7.43 ± 0.03  pO2 (mmHg) 63.2–108 82.6 ± 10.5  pCO2 (mmHg) 32.8–49.1 95.4 ± 2.7  Saturation (%) 90.9–97.3 95.4 ± 1.8  IgG (g/l, reference: 7.0–16.0) 7.1–21.6 10.8 ± 2.7  IgM (g/l, reference: 0.4–2.3) 0.28–3.54 1.38 ± 0.8  IgA (g/l, reference: 0.7–4.0) 0.51–6.9 2.57 ± 1.3  CRP (mg/l, reference: 0.0–5.0): 0.51–80.48 7.23 ± 16.7   0.0–5.0 (n = 25) 0.51–4.15 1.51 ± 0.93   5.1–80.48 (n = 6) 5.8–80.48 31.10 ± 28.83  Changes in CT scan 30/30 100  Changes in X-ray 19/20 95  Changes observed during bronchofiberoscopy 14/17 82.4 Data were presented as: No. of positive samples/No. of examinations or value range; mean ± SD Abbreviations: CT computed tomography, CRP C-reactive protein, FEV 1 forced expiratory volume in one second, FVC forced vital capacity, IgA, IgM, IgG immunoglobulin A, M, G, NP nasopharynx, SP sputum; susceptibility to “fatigue” was associated with a short walk on a flat surface, or with stair walk to the first or second floors. “Fatigue” was self reported as being unable to go any further There was a positive correlation (P < 0.1) between gender and a tumor of the lung/mediastinum (c = 0.4) or fluid in the pleural cavity (c = 0.32) in patients with sarcoidosis (Table 2). These changes mostly occurred in males rather than in females. A positive correlation was demonstrated between gender and CRP (C-reactive protein; P = 0.01, c = 0.44), and in elderly patients between the disease and visible changes (e.g. sclerosis, thickening or fibrosis) in the imaging diagnostics of the chest (P < 0.1, c = 0.34–0.45). A negative correlation between age and IgM (P < 0.1, c = -0.4) values was detected. Besides, changes in levels of IgG and IgA immunoglobulins were observed.Table 2 Patient characteristics by age and gender correlations with clinical conditions, laboratory findings, and complications Tests Test results Age (years) Gender P value Correlation P value Correlation Clinical manifestation  Sclerotic changes in lung 0.84 0.04 0.64 0.09  Enlargement of the mediastinal lymph nodes 0.45 −0.14 0.9 0.02  Tuberculosis like findings 0.23 0.23 0.23 0.23 Chest radiography  Micronodules 0.81 −0.05 0.63 0.12  Pulmonary shading/infiltrations 0.05 0.45* 0.28 −0.26  Tumor of the lung/mediastinum 0.36 0.22 0.08 0.4* Chest CT  Changes sclerosis/thickening/fibrosis 0.61 0.22 0.64 0.04  Lymphadenopathy 0.06 0.34* 0.16 −0.27  Fluid in the pleural cavity 0.44 0.15 0.09 0.32* Changes observed during bronchofiberoscopy  Infiltration of bronchial wall/macrophages 0.38 0.23 0.66 −0.12  Scraps of bronchial epithelial/mucus 0.28 0.28 0.39 0.22  Easy bleeding capillaries 0.17 −0.35 0.33 0.25 Laboratory findings  pH 0.77 −0.06 0.68 −0.08  IgG (g/l) 0.74 −0.06 0.41 0.15  IgM (g/l) 0.25 −0.4* 0.27 −0.2  IgA (g/l) 0.17 0.25 0.47 0.14  CRP (mg/l) 0.21 0.23 0.01 0.44** Abbreviations: CT computed tomography, CRP C-reactive protein The significance levels were marked as: *P < 0.1, and **P < 0.05) The prevalence of haemophili-positive clinical samples In total, 30/31 (96.8 %) patients with sarcoidosis were colonized by H. influenzae and/or H. parainfluenzae. H. influenzae was isolated in 14/30 (46.7 %) patients - in 1/14 (7.1 %) only in nasopharynx, in 8/14 (57.1 %) only in the sputum and in 5/14 (35.7 %) patients in both samples (Fig. 1a). H. parainfluenzae was isolated in 30/30 (100 %) patients - in 2/30 (6.7 %) only in nasopharynx, in 2/30 (6.7 %) only in the sputum and in 26/30 (86.6 %) cases in both samples (Fig. 1b). The co-existence of H. influenzae and H. parainfluenzae was found in 14/30 (46.7 %) patients.Fig. 1 The prevalence of Haemophilus influenzae (a) and Haemophilus parainfluenzae (b) positive clinical samples in patients with sarcoidosis In comparison, nasopharynx in 31/37 (83.8 %) of healthy people was colonized by both species: in 1/31 (3.2 %) - by H. influenzae, in 31/31 (100 %) - by H. parainfluenzae. There were significant differences in the frequency of nasopharynx colonization by H. influenzae (P = 0.05), but not by H. parainfluenzae (P = 0.49) in patients with sarcoidosis and in healthy people. Negative correlations (a simple regression) were found between H. influenzae presence in the sputum and O2 saturation (P < 0.01, c = -0.52) or easily bleeding capillaries (P < 0.05, c = -0.53), and between H. influenzae presence both in the nasopharynx and in the sputum and the sclerotic changes in lung (P < 0.05, c = -0.4 and c = -0.44, respectively). Positive correlations were found between H. parainfluenzae isolation in the sputum and the presence of fluid in the pleural cavity (P < 0.01, c = 0.56), pH value (P < 0.05, c = 0.39), or O2 saturation (P < 0.1, c = 0.32) and tuberculosis like findings (P < 0.1, c = 0.35). Negative correlations were found between H. parainfluenzae isolation in the sputum and FVC (Forced Vital Capacity; P < 0.1, c = -0.35). Number of haemophili isolates From one to six different phenotypes of haemophili were isolated in the nasopharynx and/or in the sputum samples in each single patient. The growth morphology and biochemical features distinguished phenotypes of isolates. In total, 125 haemophili isolates were found in 62 nasopharyngeal swabs and sputum samples obtained in patients with sarcoidosis. In 30 haemophili-positive nasopharyngeal samples 59/125 (47.2 %) isolates were identified: 6/59 (10.2 %) as H. influenzae, 47/59 (79.7 %) as H. parainfluenzae and 6/59 (10.2 %) as other Haemophilus spp. In 30 haemophili-positive samples taken in the sputum, 66/125 (52.8 %) isolates were identified: 13/66 (19.7 %) as H. influenzae, 49/66 (74.2 %) as H. parainfluenzae and 4/66 (6.1 %) as other Haemophilus spp. Likewise, in 31 haemophili-positive samples obtained in the nasopharynx of healthy people 67 isolates were identified: 1/67 (1.5 %) as H. influenzae, 65/67 (97.0 %) as H. parainfluenzae, and 1/67 (1.5 %) as other Haemophilus spp. H. parainfluenzae was the main species isolated from the nasopharynx in both groups. There were statistically significant differences between the number of H. influenzae (P = 0.05) or H. parainfluenzae (P = 0.0032) isolates in patients with sarcoidosis and in healthy people. Biotypes of H. influenzae and H. parainfluenzae isolates A total of 19 H. influenzae and 96 H. parainfluenzae strains isolated from patients with sarcoidosis were assigned to eight biotypes. Biotypes III and VI occurred in 47.4 % and 26.3 % H. influenzae isolates respectively, mainly in the sputum (Table 3). Biotypes I (60.4 %) and II (30.2 %) constituted most of the H. parainfluenzae isolates both in the nasopharynx and in the sputum. For comparison, thedifferences were observed in nasopharyngeal H. parainfluenzae biotypes I-IV identified in healthy people and in patients with sarcoidosis (biotype I: 58.5 % vs. 26 %, P = 0.6998; biotype II: 20 % vs. 16 %, P = 0.1864; biotype III: 12.3 % vs. 5.2 %, P = 1.000 and biotype IV: 6.2 % vs. 1 %, P = 0.397, respectively).Table 3 Distribution of Haemophilus influenzae and Haemophilus parainfluenzae biotypes in healthy people and in patients with sarcoidosis Healthy people Patients with sarcoidosis Biotype Nasopharynx Nasopharynx Sputum Total No. (%) of isolates Haemophilus influenzae n = 1 n = 19  II 0 (0) 3 (15.8) 0 (0) 3 (15.8)  III 0 (0) 2 (10.5) 7 (36.8) 9 (47.4)  VI 1 (100) 1 (5.3) 4 (21.1) 5 (26.3)  VIII 0 (0) 0 (0) 2 (10.5) 2 (10.5) Haemophilus parainfluenzae n = 65 n = 96  I 38 (58.5) 25 (26.0) 33 (34.4) 58 (60.4)  II 13 (20.0) 15 (15.6) 14 (14.6) 29 (30.2)  III 8 (12.3) 5 (5.2) 1 (1.04) 6 (6.3)  IV 4 (6.2) 1 (1.04) 0 (0) 1 (1.04)  VI 1 (1.5) 1 (1.04) 0 (0) 1 (1.04)  VII 0 (0) 0 (0) 1 (1.04) 1 (1.04)  VIII 1 (1.5) 0 (0) 0 (0) 0 (0) Antimicrobials sensitivity of H. influenzae and H. parainfluenzae isolates As shown in Table 4, only 4/19 (21.1 %) H. influenzae isolates from patients with sarcoidosis were resistant to the antimicrobials - 2/19 (10.5 %) were trimethoprim/sulfamethoxazole-resistant and 2/19 (10.5 %) were tetracycline-resistant (Table 4). In contrast, 53/96 (55.2 %) H. parainfluenzae isolates were resistant to the antimicrobials – 18/96 (18.8 %) were trimethoprim/sulfamethoxazole-resistant and 15/96 (15.6 %) were tetracycline-resistant. In addition, 20/96 (20.9 %) H. parainfluenzae isolates were resistant to beta-lactams, including 8/96 (8.3 %) ampicillin-resistant ones.Table 4 Distribution of antimicrobial resistance in Haemophilus influenzae and Haemophilus parainfluenzae isolates in healthy people and in patients with sarcoidosis No. (%) of resistant isolates Species Place of isolation Sxt Te Ctx Caz Sam AmC Am Healthy people Haemophilus influenzae (n = 1) Nasopharynx (n = 1) 0 0 0 0 0 0 0 Haemophilus parainfluenzae (n = 65) Nasopharynx (n = 65) 8 (12.3) 4 (6.2) 4 (6.2) 5 (7.7) 3 (4.6) 1 (1.5) 10 (15.4) Patients with sarcoidosis Haemophilus influenzae (n = 19) Nasopharynx (n = 6) 1 (16.7) 0 0 0 0 1 (16.7) 0 Sputum (n = 13) 1 (7.7) 2 (15.4) 0 0 0 0 0 Haemophilus parainfluenzae (n = 96) Nasopharynx (n = 47) 10 (21.3) 9 (19.1) 2 (4.3) 3 (6.4) 0 0 6 (12.8) Sputum (n = 49) 8 (16.3) 6 (12.2) 1 (2.0) 5 (10.2) 0 0 2 (4.1) Abbreviations: Am ampicillin, AmC amoxycillin/clavulanate, Caz ceftazidime, Ctx cefotaxime, Sam ampicillin-sulbactam, Sxt trimethoprim/sulfametoxazole, Te tetracycline The presence of nasopharyngeal H. parainfluenzae isolates resistant to tetracycline (4/65; 6.2 %) and to trimethoprim/sulfametoxazole (8/65; 12.3 %) was shown in healthy people. In addition, 23/65 (35.4 %) isolates were resistant to beta-lactams, including 10/65 (15.4 %) ampicillin-resistant ones. The difference in presence of nasopharyngeal H. parainfluenzae isolates resistant to tetracycline in patients with sarcoidosis and in healthy people (19.1 % vs. 6.2 %, P = 0.041) was statistically significant. The differences in resistance to trimethoprim/sulfametoxazole (21.3 % vs. 12.3 %, P = 0.297) and to beta-lactams (23.4 % vs. 35.3 %, P = 0.214), including ampicillin (12.8 % vs. 15.4 %, P = 1.000) were also found, but they were not statistically significant. The ampicillin-resistant isolates selected in patients with sarcoidosis (MICAm ≥ 6 μg/ml) were beta-lactamase-positive, and were found within I and III biotypes (Table 5). High MICs for ampicillin were detected in two isolates taken in the nasopharynx (MIC = 128 and >256 μg/ml), and in one isolate taken in the sputum (MIC = 48 μg/ml). Among H. parainfluenzae isolates, 3/96 (3.1 %) isolates were resistant to beta-lactams, tetracycline, and trimethoprim/sulfametoxazole (MDR-strains).Table 5 Distribution of beta-lactamase positive ampicillin-resistant Haemophilus parainfluenzae isolates in patients with sarcoidosis Biotype Nasopharynx Sputum Resistance profile (No. of isolates) MICAm (μg/ml) Resistance profile (No. of isolates) MICAm (μg/ml) I Am (1) 24 AmTe (1) 4 AmSxt (1) >256 AmTe (1) 48 AmCtxCazSxt (1) 8 AmTeSxt (1) 128 III AmCtxCazTeSxt (1) 6 AmTeSxt (1) 6 Abbreviations: Am ampicillin, Caz ceftazidime, Ctx cefotaxime, Te tetracycline, Sxt trimethoprim/sulfametoxazole Ampicillin breakpoints used for interpretation of minimal inhibitory concentration (MICAm): susceptible ≤1 μg/ml, intermediate = 2 μg/ml and resistant ≥ 4 μg/ml, according to [25] Biofilm formation by H. influenzae and H. parainfluenzae isolates In total, 69 randomly selected H. influenzae and H. parainfluenzae isolates (35 - nasopharyngeal, 34 - from the sputum) taken in patients with sarcoidosis were screened for biofilm-production using the CV method. According to Table 6, 8/12 (66.7 %) isolates of H. influenzae and 39/57 (68.4 %) of H. parainfluenzae were found to be biofilm-producers (P = 0.407). All 8/12 (66.7 %) H. influenzae isolates were weak-producers, irrespective of the clinical specimens. Among 19 biofilm-forming H. parainfluenzae isolates from nasopharynx, the ability to biofilm formation ranged from weak (11/19, 57.8 %) to strong (1/19, 5.3 %) and very strong (7/19, 36.8 %).Table 6 Distribution of Haemophilus influenzae and Haemophilus parainfluenzae biofilm-producers in patients with sarcoidosis and in healthy people Category of biofilm producers OD570 range No. (%) of isolates Haemophilus influenzae Haemophilus parainfluenzae Healthy people Nasopharynx (n = 1) Nasopharynx (n = 65) Non-producers 0.0 0 (0.0) 8 (12.3) Weak 0.01–0.24 1 (100) 15 (23.1) Moderate 0.25–0.48 0 (0.0) 3 (4.6) Strong 0.49–0.96 0 (0.0) 12 (18.5) Very strong ≥0.97 0 (0.0) 27 (41.5) Patients with sarcoidosis Nasopharynx (n = 4) Sputum (n = 8) Nasopharynx (n = 31) Sputum (n = 26) Non-producers 0.0 1 (25.0) 3 (37.5) 12 (38.7) 6 (23.1) Weak 0.01–0.24 3 (75.0) 5 (62.5) 11 (35.5) 9 (34.6) Moderate 0.25–0.48 0 (0.0) 0 (0.0) 0 (0.0) 3 (11.5) Strong 0.49–0.96 0 (0.0) 0 (0.0) 1 (3.2) 4 (15.4) Very strong ≥0.97 0 (0.0) 0 (0.0) 7 (22.6) 4 (15.4) Distribution of biofilm-producers was detected using the crystal-violet (CV) method. The cut-off OD (ODc; here: 0.1 ± 0.02) was defined as three standard deviations above the mean OD570 of the negative control. The categories of biofilm-producers: 0 - non-producers (OD ≤ ODc); 1, weak (ODc < OD ≤ 2xODc); 2, moderate (2xODc < OD ≤ 4xODc); 3, strong (4xODc < OD ≤ 8xODc); 4, very strong (8xODc < OD) producers, according to [26] Additionally, reference strains were included as biofilm-producers: H. influenzae ATCC10211as weak biofilm-producer (OD570 = 0.098 ± 0.05), H. parainfluenzae ATCC7901 as strong biofilm-producer (OD570 = 0.682 ± 0.07), and both H. parainfluenzae ATCC33392 and H. parainfluenzae ATCC51505 as very strong biofilm-producers (OD570 = 1.191 ± 0.04 and OD570 = 1.026 ± 0.08, respectively). In 65 H. parainfluenzae isolates obtained from nasopharynx of healthy people, in 57/65 (87.7 %) isolates the ability for biofilm formation was identified (Table 6). Among 57 biofilm-positive H. parainfluenzae isolates 15/57 (26.3 %), 3/57 (5.3 %), 12/57 (21.1 %), and 27/57 (47.4 %) ones were classified as weak, moderate, strong or very strong biofilm-producers, respectively. Significant differences between the ability to biofilm production in nasopharyngeal H. parainfluenzae isolates taken in patients with sarcoidosis and in healthy people were shown (P = 0.006). Among biofilm-positive H. parainfluenzae isolates taken in cases and in healthy people significant differences were observed between weak biofilm-producers and the group of isolates classified as moderate, strong and very strong biofilm-producers (P = 0.024). Morphometric parameters of biofilm formed by 7 H. influenzae and 18 H. parainfluenzae isolates taken from the patients with sarcoidosis were assessed by means of CLSM technique (Table 7). The biofilm formed by H. influenzae isolates had the thickness of 14.2 ± 3.5 μm and covered 71.6 ± 5.6 % of the measured area. The content of living cells was from 8.6 to 95.2 % (average 71.6 ± 4.4 %) and it was 75.2 ± 4.8 % in the biofilm area. The biofilm formed by H. parainfluenzae isolates had the thickness of 20.02 ± 4.3 μm and covered about 50.6 ± 4.8 % of the measured area. The content of living cells was from 42.8 to 99.9 % (average 80.9 ± 4.8 %) and it was about 77.9 ± 8.98 % in the biofilm area.Table 7 Morphometric parameters of biofilm formed by Haemophilus influenzae and Haemophilus parainfluenzae isolates taken from patients with sarcoidosis Haemophili isolates selected from Thickness of biofilm ± SD (μm) Measurement area covered by biofilm (%) ± SD Content of live cells (%) in biofilm ± SD Biofilm area covered by live cells (%) ± SD Haemophilus influenzae Nasopharynx (n = 3) 11.3 ± 3.2 68.9 ± 2.5 58.4 ± 2.97 66.99 ± 9.1 Sputum (n = 4) 16.4 ± 3.7 73.6 ± 7.96 81.5 ± 5.4 81.4 ± 1.6 All (n = 7) 14.2 ± 3.5 71.6 ± 5.6 71.6 ± 4.4 75.2 ± 4.8 Haemophilus parainfluenzae Nasopharynx (n = 10) 20.5 ± 4.9 47.1 ± 5.9 76.8 ± 4.2 72.5 ± 12.5 Sputum (n = 8) 19.4 ± 3.7 54.95 ± 3.5 86.0 ± 5.4 84.6 ± 4.7 Total (n = 18) 20.0 ± 4.3 50.58 ± 4.8 80.9 ± 4.8 77.9 ± 8.98 Averages morphometric parameters of biofilm formed by Haemophilus influenzae and Haemophilus parainfluenzae isolates was calculated on the basis of data assessed using confocal laser scanning microscopy (CLSM) images after 24 h incubation The biofilm formed by one of H. parainfluenzae nasopharyngeal isolate taken from a patient with sarcoidosis was revealed by CLSM image (Fig. 2). The biofilm area formed by living (Fig. 2a) and dead (Fig. 2c) cells is presented. The textural parameters were detected as the grey scale intensity of the biofilm formed by live (Fig. 2b) and dead (Fig. 2d) cells.Fig. 2 Biofilm formed by living and dead cells of Haemophilus parainfluenzae SHpiNP18C strain. Expanations: Biofilm formed by living (a, b) and dead (c, d) cells of Haemophilus parainfluenzae SHpiNP18C and detected using two-dimensional (2d: a, c) or three-dimensional (3d: b, d) CLSM image with gray level values varying from 0 to 250 (b, d) after 24 h of incubation. The layer of living cells was located on or between the structures formed by dead cells. The living and dead cells forming biofilm were detected in the green and red canals, respectively, and the structure of the biofilm was shown in XZ plan on the basis of 3D scan Discussion Several environmental agents interacting with social and genetic factors and including immune responses to microbial components rather than an infection per se have been considered to play a role in the pathogenesis of sarcoidosis [1, 2, 4]. In our study changes in the healthy conditions, in the imaging diagnostics of the chest and in levels of three major immunoglobulins (IgG, IgM and IgA) as well as CRP value with respect to sarcoidosis and the age and gender were shown (Tables 1 and 2). Similar observations were also done by other authors [28–31]. Hiperglobulinemia is frequently observed in patients with sarcoidosis [30]. Changes observed in this group of patients [29, 30] and in elderly people [31] suggested an association between age, race, sex and the immune system’s defense as well as microbials presence. According to Buckley et al. [29], in patients with sarcoidosis the increase in IgG was significant in white patients, and the increase in IgM concentration was significant only in black patients, especially in black woman. Cagatay et al. [32] found higher than normal laboratory values for immunoglobulins IgG, IgA and IgM. They have noted that IgG and IgA levels were significantly higher in the group of routinely checked patients without any clinical symptoms and during established the activity of the disease. Drent et al. [33] have identified a high CRP concentration associated with severe fatigue in sarcoidosis. The human microbiota has multidirectional effects on the host’s health and changes in its composition may have important consequences for human pathophysiology and disease development [5, 7, 8, 34]. According to literature, bacterial 16S rRNA from H. influenzae as well as Moraxella catarrhalis was detected in sarcoid fluid samples [35]. Cantwell’s review [3] pointed to the presence of other bacterial species (e.g. Mycobacterium spp., staphylococci, Propionibacterium acnes) in the biopsy of sarcoid tissue, blood, and skin samples taken from patients with sarcoidosis. According to our results, higher rate of nasopharyngeal colonization by H. influenzae, but not by H. parainfluenzae (P = 0.05 vs. P = 0.49) was found in patients with sarcoidosis as compared to colonization in healthy people. It suggests that sarcoidosis can be regarded as a factor predisposing for colonization by H. influenzae. Besides, H. influenzae was frequently found in the sputum samples taken in patients with sarcoidosis even if this bacterial species was absent in the nasopharynx. In our opinion, it suggests the role of H. influenzae in the lower respiratory tract colonization and inflammation particularly in chronic diseases, e.g. in patients with sarcoidosis. In the literature [36], it has been documented that a combination of pathogenic mechanisms of bacteria and defects in host defense may allow this species to their migration into the lower respiratory tract. This may result in chronic colonization and/or in acute exacerbations of airway disease. Despite the high and a similar number of people colonized by H. parainfluenzae in the nasopharynx of healthy people and patients with sarcoidosis (83.8 % vs. 96.8 %), we observed a decreased number of this species isolates with different phenotypes, differentiated on the basis of growth morphology, biochemical characteristics and biotypes. In 30 patients with sarcoidosis 47 isolates of H. parainfluenzae were identified, while in 31 healthy people - 65 isolates of this species. H. parainfluenzae isolates with biotypes I and II were found to occur most frequently, similarly in patients with sarcoidosis and in healthy people (Table 3). As found by other authors [37–39], these biotypes constituted most of H. parainfluenzae isolates in patients with other respiratory diseases such as chronic bronchitis or cystic fibrosis. During our studies we compared the antimicrobial susceptibility of nasopharyngeal H. parainfluenzae isolates obtained in patients with sarcoidosis and in healthy people (Table 4). Resistance of these isolates to tetracycline (P = 0.041) and trimethoprim/sulfametoxazole (P = 0.297) in patients with sarcoidosis was higher compared to that in healthy people. In contrast, resistance to beta-lactams (P = 0.214), including ampicillin (P = 1.000) was lower in patients with sarcoidosis compared to that in healthy people. These differences may be due to higher consumption of a given group of antimicrobials in a defined population. The growing antibiotic resistance and reduction or elimination their effectiveness is one of the world’s most pressing public health problems [40]. Beta-lactam antibiotics are the most widely used antimicrobial agents during treatment of both community-acquired and hospital infections. The resistance to this group of antimicrobials in Haemophilus spp. usually is mediated by the production of beta-lactamases and the presence of altered penicillin-binding protein (PBP) with lowered affinity for these antibiotics as a target site [17, 41]. The ampicillin-resistant, beta-lactamase positive H. parainfluenzae isolates from patients with sarcoidosis were found in our studies (Tables 4 and 5). This may have some implications including the possibility to exchange resistance genes within microorganisms [40–45]. According to literature, over the past years many authors detected beta-lactamases mainly in H. influenzae and rarely in H. parainfluenzae isolates taken from patients with respiratory tract infections as well as from healthy people [25, 26, 46]. It was shown that DNA mutation and rapid multiplication as well as transformation can be important mechanisms in the spread of drug resistance in haemophili, including the ampicillin resistance due to beta-lactamase production [45, 46]. It seems that especially efficient in transformation were H. parainfluenzae cells with a highest ability to develop competence and transfer of resistance genes occurs via free DNA (from dead or lysed cells) during natural transformation from the medium by competent cells. It may explain the acquisition of resistance or resistance gene exchange with other microorganisms. Besides, resistance and reduced susceptibility to beta-lactams mediated by altered PBPs is also important in many bacterial pathogens, including beta-lactamase negative H. influenzae [47]. For this reason, there are different events that may contribute to the emerge of resistance: the acquisition of resistance genes (e.g. beta-lactamases) by conjugation or transformation; and inter-species recombination of the ftsI gene [47, 48]. According to Gromkova et al. [46], most efficient in transformation among H. parainfluenzae strains was biotype II, followed by biotype I. It was shown in this paper that H. parainfluenzae isolates selected in patients with sarcoidosis compared to isolates selected in healthy people had a lower ability for biofilm production (Table 6). It is possible that a reduction in the number of H. parainfluenzae strains capable of biofilm formation may contribute to an increased colonization by certain opportunistic pathogens like H. influenzae (the present results) or by other bacteria [3, 4]. The CLSM technique revealed that biofilms formed in vitro by H. parainfluenzae and H. influenzae isolates taken from patients with sarcoidosis (Table 7, Fig. 2) had high content of live cells (average 72 to 81 %), suggesting the possibility of bacterial persistence and dispersal in vivo. Biofilm may be regarded as a pathogenic or protective factor depending on the conditions [18, 21]. Nontypeable H. influenzae [NTHi] biofilms were observed for both bacteria colonizing the tissue of human, as an etiologic agent which causes the infection or exacerbation of chronic respiratory diseases [49–51]. On the other hand, Clancy and Dunkley [52] showed that oral NTHi could enhance the mucosal protection and prevent exacerbations of a chronic obstructive pulmonary disease. On the basis of our results we propose, that haemophili, mainly H. influenzae and H. parainfluenzae, would be microorganisms indicative in respiratory microbiota changes as well as healthy condition in patients with chronic diseases. We observed higher frequency of H. influenzae colonization in patients with sarcoidosis compared to healthy people (P = 0.05). All H. influenzae isolates were weak- or non-biofilm producers independently to source of isolation and peoples’ health condition. Moreover, despite the high prevalence of H. parainfluenzae in the nasopharynx of people from both studied groups (P = 0.49), less number of isolates of this species obtained in nasopharyngeal samples in patients with sarcoidosis as compared to healthy people were classified as biofilm-producers (61.3 % vs. 87.7 %, P = 0.006), especially as strong and very strong biofilm-producers (25.8 % vs. 60 %, P = 0.002). Conclusions The obtained results suggest that sarcoidosis, associated with many different factors, may be partially due to the respiratory microbiota condition. This is the first study describing qualitative and quantitative changes in the respiratory microbiota in patients with sarcoidosis with the respect to H. influenzae and H. parainfluenzae biotypes and their ability for biofilm formation. The question whether the biofilm formed by these bacterial species is a causative or just a protective factor in recurrent or chronic diseases, e.g. sarcoidosis, requires further studies. Abbreviations AmAmpicillin AmCAmoxycillin/clavulanate CazCeftazidime CRPC-reactive protein CSLMConfocal scanning laser microscopy CTComputed tomography CtxCefotaxime CVCrystal violet FEV1Forced Expiratory Volume/one second FVCForced Vital Capacity HAEMHaemophilus-chocolate-agar HTMHaemophilus Test Medium HTMSHaemophilus-Test-Medium-Supplement MALDI-TOF MSMatrix-assisted laser desorption/ionization time-of-flight mass spectrometry MDRMultidrug-resistant MICMinimal inhibitory concentration NPNasopharynx SamAmpicillin/sulbactam SPSputum SxtTrimethoprim/sulfametoxazole TeTetracycline TSBTripticasein-Soy-Broth TSB + HTMSTripticasein-Soy-Broth supplemented with Haemophilus-Test-Medium-Supplement We thank the director of the Department of Epizootiology and Clinic of Infectious Diseases, Faculty of Veterinary Medicine, University of Life Sciences in Lublin, Stanisław Winiarczyk, for his support. The authors are indebted to Beata Wojtysiak-Duma for performing serological tests, Mr Tomasz Piersiak for his participation in CLSM study, and Dorota Pietras-Ożga for her participation in MALDI-TOF MS study. Funding No external funding was obtained for this case report. Availability of data and materials All data contained within the manuscript. Authors’ contributions UK conceived, designed and coordinated the study, collected and identified haemophili species, measured biofilm, analysed the data, wrote and drafted the manuscript. PR collected patients’ protocols and the nasopharynx and sputum specimens. DSP contributed to collecting specimens and identified haemophili isolates using MALDI-TOF MS system and manuscript review. AŻ statistically analysed the data, AM analysed the data and critically revised the manuscript. All co-authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate All procedures performed in studies were in accordance with the ethical standards of the institutional research committee. The Ethics Committee of the Medical University of Lublin approved study protocol (KE-0254/75/2011). Written informed consent for participation was obtained from people who agreed to take part in the study at the time of data collection and filled out the survey. Respondents were assured of confidentiality. The study investigators were the only persons allowed to inform and counsel the volunteers after the results were obtained. ==== Refs References 1. Criado E Sánchez M Ramírez J Arguis P de Caralt TM Perea RJ Pulmonary sarcoidosis: typical and atypical manifestations at high-resolution CT with pathologic correlation Radio Graphics 2010 30 6 1567 86 2. Newman LS Rose CS Bresnitz EA Rossman MD Barnard J Frederick M A case control etiologic study of sarcoidosis: environmental and occupational risk factors Am J Respir Crit Care Med 2004 170 12 1324 30 10.1164/rccm.200402-249OC 15347561 3. Cantwell AR Bacteria in sarcoidosis and a rationale for antibiotic therapy in this disease JOIMR 2003 1 5 1 4. Chen ES Moller DR Etiology of sarcoidosis Clin Chest Med 2008 29 3 365 77 10.1016/j.ccm.2008.03.011 18539232 5. Hui AWH Lau HW Chan THT Tsui SKW The human microbiota: a new direction in the investigation of thoracic diseases J Thorac Dis 2013 5 Suppl 2 S127 S131 23977433 6. Ivanov II Littman DR Modulation of immune homeostasis by commensal bacteria Curr Opin Microbiol 2011 14 1 106 14 10.1016/j.mib.2010.12.003 21215684 7. Rogers GB Shaw D Marsh RL Carrol MP Serisier DJ Bruce KD Respiratory microbiota: addressing clinical questions, informing clinical practice Thorax 2015 70 74 81 10.1136/thoraxjnl-2014-205826 25035125 8. Piccolo M De Angelis M Lauriero G Montemurno E Di Cagno R Gesualdo L Salivary microbiota associated with immunoglobulin A nephropathy Microb Ecol 2015 70 2 557 65 10.1007/s00248-015-0592-9 25763757 9. Hilty M Burke C Pedro H Cardenas P Bush A Bossley C Disordered microbial communities in asthmatic airways PLoS One 2010 5 1 e8578 10.1371/journal.pone.0008578 20052417 10. Sekirov I Russell SL Antunes CL Finlay BB Gut microbiota in health and disease Physiol Rev 2010 90 3 859 904 10.1152/physrev.00045.2009 20664075 11. Bosch AATM Biesbroek G Trzcinski K Sanders EAM Bogaert D Viral and bacterial interactions in the upper respiratory tract PLoS Pathog 2013 9 1 e1003057 10.1371/journal.ppat.1003057 23326226 12. Cerquetti M de Filippis I McKee ML Molecular typing in bacterial infections Haemophilus spp Molecular typing in bacterial infections 2013 New York Humana Press 193 210 13. Nørskov-Lauritsen N Classification, identification, and clinical significance of Haemophilus and Aggregatibacter species with host specificity for humans Clin Microbiol Rev 2014 27 2 214 40 10.1128/CMR.00103-13 24696434 14. Kilian M Murray PR Baron EJ Jorgensen JH Landry ML Pfaller MA Haemophilus Manual of clinical microbiology 2007 Washington American Society of Microbiology 636 48 15. Middleton AM Dowling RB Mitchell JL Watanabe S Rutman A Pritchard K Haemophilus parainfluenzae infection of respiratory mucosa Respir Med 2003 97 4 375 81 10.1053/rmed.2002.1454 12693797 16. Hill SL Mitchell JL Stockley RA Wilson R The Role of Haemophilus parainfluenzae in COPD Chest 2000 117 5, suppl 1 293S 10.1378/chest.117.5_suppl_1.293S 10843959 17. García-Cobos S Moscoso M Pumarola F Arroyo M Lara N Pérez-Vázquez M Frequent carriage of resistance mechanisms to β-lactams and biofilm formation in Haemophilus influenzae causing treatment failure and recurrent otitis media in young children J Antimicrob Chemother 2014 69 9 2394 9 10.1093/jac/dku158 24891432 18. Bjarnsholt T The role of bacterial biofilms in chronic infections APMIS 2013 121 suppl 1 51 10.1111/apm.12099 23030626 19. Starner TD Zhang N Kim G Apicella A McCray PB Haemophilus influenzae forms biofilms on airway epithelia: implications in cystic fibrosis Am J Respir Crit Care Med 2006 174 2 213 20 10.1164/rccm.200509-1459OC 16675778 20. Cullen L McClean S Bacterial adaptation during chronic respiratory infections Pathogens 2015 4 66 89 10.3390/pathogens4010066 25738646 21. Burmølle M Thomsen TR Fazli M Murree-Allen K Hensley MJ Saunders NA Biofilms in chronic infections - a matter of opportunity - monospecies biofilms in multispecies infections FEMS Immunol Med Microbiol 2010 59 3 324 36 10.1111/j.1574-695X.2010.00714.x 20602635 22. Gilley RP Orihuela CJ Pneumococci in biofilms are non-invasive: implications on nasopharyngeal colonization Front Cell Infect Microbiol 2014 4 163 10.3389/fcimb.2014.00163 25414838 23. Centers for Disease Control and Prevention (CDC) Active bacterial core surveillance (ABCs) report, emerging infections program network, Haemophilus influenzae 2010 2012 24. Kosikowska U Stępien-Pyśniak D Ożga D Wernicki A Malm A Identification of Bacillus spp. colonizing the nasal mucosa of healthy adults living in the suburban area using the matrix-assisted laser desorption-ionization time-of-flight mass spectrometry (MALDI-TOF MS) system Curr Issues Pharm Med Sci 2014 27 3 137 41 25. Clinical and Laboratory Standards Institute (CLSI) Performance standards for antimicrobial susceptibility testing; twenty-fourth informational supplement; M100-S24E 2014 Wayne Clinical and Laboratory Standards Institute 26. Kosikowska U Korona-Głowniak I Niedzielski A Malm A Nasopharyngeal and adenoid colonization by Haemophilus influenzae and Haemophilus parainfluenzae in children undergoing adenoidectomy and the ability of bacterial isolates to biofilm production Medicine 2015 94 pe799 10.1097/MD.0000000000000799 27. Stepanović S Vukovic D Dakic I Savic B Svabic-Vlahovic M A modified microtiter-plate test for quantification of staphylococcal biofilm formation J Microbiol Methods 2000 40 175 9 10.1016/S0167-7012(00)00122-6 10699673 28. Arnold DF Wiggins J Cunningham-Rundles C Misbah SA Chapel HM Granulomatous disease: Distinguishing primary antibody disease from sarcoidosis Clin Immunol (Orlando, Fla) 2008 128 1 18 22 10.1016/j.clim.2008.03.510 29. Buckley CE Dorsey FC A comparison of serum immunoglobulin concentrations in sarcoidosis and tuberculosis Ann Intern Med 1970 72 37 42 10.7326/0003-4819-72-1-37 4188603 30. Buckley CE Trayer HR Serum IgD concentrations in sarcoidosis and tuberculosis Clin Exp Immunol 1972 10 257 65 4625797 31. Tavares CSMQM Bravo Junior WL Alves Leite JL Analyze the levels of immunoglobulins IgG and IgM in elderly and youngs Int J Immunol Immunother 2015 2 1 32. Cagatay T Bilir M Gulbaran M Papila C Cagatay P The immunoglobulin and complement levels in the active pulmonary sarcoidosis Kobe J Med Sci 2003 49 99 106 15141145 33. Drent M Wirnsberger RM de Vries J van Dieijen-Visser MP Wouters EF Schols AM Association of fatigue with an acute phase response in sarcoidosis Eur Respir J 1999 13 4 718 22 10.1034/j.1399-3003.1999.13d03.x 10362029 34. Segal LN Rom WN Weiden MD Lung microbiome for clinicians. New discoveries about bugs in healthy and diseased lungs Ann Am Thorac Soc 2014 11 1 108 16 10.1513/AnnalsATS.201310-339FR 24460444 35. Kelly DM Endotoxin up-regulates interleukin-18: potential role for gram-negative colonization in sarcoidosis Am J Respir Crit Care Med 2005 172 10 1299 1307 10.1164/rccm.200411-1594OC 16100009 36. King P Haemophilus influenzae and the lung (Haemophilus and the lung) Clin Transl Med 2012 1 10 10.1186/2001-1326-1-10 23369277 37. Rhind GB Gould GA Ahmad F Croughan MJ Calder MA Haemophilus biotypes in respiratory disease Thorax 1987 42 151 2 10.1136/thx.42.2.151 3433240 38. Taylor DC Cripps AW Clancy RL Murree-Allen K Hensley MJ Saunders NA Biotypes of Haemophilus parainfluenzae from the respiratory secretions in chronic bronchitis J Med Microbiol 1992 36 4 279 82 10.1099/00222615-36-4-279 1560450 39. Watson KC Kerr EJ Hinks CA Distribution of biotypes of Haemophilus influenzae and H. parainfluenzae in patients with cystic fibrosis J Clin Pathol 1985 38 7 750 3 10.1136/jcp.38.7.750 3874885 40. Carlet J Jarlier V Harbarth S Voss A Goossens H Pittet D Ready for a world without antibiotics? The Pensières Antibiotic Resistance Call to Action Antimicrob Resist Infect Control 2012 1 11 10.1186/2047-2994-1-11 22958833 41. Garcia-Cobos S Arroyo M Campos J Pérez-Vázquez M Aracil B Cercenado E Novel mechanisms of resistance to beta-lactam antibiotics in Haemophilus parainfluenzae : beta-lactamase-negative ampicillin resistance and inhibitor-resistant TEM beta-lactamases J Antimicrob Chemother 2013 68 5 1054 9 10.1093/jac/dks525 23335113 42. Thomas CM Nielsen KM Mechanisms of, and barriers to, horizontal gene transfer between bacteria Nat Rev Microbiol 2005 3 9 711 21 10.1038/nrmicro1234 16138099 43. Pfeifer Y Meisinger I Brechtel K Grobner S Emergence of a multidrug-resistant Haemophilus influenzae strain causing chronic pneumonia in a patient with common variable immunodeficiency Microb Drug Resist 2013 19 1 1 5 10.1089/mdr.2012.0060 23095085 44. Tinguely R Seiffert SN Furrer H Perretenb V Droza S Endimiania A Emergence of extensively drug-resistant Haemophilus parainfluenzae in Switzerland Antimicrob Agents Chemother 2013 57 6 2867 9 10.1128/AAC.00221-13 23545526 45. Tristram SG Pitout MJ Forward K Campbell S Nichols S Davidson RJ Characterization of extended-spectrum beta-lactamase-producing isolates of Haemophilus parainfluenzae J Antimicrob Chemother 2008 61 509 14 10.1093/jac/dkm523 18245789 46. Gromkova RC Mottalini TC Dove MG Genetic transformation in Haemophilus parainfluenzae clinical isolates Curr Microbiol 1998 37 2 123 6 10.1007/s002849900349 9662612 47. Søndergaard A Witherden EA Nørskov-Lauritsen N Tristram SG Interspecies transfer of the penicillin-binding protein 3-encoding gene ftsI between Haemophilus influenzae and Haemophilus haemolyticus can confer reduced susceptibility to beta-lactam antimicrobial agents Antimicrob Agents Chemother 2015 59 4339 42 10.1128/AAC.04854-14 25918135 48. Witherden EA Bajanca-Lavado MP Tristram SG Nunes A Role of inter-species recombination of the ftsI gene in the dissemination of altered penicillin-binding-protein-3-mediated resistance in Haemophilus influenzae and Haemophilus haemolyticus J Antimicrob Chemother 2014 69 1501 9 10.1093/jac/dku022 24562614 49. Moxon ER Sweetman WA Deadman ME Ferguson DJ Hood DW Haemophilus influenzae biofilms: hypothesis or fact? Trends Microbiol 2008 16 3 95 100 10.1016/j.tim.2007.12.005 18280163 50. Swords WE Nontypeable Haemophilus influenzae biofilms: role in chronic airway infections Front Cell Infect Microbiol 2012 25 2 97 22919686 51. Yonker LM Cigana C Hurley BP Bragonzi A Host-pathogen interplay in the respiratory environment of cystic fibrosis J Cyst Fibros 2015 14 4 431 9 10.1016/j.jcf.2015.02.008 25800687 52. Clancy RL Dunkley ML Oral non-typable Haemophilus influenzae enhances physiological mechanism of airways protection Clin Exp Immunol 2010 161 1 127 33 20408861

PMC005xxxxxx/PMC5000414.txt

==== Front BMC Complement Altern MedBMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 130710.1186/s12906-016-1307-6Research ArticleTurmeric extract and its active compound, curcumin, protect against chronic CCl4-induced liver damage by enhancing antioxidation Lee Hwa-Young 1Kim Seung-Wook 2Lee Geum-Hwa 1Choi Min-Kyung 1Jung Han-Wool 1Kim Young-Jun 3Kwon Ho-Jeong 4Chae Han-Jung 82-63-270-3092hjchae@chonbuk.ac.kr 11 Department of Pharmacology and New Drug Development Institute, Chonbuk National University Medical School, Jeonju, Chonbuk 561-180 Republic of Korea 2 CS1 Center, Ottogi Research Center, Ottogi Corporation, Kyeonggi-do, Republic of Korea 3 Food Safety Center, Ottogi Corp, 49 Heungan-daero 395 beon-gil, Dongan-gu, Anyang-si, Gyeonggi-do 14060 South Korea 4 Chemical Genomics National Research Laboratory, Department of Biotechnology, Translational Research Center for Protein Function Control, College of Life Science and Biotechnology, Yonsei University, Seoul, 120-752 Republic of Korea 26 8 2016 26 8 2016 2016 16 1 31615 3 2016 20 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Curcumin, a major active component of turmeric, has previously been reported to alleviate liver damage. Here, we investigated the mechanism by which turmeric and curcumin protect the liver against carbon tetrachloride (CCl4)-induced injury in rats. We hypothesized that turmeric extract and curcumin protect the liver from CCl4-induced liver injury by reducing oxidative stress, inhibiting lipid peroxidation, and increasing glutathione peroxidase activation. Methods Chronic hepatic stress was induced by a single intraperitoneal injection of CCl4 (0.1 ml/kg body weight) into rats. Turmeric extracts and curcumin were administered once a day for 4 weeks at three dose levels (100, 200, and 300 mg/kg/day). We performed ALT and AST also measured of total lipid, triglyceride, cholesterol levels, and lipid peroxidation. Result We found that turmeric extract and curcumin significantly protect against liver injury by decreasing the activities of serum aspartate aminotransferase and alanine aminotransferase and by improving the hepatic glutathione content, leading to a reduced level of lipid peroxidase. Conclusions Our data suggest that turmeric extract and curcumin protect the liver from chronic CCl4-induced injury in rats by suppressing hepatic oxidative stress. Therefore, turmeric extract and curcumin are potential therapeutic antioxidant agents for the treatment of hepatic disease. Electronic supplementary material The online version of this article (doi:10.1186/s12906-016-1307-6) contains supplementary material, which is available to authorized users. Keywords CurcuminCarbon tetrachlorideOxidative stressGlutathioneLipid peroxidaseissue-copyright-statement© The Author(s) 2016 ==== Body Background Liver diseases constitute a major global problem. Carbon tetrachloride (CCl4) is a well-known hepatotoxin that is widely used to induce acute or chronic toxic liver injury in a large range of laboratory animals [1, 2]. Chronic hepatotoxicity induces oxidative damage, necrosis, and inflammation in the liver [3]. The best-characterized model of xenobiotic-induced free radical-mediated liver disease is CCl4-induced liver damage in rats [4]. A recent study reported that CCl4 induces hepatic damage via reactive oxygen species (ROS) [5]. Elevated ROS production, together with inhibition of the antioxidant system, can generate a state of oxidative stress that leads to cell damage [6]. Curcuma longa Linn (CLL, turmeric), an herb widely farmed in Asia, is a primary constituent of a traditional Chinese medicine [7] that has been used effectively for centuries to treat liver diseases in China. Curcumin, best known as a yellow pigment in CLL, has been found to have antioxidant, anti-inflammatory, anti-hepatotoxic, and anti-cancer properties [8–10]. Recently, curcumin has been indicated as a potential treatment for liver damage through mediation of various signaling pathways. It decreased the expression of pro-inflammatory mediators through down regulation of toll-like receptor 4 (TLR4) and TLR2 expression in CCl4-induced rat model of fibrogenesis [11] and Curcumin could also remarkably attenuate the severity of CCl4-induced liver damage through suppression of TGF-β1/Smad signaling pathway and CTGF expression [12]. Another studies demonstrated that curcumin inhibited activation of HSC in vitro by reducing cell proliferation, inducing apoptosis and suppressing ECM gene expression [13]. Here we investigated the role of turmeric and its active compound, curcumin, in liver protection in a rat model of CCl4-induced damage. We hypothesized that turmeric extract and curcumin protect the liver from CCl4-induced damage by reducing oxidative stress, decreasing lipid peroxidation, and suppressing glutathione activation. We demonstrate that the hepatoprotective mechanisms of turmeric extract and curcumin involve the regulation of oxidative stress. Methods Materials CCl4 and Curcumin were purchased from Sigma-Aldrich (St. Louis, MO, USA). Malondialdehyde (MDA) and glutathione (GSH) detection kits were obtained from BioVision (Mountain View, CA, USA). Aspartate aminotransferase (AST), alanine aminotransferase (ALT), total cholesterol, triglyceride, HDL cholesterol, and LDL-cholesterol detection kits were obtained from Asan Pharmaceutical Company (Seoul, Korea). Animal Treatment and Care Seventy male Sprague-Dawley rats weighing 250–270 g were obtained from central lab. animal Inc. (Seoul, Korea) and divided into 7 groups. Rats were intraperitoneally (i.p.) injected with a mixture of CCl4 (0.1 mL/100 g body weight) and olive oil [1:1(v/v)] every other day for 4 weeks. Rats were orally administered turmeric at doses of 100, 200, and 300 mg/kg body weight. Curcumin was given once daily at a dose of 200 mg/kg. All rats were fed a chow diet and kept at 22–23 °C under a 12 h dark/light cycle. The control animals were handled similarly, including i.p. injection with the same volume of olive oil and oral administration of the same volume of PBS. After the last CCl4 injection, the rats were anesthetized with diethyl ether (Sigma) and sacrificed. All animal procedures in this study were performed in accordance with the regulations described in the Care and Use of Laboratory Animals guide of Chonbuk National University. All procedures were also approved by the Institutional Animal Care and Use Committee of Chonbuk National University for the animal center (IACUC protocol CBU 150608-25). Sample collection Liver and blood samples were collected from all sacrificed animals. Whole blood was immediately placed on ice in a centrifuge tube for 30 min and spun by centrifugation at 7,168 × g for 10 min. Serum was transferred to 1.5 mL tubes and stored at −75 °C. All harvested liver tissue samples were immediately stored at -75 °C. Histologic analysis Liver samples were fixed in 10 % formalin and embedded in paraffin. Liver sections were incubated for 10 min in 0.5 % thiosemicarbazide, stained with 0.1 % Sirius red F3B in saturated picric acid for 1 h, and washed with acetic acid (0.5 %). Sections were visualized using a Nikon Eclipse E600 microscope (Kawasaki, Kanagawa, Japan) at 40 × magnification, and relative areas of fibrosis (% positive areas for Sirius red staining) were quantified by histomorphometry using a computerized image analysis system (AnalySIS, Soft Imaging System, Munster, Germany). Hepatic steatosis was assessed by Oil Red O staining. Briefly, liver cryosections were fixed for 10 min in 60 % isopropanol, followed by staining with 0.3 % Oil Red O in 60 % isopropanol for 30 min and washing with 60 % isopropanol. Sections were counterstained with Gill’s hematoxylin, washed with 4 % acetic acid, and mounted in an aqueous solution. Stained sections were quantified by histomorphometry. Blood biochemical marker assays Serum ALT and AST activities were measured by a colorimetric procedure using commercially available detection kits. Measurement of total lipid, triglyceride, and cholesterol levels For lipid determination, liver homogenates were extracted according to the modified Bligh and Dyer procedure [14]. Briefly, samples were homogenized with chloroform-methanol-water (8:4:3), shaken at 37 °C for 1 h, and spun by centrifugation at 1,100 × g for 10 min. The bottom layer was collected for hepatic lipid analysis. Triglyceride, total cholesterol, and total lipid contents were measured using kits from Asan Pharmaceutical Company in accordance with the manufacturer’s instructions. Analyses of lipid peroxidation Lipid peroxidation was assessed in liver tissue using a lipid hydroperoxide assay kit purchased from Cayman Chemicals (Ann Arbor, MI, USA). In this assay, lipid hydroperoxide was extracted from the samples into chloroform using the extraction buffer provided by the manufacturer. The chromogenic reaction was carried out at room temperature for 5 min and the absorbance of each well was read at 500 nm using a 96 well plate spectrometer (SpectraMax 190). 13-Hydroperoxy-octadecadienoic acid was used as the standard. The cellular levels of lipid hydroperoxide were calculated as described by the manufacturer. Measurement of glutathione The levels of hepatic glutathione (GSH) and oxidized GSH (GSSG) were determined using a commercially available kit (BioVision). Briefly, tissue was homogenized and sonicated in 0.5 ml of ice cold buffer [50 mM MES (pH 7.0) with 1 mM EDTA] and then spun by centrifugation at 10,000 × g for 15 min at 4 °C. The supernatant was collected and deproteinized using the reagent supplied in the kit. Next, 50 μL of sample were mixed with 150 μl of assay cocktail in each well and incubated in the dark on a plate shaker. After 30 min, the absorbance of each well was measured at 404 nm. The concentration of total GSH was calculated according to the equation in the manufacturer’s protocol. Sirius red collagen staining Liver tissue was fixed with 10 % formalin and embedded in paraffin. Thin sections were deparaffinized and stained with picro-Sirius red for 60 min at room temperature. Sections on slides were dehydrated sequentially in 100 % ethanol and xylene, and then mounted in Permount (Thermo Fisher Scientific, Waltham, MA, USA). Representative views of liver sections are shown. Statistical analysis Results are presented as the mean ± SEM. The MicroCal Origin software (Northampton, MA, USA) was used for all statistical calculations. Differences were tested for significance using one-way analysis of variance (ANOVA) with Duncan’s multiple range test. Statistical significance was set at p < 0.05. Results Validation of the HPLC method Linearity The slopes, y-intercepts, and correlation coefficients (r2) obtained from regression analysis were as follows: The calibration curves were linear in the tested concentration ranges. The regression equations were y = 47.53x – 1.02 (r2 = 0.9999), for bisdemethoxycurcumin (BDMC); y = 44.92x – 14.62 (r2 = 0.9991), for demethoxycurcumin (DMC); and y = 45.52x – 67.42 (r2 = 0.9992), for curcumin, respectively. The correlation coefficients were all greater than 0.999, indicating high degrees of correlation and good linearity of the method. Precision and accuracy The intra-day and inter-day precision and accuracy for determination of curcuminoids are given in Table 1. The % RSD values for intra-day precision were 0.13, 0.10, and 0.13 for curcumin, DMC, and BDMC, respectively, and those for inter-day precision were 0.30, 0.48, and 0.59, respectively. The low values of % RSD (<0.59 %) reflect the high precision of the method. The percentage recoveries for intra-day accuracy were 95.84 ± 0.12, 98.15 ± 0.09, and 101.08 ± 0.13, respectively, and those for inter-day accuracy were 95.78 ± 0.29, 97.96 ± 0.47, and 101.74 ± 0.60, respectively. All percentage recoveries were within 95.78 – 101.083 %, indicating the good accuracy of the method.Table 1 The intra-day and inter-day precision and accuracy for determination of curcuminoids Present (g/100 g) Added (g/100 g) Found (g/100 g) Accuracy (%) Precision (RSD%) Intra-day  Bisdemethoxycurcumin 0.095 0.096 0.097 ± 0.0001 101.08 0.13  Demethoxycurcumin 0.247 0.262 0.257 ± 0.0003 98.15 0.10  Curcumin 0.993 1.066 1.022 ± 0.002 95.84 0.13 Inter-day  Bisdemethoxycurcumin 0.095 0.096 0.098 ± 0.001 101.74 0.59  Demethoxycurcumin 0.247 0.262 0.257 ± 0.001 97.96 0.48  Curcumin 0.993 1.066 1.021 ± 0.003 95.78 0.30 Analysis of compounds in turmeric extract HPLC analysis was used to identify the two major compounds in turmeric extract as curcumin (901.63 ± 5.37 mg/100 g), BDMC (108.28 ± 2.89 mg/100 g), and DMC (234.85 ± 1.85 mg/100 g) curcuminoids (1244.76 ± 3.86 mg/100 g) (Table 2). The representative HPLC chromatogram is presented in Additional file 1: Figure S1.Table 2 Analysis of turmeric extract Component mg/100 g Curcumin 901.6 ± 5.4 Curcuminoids 1244.8 ± 3.9 BDMC 108.3 ± 2.9 DMC 234.9 ± 1.9 Turmeric extract and its active compound, curcumin, regulate body weight during chronic CCl4 exposure After 4 weeks, the total body weight of each rat was measured. The average weight gain of rats in the control group was normal. In contrast, rats in the chronic CCl4 exposure group exhibited significantly increased weight gain compared with rats in the control group, whereas administration of turmeric extract significantly inhibited this increase (Table 3). Administration of turmeric extract at 300 mg/kg lessened body weight gain to a greater extent than that at 100 or 200 mg/kg. Moreover, administration of 200 mg/kg curcumin also significantly inhibited body weight gain. Daily food intake was monitored in all groups. The average daily food consumption in each group is shown in Table 3; importantly, no significant differences were observed between any of the groups regarding food intake. These findings suggest that both turmeric extract and curcumin attenuate CCl4-induced weight gain.Table 3 Effects of turmeric extract and curcumin on food intake and body weight Group Food intake (g/day) Body weight (g) Control 22.1 ± 3.1 335 ± 11.2 Curcumin 23.2 ± 3.3 335 ± 10.5 300 mg/kg turmeric extract 22.0 ± 2.3 351 ± 12.2 CCl4 22.3 ± 3.2 401 ± 20.4 CCl4 + curcumin 19.9 ± 3.2 376.0 ± 18.5 CCl4 + 100 mg/kg turmeric extract 24.6 ± 2.0 400.4 ± 19.5 CCl4 + 200 mg/kg turmeric extract 23.1 ± 1.3 381.1 ± 20.3* CCl4 + 300 mg/kg turmeric extract 21.2 ± 2.1 362.5 ± 20.2* *p < 0.05 vs. the CCl4 group Both turmeric extract and curcumin regulate CCl4-induced lipid accumulation The serum levels of triglyceride, total cholesterol, LDL-C, and HDL-C were also measured (Table 4). Compared with control mice, mice in the CCl4 group showed decreased TG, total cholesterol, and LDL-C levels. However, the levels of HDL-C were not different between the two groups. Compared with mice in the CCl4 group, mice treated with turmeric extract and mice treated with curcumin exhibited significantly increased TG, total cholesterol, and LDL-C levels.Table 4 Serum concentrations of triglyceride, total cholesterol, LDL-C, and HDL-C as determined by quantitative lipid assays Group Serum levels (mg/dl) TG Total cholesterol LDL-cholesterol HDL-cholesterol Control 82.4 ± 2.2 136.7 ± 12.8 56.6 ± 6.3 45.4 ± 6.4 Curcumin 89.4 ± 1.2 142.7 ± 6.3 60.2 ± 3.1 45.6 ± 1.6 300 mg/kg turmeric extract 83.7 ± 2.3 141.9 ± 1.5 63.2 ± 2.0 46.5 ± 3.3 CCl4 51.5 ± 7.1 94.3 ± 1.7 44.1 ± 4.0 44.5 ± 2.5 CCl4 + curcumin 77.0 ± 1.6 111.6 ± 8.3 52.2 ± 6.0 45.0 ± 9.4 CCl4 + 100 mg/kg turmeric extract 75.1 ± 1.0 95.2 ± 9.4 46.6 ± 7.1 43.8 ± 1.8 CCl4 + 200 mg/kg turmeric extract 78.1 ± 2.5* 101.3 ± 3.6* 49.6 ± 3.1* 45.4 ± 1.5 CCl4 + 300 mg/kg turmeric extract 83.0 ± 1.5* 112.3 ± 9.1* 51.5 ± 8.0* 44.8 ± 2.6 *p < 0.05 vs. the CCl4 group Both turmeric extract and curcumin protect against CCl4-induced liver damage To evaluate the protective effects of turmeric extract and curcumin against CCl4-induced lipid accumulation and liver damage, we performed histologic analyses of the livers of treated mice. CCl4 administration significantly increased the levels of the liver damage biomarkers aspartate transaminase (AST) and alanine transaminase (ALT) (Fig. 1). However, treatment with turmeric extract (300 mg/kg) or curcumin (200 mg/kg) significantly decreased the AST and ALT levels, indicating reduced liver damage compared with the CCl4 group. Morphologic analysis showed that CCl4 administration stimulated steatosis, as indicated by the appearance of lipid droplets. However, treatment with turmeric extract (300 mg/kg) or curcumin (200 mg/kg) resulted in decreased steatosis compared with the CCl4 group (Fig. 2).Fig. 1 Turmeric extract and curcumin reduce AST and ALT levels in CCl4-induced hepatic failure. Rats were intraperitoneally injected with CCl4 (0.1 mL/100 g body weight) every other day for 4 weeks. Turmeric extract (100, 200, and 300 mg/kg) and curcumin (200 mg/kg) were given once daily. Liver and blood samples were collected from all sacrificed animals. Serum levels of AST (a) and ALT (b) were determined. # p < 0.05 vs. the CCl4 group Fig. 2 Turmeric extract and curcumin protect the liver from CCl4-induced damage and lipid accumulation. Rats were intraperitoneally injected with CCl4 (0.1 mL/100 g body weight) every other day for 4 weeks. Turmeric extract (100, 200, and 300 mg/kg) and curcumin (200 mg/kg) were given once daily. Liver and blood samples were collected from all sacrificed animals. Liver tissue was fixed and stained with Oil Red O Turmeric extract and curcumin regulate CCl4-induced lipid peroxidation and antioxidant activity We next measured hepatic lipid peroxide levels (Fig. 3a). The levels of the hepatic lipid peroxides malondialdehyde (MDA) and 4-hydroxynonenal (4-HNE) were markedly increased in the CCl4 group compared with the control group. However, both turmeric extract and curcumin alleviated the increased levels of the lipid peroxides MDA and 4-HNE (Fig. 3b). We also quantified hepatic antioxidant activities (Fig. 4). Hepatic superoxide dismutase (SOD) and glutathione peroxidase (GPx) activities were markedly reduced in the CCl4 group compared with the control group; however, these effects were reversed by both turmeric extract and curcumin.Fig. 3 Turmeric extract and curcumin protect the liver from CCl4-induced lipid peroxidation. Rats were intraperitoneally (i.p.) injected with CCl4 (0.1 mL/100 g body weight) every other day for 4 weeks. Turmeric extract (100, 200, and 300 mg/kg) and curcumin (200 mg/kg) were given once daily. After liver samples were collected from all sacrificed animals, the levels of lipid peroxidation (a) and MDA + 4-HNE (b) were measured. # p < 0.05 vs. the CCl4 group Fig. 4 Turmeric extract and curcumin protect the liver from CCl4-induced oxidative stress. Rats were intraperitoneally injected with CCl4 (0.1 mL/100 g body weight) every other day for 4 weeks. Turmeric extract (100, 200, and 300 mg/kg) and curcumin (200 mg/kg) were given once daily. After liver samples were collected from all sacrificed animals, the levels of SOD (a) and GPx (b) were measured. # p < 0.05 vs. the CCl4 group Turmeric extract and curcumin regulate CCl4-induced oxidative stress in the liver Oxidative stress resulting from CCl4 treatment contributes to liver injury and stimulates hepatic fibrosis. The effects of turmeric and curcumin on CCl4-induced oxidative stress in the liver are shown in Fig. 5. In these experiments, red fluorescence from dihydroethidium indicates an increased ROS content in the liver. Chronic CCl4 exposure resulted in increased fluorescence, whereas much lower fluorescence was observed in the livers of rats treated with turmeric extract or 200 mg/kg curcumin. Moreover, these effects were dose-dependent. Glutathione (GSH) is the major low-molecular weight thiol and the most critical nonenzyme antioxidant in vitro [15]. GSH protects cells against oxidative stress-induced cellular damage by removing hydrogen peroxide (H2O2) and inhibiting lipid peroxidation [16]. The GSH/GSSG ratio is considered to be a sensitive indicator of oxidative stress [17, 18]. To evaluate the impact of turmeric extract and curcumin on oxidative stress in vivo, the levels of hepatic GSH and the ratios of reduced GSH to GSSG were determined. As shown in Fig. 6, CCl4 treatment markedly decreased the level of total GSH and the GSH/GSSG ratio. However, oral administration of turmeric and of curcumin completely prevented these CCl4-induced effects, resulting in levels of total hepatic GSH and hepatic GSH/GSS ratios similar to those of control rats. These results suggest that both turmeric extract and curcumin can protect the liver against CCl4-induced damage by attenuating oxidative stress.Fig. 5 Turmeric extract and curcumin protect the liver from CCl4-induced ROS production. Rats were intraperitoneally injected with CCl4 (0.1 mL/100 g body weight) every other day for 4 weeks. Turmeric extract (100, 200, and 300 mg/kg) and curcumin (200 mg/kg) were given once daily. a Liver tissue was isolated and loaded with 5 μM dihydroethidium. Fluorescence images were acquired. b Quantitative fluorescence density data. # p < 0.05 vs. the CCl4 group Fig. 6 Turmeric extract and curcumin protect the liver from CCl4-induced oxidative stress. Rats were intraperitoneally injected with CCl4 (0.1 mL/100 g body weight) every other day for 4 weeks. Turmeric extract (100, 200, and 300 mg/kg) and curcumin (200 mg/kg) were given once daily. a Livers were isolated and the levels of reduced glutathione (GSH) (a) and the GSH/GSSG ratios (b) were determined. # p < 0.05 vs. the CCl4 group Discussion In the present study, we showed that chronic CCl4-induced liver injury, defined as increased levels of serum markers of hepatic damage and abnormal liver morphology, was inhibited in the presence of turmeric extract and its active compound, curcumin. Moreover, the hepatic GSH/GSSG ratio was restored in the turmeric extract and its active compound, curcumin-treated group, indicating that this extract alleviates oxidative stress. Both turmeric extract and its active compound, curcumin, remarkably decreased CCl4-induced liver damage. Curcumin is a polyphenolic compound found in the dietary spice Curcuma longa Linn. Due to its wide range of biological and biochemical activities, the therapeutic effects of curcumin are being investigated in various disease models, including hepatic failure [19–21]. The main components of turmeric extract include curcumin and other curcuminoids (Table 2). Because curcumin is poorly absorbed following oral administration and is not hydrosoluble, the majority of the ingested curcumin is excreted intact after passing through the digestive tract. We thus selected a high dose of curcumin, 200 mg/kg, during experimental design. Other studies used a similar dose of curcumin [22–24]. In these studies, curcumin was reported to be metabolized at the intestinal mucosa and liver, with curcumin plasma levels detectable only when it was administered at a gram-based high dose; ≥ 1 g [25, 26]. In this study, orally administered curcumin was expected to be metabolized to a greater degree in the CCl4-exposed condition. Based on other references [22–24], we selected a high dose of curcumin, 200 mg/kg. Throughout this study, curcumin was used as a functional component control. The CCl4-induced experimental model of liver damage is very useful for the study of hepatotoxic effects [27] since CCl4 consistently produces liver damage in many species, including nonhuman primates [28]. In the liver, CCl4 metabolism stimulates lipid peroxidation and upregulates ROS production [29], which is important because peroxidated lipids and ROS cause hepatocyte necrosis, induce inflammation, and further stimulate the progression of hepatic fibrosis. In the progression of chronic hepatic failure, lipid accumulation is also frequently observed and is ultimately linked to the fibrosis pattern. However, elevated serum ALT and AST levels, clear markers of liver injury, were not observed in the turmeric extract-treated group (Fig. 1a, b). Similarly, curcumin also inhibited the increases in aminotransferase levels. Other studies have also demonstrated that curcumin administration can reverse the elevated serum AST and ALT levels in models of hepatotoxicity [30]. The current study showed that both turmeric extract and curcumin alleviate CCl4-induced upregulation of serum ALT and AST, indicating liver protection from CCl4-induced toxicity. In addition, turmeric extract and curcumin attenuated CCl4-mediated hepatic lipid accumulation (Fig. 2). The levels of serum TG, total cholesterol, and LDL were significantly decreased in the CCL4 group, whereas they were restored in the turmeric extract-treated group (Table 4). The hepatic lipid profiles (Table 5) were consistent with the serum profiles and the Oil Red O staining data (Fig. 2). Moreover, curcumin exerted a protective effect similar to that of the highest dose of the extract. This study shows that turmeric extract and curcumin protect the liver from CCl4-induced hepatic lipid accumulation. However, hepatic fibrosis was not observed in this model (data not shown). We conclude that more severe stress conditions and a longer time frame than this 4-weeks system are required to investigate aspects of fibrosis.Table 5 Hepatic concentrations of triglyceride, total cholesterol, LDL-C, and HDL-C as determined by quantitative lipid assays Group Liver levels (mg/dl) TG Total cholesterol LDL-cholesterol HDL-cholesterol Control 60.6 ± 0.1 70.6 ± 5.2 50.5 ± 4.3 44.3 ± 2.5 Curcumin 68.1 ± 0.1 68.9 ± 1.3 48.5 ± 8.5 47.1 ± 4.6 300 mg/kg turmeric extract 61.3 ± 0.2 71.3 ± 1.6 45.8 ± 5.5 45.6 ± 1.3 CCl4 92.7 ± 0.2 99.6 ± 1.5 84.6 ± 7.6 44.9 ± 7.8 CCl4 + curcumin 78.9 ± 0.2* 75.4 ± 1.5* 62.5 ± 5.6* 46.2 ± 1.5 CCl4 + 100 mg/kg turmeric extract 87.1 ± 0.1 92.7 ± 2.0 82.2 ± 5.5 45.6 ± 5.6 CCl4 + 200 mg/kg turmeric extract 77.3 ± 0.1* 88.3 ± 4.6* 75.5 ± 7.6* 45.0 ± 2.8 CCl4 + 300 mg/kg turmeric extract 68.3 ± 0.1* 78.1 ± 1.4* 61.5 ± 7.6* 46.5 ± 1.9 *p < 0.05 vs. the CCl4 group Oxidative stress is closely associated with hepatic lipid accumulation and liver failure [31]. Lipid peroxidation is markedly suppressed in the livers of rats treated with antioxidants such as flavonoids and vitamins [32]. We hypothesized that turmeric extract and curcumin protect the liver against CCl4-induced injury and hepatic lipid accumulation by decreasing oxidative stress. In support of our hypothesis, treatment with turmeric extract and treatment with curcumin both reduced oxidative stress, as demonstrated by the reductions in the CCl4-mediated increases in lipid peroxidation, malondialdehyde (MDA), and 4-hydroxynonenal (4-HNE) levels (Fig. 3a, b). Our results show that turmeric extract and curcumin not only increase the level of total hepatic GSH, but also markedly improve the GSH/GSSG ratio (Fig. 6). Although many factors have been implicated in CCl4-induced liver damage, oxidative stress and ROS production are thought to be of primary importance. Oxidative stress is defined as an imbalance between ROS production and removal [33]. ROS, which are generated as products of oxidative metabolism, frequently damage cellular macromolecules such as DNA and lipids. In the present study, the accumulation of ROS in the liver after CCl4 treatment was ameliorated by treatment with turmeric extract and treatment with curcumin (Fig. 5a, b). These results strongly suggest that turmeric extract and curcumin protect against hepatic functional disturbances, including hepatic dysmetabolism, through ROS regulation. Conclusions This study suggests that turmeric extract and curcumin are both highly effective in preventing chronic CCl4-induced liver damage and provides new insights into the potential pharmacologic targets of curcumin in the prevention of liver disease. Additional file Additional file 1: Figure S1. Chromatograms of blank (A), standard (B), turmeric extract (C). (PDF 164 kb) Abbreviations 4-HNE4-hydroxynonenal ASTAspartate transaminase CCl4Carbon tetrachloride CLLCurcuma longa L. turmeric GPxGlutathione peroxidase MDAMalondialdehyde ROSReactive oxygen species SODSuperoxide dismutase Acknowledgements This work was supported by the Food Functionality Evaluation Program of the Ministry of Agriculture, Food and Rural Affairs. Availability of data and materials The datasets supporting the conclusions of this article are included within the article and all datasets supporting our findings are available. Authors’ contributions HYL, GHL, HWJ, and SWK participated in the design of the research. HYL, MKC, GHL, HJK, and HJC carried out the experiments, analyzed the data and wrote the paper. All authors read and approved the final manuscript. Competing interest The authors declare that they have no competing interest. Ethics approval and consent to participate All animal procedures in this study were performed in accordance with the regulations described in the Care and Use of Laboratory Animals guide of Chonbuk National University. All procedures were also approved by the Institutional Animal Care and Use Committee of Chonbuk National University for the animal center (IACUC protocol CBU 150608-25). ==== Refs References 1. Ma JQ Ding J Zhang L Liu CM Hepatoprotective properties of sesamin against CCl4 induced oxidative stress-mediated apoptosis in mice via JNK pathway Food Chem Toxicol 2014 64 41 48 10.1016/j.fct.2013.11.017 24287204 2. Kaneko M Nagamine T Nakazato K Mori M The anti-apoptotic effect of fucoxanthin on carbon tetrachloride-induced hepatotoxicity J Toxicol Sci 2013 38 1 115 126 10.2131/jts.38.115 23358145 3. Savransky V Reinke C Jun J Bevans-Fonti S Nanayakkara A Li J Myers AC Torbenson MS Polotsky VY Chronic intermittent hypoxia and acetaminophen induce synergistic liver injury in mice Exp Physiol 2009 94 2 228 239 10.1113/expphysiol.2008.044883 19028810 4. Rechnagel RO Glende EA Jr Carbon tetrachloride hepatotoxicity: an example of lethal cleavage CRC Crit Rev Toxicol 1973 2 3 263 297 10.3109/10408447309082019 4357489 5. Albano E Alcohol, oxidative stress and free radical damage Proc Nutr Soc 2006 65 3 278 290 10.1079/PNS2006496 16923312 6. Koch OR Pani G Borrello S Colavitti R Cravero A Farre S Galeotti T Oxidative stress and antioxidant defenses in ethanol-induced cell injury Mol Asp Med 2004 25 1-2 191 198 10.1016/j.mam.2004.02.019 7. Xu Y Ku BS Yao HY Lin YH Ma X Zhang YH Li XJ The effects of curcumin on depressive-like behaviors in mice Eur J Pharmacol 2005 518 1 40 46 10.1016/j.ejphar.2005.06.002 15987635 8. Selvam R Subramanian L Gayathri R Angayarkanni N The anti-oxidant activity of turmeric (Curcuma longa) J Ethnopharmacol 1995 47 2 59 67 10.1016/0378-8741(95)01250-H 7500637 9. Aggarwal BB Sundaram C Malani N Ichikawa H Curcumin: the Indian solid gold Adv Exp Med Biol 2007 595 1 75 10.1007/978-0-387-46401-5_1 17569205 10. Corson TW Crews CM Molecular understanding and modern application of traditional medicines: triumphs and trials Cell 2007 130 5 769 774 10.1016/j.cell.2007.08.021 17803898 11. Tu CT Yao QY Xu BL Wang JY Zhou CH Zhang SC Protective effects of curcumin against hepatic fibrosis induced by carbon tetrachloride: modulation of high-mobility group box 1, Toll-like receptor 4 and 2 expression Food Chem Toxicol 2012 50 9 3343 3351 10.1016/j.fct.2012.05.050 22683883 12. Yao QY Xu BL Wang JY Liu HC Zhang SC Tu CT Inhibition by curcumin of multiple sites of the transforming growth factor-beta1 signalling pathway ameliorates the progression of liver fibrosis induced by carbon tetrachloride in rats BMC Complement Altern Med 2012 12 156 10.1186/1472-6882-12-156 22978413 13. Xu J Fu Y Chen A Activation of peroxisome proliferator-activated receptor-gamma contributes to the inhibitory effects of curcumin on rat hepatic stellate cell growth Am J Physiol Gastrointest Liver Physiol 2003 285 1 G20 30 10.1152/ajpgi.00474.2002 12660143 14. Bligh EG Dyer WJ A rapid method of total lipid extraction and purification Can J Biochem Physiol 1959 37 8 911 917 10.1139/o59-099 13671378 15. Wu G Fang YZ Yang S Lupton JR Turner ND Glutathione metabolism and its implications for health J Nutr 2004 134 3 489 492 14988435 16. Blair IA Endogenous glutathione adducts Curr Drug Metab 2006 7 8 853 872 10.2174/138920006779010601 17168687 17. Fridovich I The biology of oxygen radicals Science 1978 201 4359 875 880 10.1126/science.210504 210504 18. Johnson SJ Hines JE Burt AD Immunolocalization of proliferating perisinusoidal cells in rat liver Histochem J 1992 24 2 67 72 10.1007/BF01082441 1577623 19. Gaedeke J Noble NA Border WA Curcumin blocks multiple sites of the TGF-beta signaling cascade in renal cells Kidney Int 2004 66 1 112 120 10.1111/j.1523-1755.2004.00713.x 15200418 20. Sharma RA Ireson CR Verschoyle RD Hill KA Williams ML Leuratti C Manson MM Marnett LJ Steward WP Gescher A Effects of dietary curcumin on glutathione S-transferase and malondialdehyde-DNA adducts in rat liver and colon mucosa: relationship with drug levels Clin Cancer Res 2001 7 5 1452 1458 11350917 21. Punithavathi D Venkatesan N Babu M Curcumin inhibition of bleomycin-induced pulmonary fibrosis in rats Br J Pharmacol 2000 131 2 169 172 10.1038/sj.bjp.0703578 10991907 22. Abdelmegeed MA Banerjee A Yoo SH Jang S Gonzalez FJ Song BJ Critical role of cytochrome P450 2E1 (CYP2E1) in the development of high fat-induced non-alcoholic steatohepatitis J Hepatol 2012 57 4 860 866 10.1016/j.jhep.2012.05.019 22668639 23. Zhao Y Ma X Wang J He X Hu Y Zhang P Wang R Li R Gong M Luo S Curcumin protects against CCl4-induced liver fibrosis in rats by inhibiting HIF-1alpha through an ERK-dependent pathway Molecules 2014 19 11 18767 18780 10.3390/molecules191118767 25407718 24. Fu Y Zheng S Lin J Ryerse J Chen A Curcumin protects the rat liver from CCl4-caused injury and fibrogenesis by attenuating oxidative stress and suppressing inflammation Mol Pharmacol 2008 73 2 399 409 10.1124/mol.107.039818 18006644 25. Sharma RA Euden SA Platton SL Cooke DN Shafayat A Hewitt HR Marczylo TH Morgan B Hemingway D Plummer SM Phase I clinical trial of oral curcumin: biomarkers of systemic activity and compliance Clin Cancer Res 2004 10 20 6847 6854 10.1158/1078-0432.CCR-04-0744 15501961 26. Epelbaum R Schaffer M Vizel B Badmaev V Bar-Sela G Curcumin and gemcitabine in patients with advanced pancreatic cancer Nutr Cancer 2010 62 8 1137 1141 10.1080/01635581.2010.513802 21058202 27. Saba AB Oyagbemi AA Azeez OI Amelioration of carbon tetrachloride-induced hepatotoxicity and haemotoxicity by aqueous leaf extract of Cnidoscolus aconitifolius in rats Niger J Physiol Sci 2010 25 2 139 147 22314953 28. Yoshida T Adachi E Nigi H Fujii S Yanagi M Changes of sinusoidal basement membrane collagens in early hepatic fibrosis induced with CCl4 in cynomolgus monkeys Pathology 1999 31 1 29 35 10.1080/003130299105494 10212919 29. Basu S Carbon tetrachloride-induced lipid peroxidation: eicosanoid formation and their regulation by antioxidant nutrients Toxicology 2003 189 1-2 113 127 10.1016/S0300-483X(03)00157-4 12821287 30. Naik SR Thakare VN Patil SR Protective effect of curcumin on experimentally induced inflammation, hepatotoxicity and cardiotoxicity in rats: evidence of its antioxidant property Exp Toxicol Pathol 2011 63 5 419 431 10.1016/j.etp.2010.03.001 20363603 31. Poli G Parola M Oxidative damage and fibrogenesis Free Radic Biol Med 1997 22 1-2 287 305 10.1016/S0891-5849(96)00327-9 8958154 32. Parola M Robino G Oxidative stress-related molecules and liver fibrosis J Hepatol 2001 35 2 297 306 10.1016/S0168-8278(01)00142-8 11580156 33. Behrman HR Kodaman PH Preston SL Gao S Oxidative stress and the ovary J Soc Gynecol Investig 2001 8 1 Suppl Proceedings S40 42 10.1016/S1071-5576(00)00106-4 11223371

PMC005xxxxxx/PMC5000415.txt

==== Front BMC Public HealthBMC Public HealthBMC Public Health1471-2458BioMed Central London 343810.1186/s12889-016-3438-6Research ArticleHow are self-rated health and diagnosed disease related to early or deferred retirement? A cross-sectional study of employees aged 55-64 http://orcid.org/0000-0002-3193-205XNilsson Kerstin kerstin.nilsson@med.lu.se 12Hydbom Anna Rignell anna.rignell-hydbom@med.lu.se 1Rylander Lars lars.rylander@med.lu.se 11 Divison of Occupational and Environmental Medicine, Lund University, Box 188, SE-221 85 Lund, Sweden 2 Swedish University of Agricultural Sciences, Work Sciences, Business Economics and Environmental Psychology, Box 88, SE-230 53 Alnarp, Sweden 26 8 2016 26 8 2016 2016 16 1 88610 6 2016 3 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background More people will probably continue working into old age in the future due to the increased size of aging populations in many countries. We therefore need to know more about older workers’ health in relation to their work situation and retirement. This study is a part of a theoretical development of older workers’ situations. Older workers’ situations are theoretically themed in nine areas by the authors of this study. The aims of the study were to investigate the relationship between: i) diagnosed disease and factors in older workers’ situations, theoretically themed in nine areas; ii) self-rated health and factors in older workers’ situations, theoretically themed in nine areas; iii) diagnosed disease and self-rated health; and iv) the relationships between these health measures and retirement. Methods A questionnaire-based cross-sectional study, using logistic regression, with 1,756 health care personnel aged 55–64 years. The questionnaire used gave an overview of most different areas in the older workers’ situations. Result There was a difference in the participants’ frequency of objectively specified diagnosed disease and their subjectively experienced self-rated health. A bad self-rated health was related higher to early retirement than diagnosed diseases. In the multivariate model, having ‘Diagnosed disease’ was not significantly related to whether older workers thought they could not work beyond 65 years of age. A bad ‘Self-rated health’ was also more highly related to whether older workers thought they could not work beyond 65 years, than if the respondents stated that a ‘Diagnosed disease is a hindrance in my daily work’ in the multivariate model. Conclusion This study showed an important difference between older workers’ own experiences and the effect of their self-rated health and their diagnosed diseases. Subjective self-rated health seems to be more important to people’s retirement planning than diagnosed disease. The most important factors affecting older workers’ self-rated health was the degree to which they felt physically and mentally fatigued, their possibilities for revitalization, and issues of work satisfaction, age discrimination and attitudes of managers to them as seniors. Keyword Diagnosed diseaseExtended work lifeHealthOlder workersRetirementSelf-rated healthissue-copyright-statement© The Author(s) 2016 ==== Body Background The social occupational pension system was introduced to make it economically possible for people to withdraw from working life when their health declined as a result of old age. To extend working life for older workers is nowadays high on the agenda in many countries considering their social pension system, because the old age dependency ratio will increase along with the rising proportion of older people in most industrialised countries [1–5]. Due to that, many countries will have to increase the number of working hours in the economy in order to maintain the welfare state and, will therefore have to motivate people to work into older age [1–4]. Health is described as one of the most important factors affecting disability and retirement age [6–18]. Defining a person’s health is complex and due to its different definitions, health can be rated in different ways. Some of the most commonly used health measures in investigations are self-rated health and diagnosed diseases [27–29]. Diagnosed diseases as determined by a physician is an objective way to rate health, but self-rated health is an individual’s subjective experience of their own health [30]. The social insurance system in many countries uses diagnosis to define a person’s eligibility for sick leave and retirement, although self-rated health has been found to be a better predictor for disability retirement than medical diagnosis in some studies [31–33]. People in their 50s reported a lower level of good self-rated health than their diagnosed diseases, but 10 years later the same persons reported a higher level of good self-rated health as well as a greater amount of diagnosed disease [38]. Some studies have also stated that self-rated health is a better predictor of mortality than more objective health parameters like diagnosed disease phase [34–37]. Retirement is described to be both positive and negative to an individual’s health [19]. Some of those who participate in working life at an older age (65+) perceive themselves to have better mental and physical health due to their meaningful tasks and physical activities in work [11, 20–22]. Others report that mental and physical fatigue decrease after retirement [11, 23, 24]. Some studies have stated that health improves after retirement among those with a poor work environment and who had health complaints before retirement, but this is less apparent among workers experiencing low demands, high satisfaction and a high occupational grade in working life [25, 26]. However, a person’s health status is important to whether they can participate and continue working, and it is their first concern regarding retirement decisions [11]. Regarding the complexity of health definitions, deeper knowledge is required regarding the relationships and differences between diagnosed disease and self-rated health [28, 29, 36, 38]. It is currently unknown whether diagnosed health or self-rated health is the best predictor of retirement. Different factors of importance to work life participation have been divided into theoretical themes [39, 40] including where those theoretical themes comprise nine areas [41–44]. The overall aim of this study was to explore if diagnosed health or self-rated health was a better predictor of retirement age. The objectives were to investigate the relation between: i) diagnosed disease and factors in older workers’ situations within nine areas of theoretical themes; ii) self-rated health and factors in older workers’ situations within nine areas of theoretical themes; iii) diagnosed disease and self-rated health; and iv) the relationships between these health measures and retirement. Methods Study population The study population consisted of 2,822 employees aged 55–64 years and identified in the official employment register for healthcare and medical services in Scania, the most southern county in Sweden. The study design was cross-sectional, and these individuals were sent a questionnaire by post to measure their attitudes and self-stated potential to work an extended working life (see detailed description below). After two reminders, 1,815 individuals responded, which corresponds to a response rate of 64 %. However, 59 individuals who did not answer the questions about their self-rated health and diagnosed disease were excluded from the study, which left 1,756 individuals in the final study population. The median age among the participants was 58 years, which was similar to that of non-participants (58 years). The majority of the participants who received a questionnaire were women (85 %); this reflects that more women than men work in healthcare and medical services. In addition, the response rate was better among women (68 %) than men (59 %). Among the participants with a partner, 61 % had a partner who worked full-time, 13 % had a partner who worked part-time, and 26 % had a partner who did not work at all. The distribution of occupations among the participants was: 21 % nurses, 20 % medical secretaries, 17 % physicians, 16 % nursing assistants, 7 % physiotherapists, 7 % welfare officers, 4 % psychologists, 5 % psychiatric nursing assistants, and 3 % occupational therapists. The proportion of non-responders was highest among physicians (25 %), and lowest among physiotherapists, occupational therapists, psychologists and welfare officers (14 %). Questionnaire Earlier research and theories regarding factors important to older workers and retirement present a complex picture with many facets [39, 40]. To cover most of the known areas of interest, a questionnaire was developed from a literature review and from results of previous studies by the research team regarding different areas important to older workers, retirement and extended working life. The questions in the questionnaire were subdivided into nine theoretical themes. These themes and the number of items in each were: (i) Physical and mental health (9 variables); (ii) Personal finances incentives (7 variables); (iii) Physical work environment (3 variables); (iv) Mental work environment (9 variables); (v) Work pace and working time (9 variables); (vi) Attitude to seniors in the organisation (7 variables); (vii) Possibility for skill development (9 variables); (viii) Motivation and work satisfaction (7 variables); and (ix) Family/leisure pursuits and attitude to retirement in society (5 variables). The majority of the chosen questions in the questionnaire had previously been validated in other studies. A pilot study was performed to ensure that the final questionnaire fulfilled the objectives. In the analysis, the basis for the ‘self-rated health’ variable was the following variable in the questionnaire: ‘I am currently experiencing good health/well-being’. The response options were dichotomised from four to two variables, i.e. from both highly agrees and agrees to just ‘Highly agree’, and from both partly disagree and disagree to just ‘Disagree’. The variable ‘diagnosed disease’ was based on the variable ‘I have a diagnosed disease’, with the response opinions ‘Yes’ or ‘No’. To take the analysis one step forwards, the variable ‘I have a diagnosed disease which is a hindrance in my daily work’ was also used in the analyses. Those response options were also dichotomised from four to two variables, i.e. from highly agrees and agrees to ‘Highly agree’, and from partly disagree and disagree to ‘Disagree’. Those who did not answer this question were added to those who disagreed, because the question was a follow-up question to the earlier described question ‘I have a diagnosed disease’ (Yes/No). A categorisation of diagnoses was not made in this study. The variable ‘Planned retirement’ was based on a question as to whether the participant believed he/she could work until 55–59, 60–64, 65, or ≥66 years of age. The response options were dichotomised at 65 years of age (i.e., working until <65 versus ≥65 years of age). The reason for doing that, was because 65 years is the most common retirement age in Sweden, and to leave working life before is seen as an early retirement. On the other questions, the participants had five response options ranging from ‘Fully agree’ to ‘Fully disagree’. In the present study, the answers to each variable were sorted into three categories: ‘Disagree’, ‘Partly agree’ and ‘Agree’. Statistical analysis Regarding the four aims of the study, we used the following strategy for the statistical analyses: Relationship between the health ratings The relationship between the two health ratings (self-rated health and diagnosed disease) was presented in frequencies. The difference between self-rated and diagnosed disease was analysed in cross-tabulation and tested with McNemar’s test. In addition, separate analyses were performed for subgroups based on gender, age, occupation and planned retirement. Relationship between the nine themes and the health ratings: diagnosed disease and self-rated health This was analysed and tested with logistic regression models (generating odds ratios [OR], 95 % confidence intervals [CI], and P-values) to identify statements within the nine themes that were associated with the health outcomes. For each of the outcome variables, we used the following analytical strategy: Analyses within each of the nine themes First, univariate analyses were carried out; in other words, the relationship was evaluated for one variable at a time, and variables with P < 0.05 in the regression analyses were retained. In the second step, the variable with the lowest P-value (if P < 0.05) was retained and all other statements were tentatively included, one variable at a time. In the third step, the two statements with the lowest P-values (if both P < 0.05) were retained and the remaining statements were tentatively included one variable at a time. This procedure continued as long as P < 0.05 for all variables included. Analyses including all nine themes The analysis started by including the selected variables from theme (i) and theme (ii) in a multivariate model (i.e., the themes ‘Physical and mental health’ and ‘Personal finances incentives’). Variables with P < 0.05 were retained in the model for the next step, which also included the selected variables from theme (iii). This procedure continued until all nine themes were included in a final model. To check once more whether the model was robust, the variables excluded from the nine themes were tested one at a time with the final model. Relationships between the health ratings, the nine themes and planned retirement Logistic regression analysis was used for evaluating the relationships between the health ratings and to check whether the participant believed he/she could work until <65 versus ≥65 years of age. Both univariate (i.e., one health measure at a time) and multivariate analyses were performed. Results Relationships with health ratings The relationship between the two health ratings (self-rated health and diagnosed disease) was tested by a cross tabulation. A total 25 % stated that they have a diagnosed disease and low self-rated health; 8 % stated no diagnoses, but low self-rated health; 20 % stated that they have a diagnosed disease, but a high self-rated health; 48 % stated no diagnoses and a high self-rated health (P < 0.001). In the study population, 53 % thought they could work until 55–64 years of age, and 47 % thought they could work until 65 years of age or beyond (Table 1). Among those with high self-rated health, 62 % believed they could work until 65 years of age. Of those who had a diagnosed disease, 46 % thought they could work until 65 years of age or beyond. Furthermore, 55 % stated that their diagnosed disease was caused by their work, and 62 % stated that their diagnosed disease was an obstacle in their daily work.Table 1 Proportion of respondents with high or low self-rated health and diagnosed disease; their age groups, self-estimated retirement age, gender and occupation High self-rated health Low self-rated health No diagnosed disease Diagnosed disease Median age in the group 58 years 59 years 58 years 59 years 55–59 years of age (n = 1060) 69 % 31 % 58 % 42 % 60–64 years of age (n = 696) 65 % 35 % 52 % 48 % Genderα Women (n = 1486) 67 % 33 % 56 % 44 % Men (n = 263) 71 % 29 % 58 % 42 % Occupationα Physician (n = 294) 67 % 33 % 55 % 45 % Nurse (n = 362) 69 % 31 % 59 % 41 % Nursing assistant (n = 273) 70 % 30 % 56 % 44 % Physiotherapist (n = 118) 68 % 32 % 54 % 46 % Occupational therapist (n = 49) 65 % 38 % 74 % 26 % Welfare officer (n = 116) 62 % 38 % 62 % 38 % Psychologist (n = 75) 67 % 33 % 63 % 37 % Psychiatric nursing assistant (n = 82) 54 % 46 % 45 % 55 % Medical secretary (n = 342) 69 % 31 % 55 % 45 % Non-defined occupation (n = 42) 69 % 31 % 38 % 62 % Men, and those in the age group 55–59 years, had a higher frequency of self-rated health, whereas women and those aged 60–64 years had more diagnosed disease. The occupation with most diagnosed disease was psychiatric nursing assistant, while nursing assistant had the highest self-rated health. Relationship between the nine themes and the health ratings Statements associated with low self-rated health In the final multivariate model, eight statements from five of the nine themes were associated with low self-rated health (Table 2): two statements from each of the themes ‘physical and mental health’, ‘work pace and working time’ and ‘attitude to seniors in the organisation’, and one statement from each of the themes ‘physical work environment’ and ‘motivation and work satisfaction’. The highest ORs were observed for agreement with the statement: ‘I feel physically worn out’ (OR 4.1, 95 % CI 2.8–6.1) and partial agreement with ‘I am satisfied with my present work situation’ (OR 2.9, 95 % CI 1.8–4.4). The OR changed only marginally when the background variables of age, gender, marital status and occupation, were included in the final models to check the risk of possibilities of confounders (data not shown).Table 2 Distribution regarding ‘self-rated health’ outcomes for statements in the final multivariate model. Increased odds ratio (OR) indicates that the individual experienced low self-rated health (95 % confidence intervals (CI); Ref = reference) Area Statement Agree/disagree with the statement Univariate estimate Multivariate model OR 95 % Cl OR 95 % Cl Physical and mental health I feel mentally worn out. Disagree Ref. Ref. Partly agree 5.0 3.8–6.6 1.6 1.1–2.3 Agree 9.0 6.9–12 1.8 1.3–2.6 I feel physically worn out. Disagree Ref. Ref. Partly agree 5.8 4.4–7.6 2.5 1.8–3.6 Agree 13 9.2–17 4.1 2.8–6.1 Physical work environment I can usually manage the physical working demands of my daily work. Agree Ref. Ref. Partly agree 3.2 2.1–4.9 1.9 1.1–3.2 Disagree 3.1 2.2–4.4 1.7 1.1–2.7 Working pace and working time I seldom feel rested. Disagree Ref. Ref. Partly agree 3.6 2.7–4.9 1.6 1.1–2.2 Agree 7.9 6.2–10 1.9 1.3–2.7 I feel that I get enough rest/relaxation between my working days. Agree Ref. Ref. Partly agree 9.5 7.3–12 2.3 1.6–3.3 Disagree 4.8 3.6–6.4 2.1 1.5–3.0 Attitude to senior in the organization The manager’s attitude towards senior employees is positive at my workplace. Agree Ref. Ref. Partly agree 2.5 1.8–3.5 0.7 0.4–1.0 Disagree 2.5 1.9–3.3 1.4 1.0–2.0 I feel discriminated against in my workplace because of my age. Disagree Ref. Ref. Partly agree 3.8 2.5–5.6 2.3 1.4–3.9 Agree 2.8 1.9–4.1 1.6 1.0–2.6 Motivation and work satisfaction I am satisfied with my present work situation. Agree Ref. Ref. Partly agree 7.8 5.6–11 2.9 1.8–4.4 Disagree 4.6 3.6–5.9 2.1 1.6–2.9 Statements associated with diagnosed disease In the final multivariate model, four statements from three of the nine themes were associated with diagnosed disease: one variable from the theme ‘physical and mental health’, one from the theme ‘physical work environment’, and two from the theme ‘working pace and working time’ (Table 3). The statements with the highest OR values were: ‘I can usually manage physical working demands of my daily work’ (OR 2.4, 95 % CI 1.7–3.5), and ‘I feel physically worn out’ (OR 2.3, 95 % CI 1.7–3.2). The OR changed only marginally when the background variables of age, gender, marital status and occupation were included in the final model to check the risk of possible confounders (data not shown).Table 3 Distribution regarding ‘diagnosed disease’ outcomes for statements in the final multivariate model. Increased odds ratio (OR) indicates an individual with a diagnosed disease (95 % confidence interval (Cl); Ref = reference) Themes Statement Agree/ disagree with the statement Univariate estimate Multivariate model OR 95 % Cl OR 95 % Cl Physical and mental health I feel physically worn out. Disagree Ref. Ref. Partly agree 2.3 1.7–2.9 1.7 1.3–2.3 Agree 3.6 2.7–4.7 2.3 1.7–3.2 Physical work environment I can usually manage physical working demands of my daily work. Agree Ref. Ref. Partly agree 2.1 1.9–4.6 2.3 1.5–3.7 Disagree 3.1 2.2–4.4 2.4 1.7–3.5 Working pace and working time The pace of work in my daily work is too rapid. Disagree Ref. Ref. Partly agree 0.8 0.7–1.1 0.6 0.5–0.8 Agree 1.3 1.0–1.6 0.8 0.7–1.1 I feel that I get enough rest/relaxation between my working days. Agree Ref. Ref. Partly agree 2.4 1.9–3.0 1.7 1.3–2.2 Disagree 2.1 1.7–2.7 1.8 1.4–2.4 Statements associated with diagnosed disease, which is a hindrance in daily work For the variable ‘I have a diagnosed disease’, there also was a follow-up question: ‘I have a diagnosed disease which is a hindrance in my daily work’. It was therefore interesting to also investigate the relationship of that variable to the nine themes. In the final multivariate model, five statements from three of the nine themes were associated with ‘my diagnosed disease is a hindrance in my daily work’: two statements from the theme ‘physical and mental health’, one from the theme ‘physical work environment’, and two from the theme ‘personal finance incentives’ (Table 4). The statements with the highest OR values in the multivariate model were: ‘I can usually manage physical working demands of my daily work’ (OR 2.7, 95 % CI 1.5–4.7), and ‘I feel mentally worn out’ (OR 2.1, 95 % CI 1.3–3.3). The OR changed only marginally when the background variables age, gender, marital status and occupation were included in the final model to check for possibilities of confounders (data not shown).Table 4 Distribution regarding ‘I have a diagnosed disease which is a hindrance in my daily work’ outcomes for statements in the final multivariate model. Increased odds ratio (OR) indicates an individual with a diagnosed disease (95 % confidence interval (Cl); Ref = reference) Themes Statement Agree/ disagree with the statement Univariate estimate Multivariate model OR 95 % Cl OR 95 % Cl Physical and mental health I feel physically worn out. Disagree Ref. Ref. Partly agree 3.2 2.3–4.5 2.0 1.3–3.0 Agree 2.3 1.5–3.4 1.5 0.9–2.4 I feel mentally worn out. Disagree Ref. Ref. Partly agree 2.5 1.8–3.3 1.7 1.2–2.4 Agree 2.3 1.5–3.6 2.1 1.3–3.3 Personal economic incentives My intention is to work beyond age 65 to get a better pension. Agree Ref. Ref. Partly agree 1.7 1.1–2.5 1.7 1.1–2.6 Disagree 1.6 1.1–2.3 1.5 1.0–2.2 I want to work fewer hours even if I do not maintain the same personal economic standards. Disagree Ref. Ref. Partly agree 1.5 1.1–2.2 1.4 1.0–2.0 Agree 2.3 1.7–3.2 1.9 1.3–2.6 Physical work environment I can usually manage physical working demands of my daily work. Agree Ref. Ref. Partly agree 1.2 0.7–2.3 1.1 0.6–2.2 Disagree 3.6 2.1–6.0 2.7 1.5–4.7 Relationships between the health ratings, statements and planned retirement A total 52 % (n = 919) of the participants believed they could work until 65 years of age or beyond. In univariate models, self-rated health (high vs. low; OR 3.3, 95 % CI 2.7–4.1) and diagnosed disease (yes vs. no; OR 1.7, 95 % CI 1.4–2.1) were associated with whether they thought they could work until 65 years of age or not. When the health ratings were investigated simultaneously, self-rated health (high vs. low; OR 3.2, 95 % CI 2.5–4.0) with no diagnosed disease (no vs. yes; OR 1.1, 95 % CI 0.9–1.3) was significantly associated with whether the participants thought they could work until 65 years of age or not. To take the analysis of the difference one stage further, the variable ‘I have a diagnosed disease which is a hindrance in my daily work’ was inserted. When the health ratings were investigated simultaneously, self-rated health (high vs. low; OR 2.7, 95 % CI 2.1–3.6) and having a diagnosed disease as a hindrance in daily work (high vs. low; OR 1.8, 95 % CI 1.3–2.4) were significantly associated with whether the participants thought they could work until 65 years of age or not, though not having a diagnosed disease was not (no vs. yes; OR 1.3, 95 % CI 0.9–1.8). However, self-rated health was still more highly associated with whether the participants thought they could work until 65 years of age or not. In the next step of the analysis, the background variables of age, gender, marital status and occupation were included in the final model to check the risk of confounders. Despite this, the OR changed only marginally (data not shown). Discussion This cross-sectional study examined employees’ subjective experienced self-rated health compared with objectively diagnosed disease, and the relationship of these with planned retirement. As a part of the exploration and deriving a theory development on older workers’ situations, a questionnaire grounded in previous literature and regarding nine theoretical themes was used in the study. High and low self-rated health were stated to be more strongly associated with whether older workers believed they could continue working beyond 65 or not. That was consistent even though diagnosed disease was a hindrance in the older employees’ daily work. Of the nine themes chosen to cover older workers’ lives and work situations, ‘physical and mental health’, ‘physical work environment’ and ‘working pace and working time’ were associated with diagnosed disease. However, these three themes as well as the themes ‘attitude to seniors in the organisation’ and ‘motivation and work satisfaction’ were associated with self-rated health. The results showed that 27–29 % of the respondents gave inconsistent answers regarding self-rated health and diagnosed disease. That showed an important difference between older workers’ own experience, and the effect of their self-rated health and their diagnosed diseases. Today, diagnosed disease is used to determine the need for economic assistance during sick leave and for a disability pension in the social insurance system in many countries [31–33]. In this study, ‘diagnosed disease as a hindrance in daily work’ was the only health parameter associated with personal economic situation, but was not as highly associated as physical work environment or mental and physical capability. In the present study, those aged 55–59 years reported higher self-rated health and less diagnosed disease than those aged 60–64 years of age. The participants with high self-rated health and no diagnosed disease were most likely to believe they could work until 65 years of age or beyond. The results indicated that self-rated health was a better predictor of extended working life than diagnosed disease. This result is similar to those of earlier studies that stated self-rated health to be a better health predictor than diagnosed disease or disability, retirement and mortality [31–37]. Therefore, these earlier studies and the results from our study seem to indicate that it is not reliable to only determine people’s health with an objective health rating of diagnosed disease for retirement or their potential to extend their working life. However, this study does not have the intention to change the formal system that grants retirement to only include subjective health ratings. The interest is in how we could make working life more sustainable for people. When self-rated health is stated to be an important factor for an extended working life, we think it is important to make interventions that could increase an individual’s self-rated health, as well as reduce their diagnosed disease. Self-rated health seems to include more information than diagnosed disease on how people experience their total work situation in relation to their health and wellbeing. Both subjective self-rated health and objectively diagnosed disease were mostly associated with older workers’ physical capability, physical work environment, working pace and opportunities to get sufficient rest and relaxation between their working days. In addition, self-rated health was associated with satisfaction in their work situations, the attitude of managers to older workers in the organisation, and whether individuals experienced age discrimination. Previous studies have shown that older people who experience good quality of life and have meaningful tasks experience better health [11, 20–22], while health among individuals with low satisfaction and low occupational grade improves after they leave working life [25]. In the present study, these factors appeared to be particularly associated with self-rated health. Secondary findings that need to be further investigated are that diagnosed disease and low self-rated health proved to be more common among women than men in this study. Earlier studies have mostly associated women’s preference for early retirement with the retirement of their older husbands [45]. However, men who live with an older woman also plan to retire early, along with their older wife [11]. Therefore, in future studies, the importance of women’s health experience in relation to planned retirement needs to be analysed in depth. Previous studies have also identified educational level, occupational status and working conditions as predictors of health, ageing and ability to participate in working life [10, 26, 46, 47]. Occupational differences therefore also need to be further investigated in a future study. A new study is already planned to analyse specific diseases that might be predictive of early or later retirement. Limitations The questionnaire was formulated from a theoretical model of ageing and themes with importance for work life participation [41–44]. The majority of the statements in the questionnaire have previously been validated in other surveys, but some questions were new and not validated. Another limitation was that although the number of participants in the study was relatively large, a potential weakness was that 36 % of the individuals contacted did not participate. Unfortunately, we were unable to evaluate this dropout rate in depth due to lack of information about the non-respondents. A further limitation of the study was that everyone included in the study population was employed, and we do not know whether people with poor health had already left the workforce. Due to this ‘healthy worker effect’, selection bias must be considered for the study. However, internal comparisons are reported to be one of the most effective ways of reducing the healthy worker effect, and the individuals in this study worked in the same workforce [48]. The respondents in the study were mostly women, and this reflects the gender distribution among people employed in the healthcare and medical sector. We do not regard this as a major shortcoming of the study, since the models were robust even after adjusting for gender, age and marital status. Furthermore, older people were described as rating their self-rated health more highly than younger people did [27, 29]. However, the study population were in the same age group (55–64 years of age), so we do not regard this as a major shortcoming. Finally, the study was a cross-sectional study and only in one working sector. Therefore, a limitation is that we do not have any information on how the participants’ health rates have developed over time, and whether the result is specific for individuals working in health care. Therefore, we will test the result in a new study, following the health development in a cohort with individuals from different occupations aged 55–75 years. Conclusion Self-experienced health is a subjective health rating, and diagnosed disease is an objective one. Self-rated health was more highly associated with older workers’ retirement planning in this study than diagnosed disease was. The most important factors for older workers’ experience of a good self-rated health were their physical and mental fatigue, potential for revitalization, having meaningful activities and occupations that gave work satisfaction, as well as experiencing age discrimination and attitude of managers to them as seniors. Related to findings in this study about older workers’ health, considerations of initiative to extend working life might: i) adjust the work pace to older workers’ capacity and/or provide extra time for rest and relaxation between working sessions; ii) combat age discrimination and engender more positive attitudes towards older workers among managers and within organisations; and iii) increase motivation and satisfaction through interesting tasks and occupations. We will, in a forthcoming intervention study, investigate if such initiatives could be a sustainable way to improve older workers’ health and possibly increase the number of work hours they can and want to contribute to the national economy. Availability of data and materials The data is not publicly available but could be requested from the corresponding author after an ethical approval to take part of the dataset. Authors’ contributions KN carried out the design of the study, performed the statistical analysis, coordination and drafted the manuscript. LR participated in the design of the study, the statistical analysis, and helped to draft the manuscript. AR-H conceived the study and helped to draft the manuscript. All of the authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent to publish Not applicable. Ethics approval and consent to participate The Regional Ethical Review Board in Lund approved the study. Those respondents that did not send back their signed written informed consent for participation with the survey were excluded from the study population, in agreement with the recommendation in the ethics approval for this study. ==== Refs References 1. Danish Labour Market Commission Velfæd kræver arbejde Welfare request work 2009 Copenhagen Labour Market Commission 2. European Commission. Demography Report. Directorate-General for Employment, Social Affairs and Inclusion. Luxembourg: Eurostat, the Statistical Office of the European Union; 2010. 3. OECD Social Indicator Old age support rate 2011 4. SOU Åtgärder för ett längre arbetsliv. Measures for an extended working life. Report 2013:25, Luxembourg: Government Offices of Sweden; 2013. (In Swedish). 5. UNFPA & HelpAge International Ageing in the Twenty-First Century: A Celebration and A Challenge 2012 New York United Nations Population Fund (UNFPA) & London, HelpAge International 6. Ahola K Sirén I Kivimäki M Ripatti S Aromaa A Lönnqvist J Work-Related Exhaustion and Telomere Length: A Population-Based Study PLoS One 2012 7 7 e40186 10.1371/journal.pone.0040186 22808115 7. Börsch-Supan A Brugiavini A Croda E The role of institutions and health in European patterns of work and retirement J of Eur Social Policy 2009 19 341 358 10.1177/1350506809341515 8. Forma P Tuominen E Väänänen-Tomppo I Who wants to continue at work? Finnish pension reform and the future plans of older workers Eur J of Social Security 2005 7 227 250 9. Koolhaas W van der Klink JJ Groothoff JW Brouwner S Towards a sustainable healthy working life: relationship between chronological age, functional age and work outcomes Eur J of Public Health 2012 22 3 424 429 10.1093/eurpub/ckr035 21459842 10. Rignell NK Hydbom A Rylander L Factors influencing the decision to extend working life or to retire Scand J Work Environ Health 2011 37 6 473 480 10.5271/sjweh.3181 21725583 11. Nilsson K Why work beyond 65? Discourse on the decision to continue working or retire early Nord J of Working Life Studies 2012 2 3 7 28 10.19154/njwls.v2i3.2361 12. Nordenmark M Stattin M Psychosocial wellbeing and reasons for retirement in Sweden Ageing & Society 2009 29 413 430 10.1017/S0144686X08008179 13. Cho PS Jang S-N S-I Health Conditions Sensitivity to Retirement and Job Loss Among Korean Middle-aged and Older Adults J of Prev Med and Public Health 2012 45 3 188 195 10.3961/jpmph.2012.45.3.188 22712046 14. Pit SW Shretha R Schofield D Passey M Health problems and retirement due to ill-health among Australian retirees age 45–64 years Health Policy 2010 94 175 181 10.1016/j.healthpol.2009.09.003 19853957 15. Roberts J. Rice N. Jones AM. Early retirement and inequality in Britain and Germany: How important is health? Health, Econometrics and Data Group. HEDG Working Paper 08/27. Luxembourg: The University of York; 2008. 16. Schuring M Mackenbach J Voorham T Burdorf A The effect of re-employment on perceived health J of Epidemiology & Community Health 2011 65 639 644 10.1136/jech.2009.103838 17. Siegrist J Wahrendorf M von dem Knesebeck O Jürges H Bösch-Supan A Quality of work, well-being and intended early retirement of older employees - baseline results from the SHARE Study European Journal of Public Health 2007 17 1 62 68 10.1093/eurpub/ckl084 16777840 18. Valdes AM Andrew T Gardner JP Kimura M Oelsner E Cherkas LF Aviv A Spector TD Obesity, cigarette smoking and telomere length in women Lancet 2005 366 662 664 10.1016/S0140-6736(05)66630-5 16112303 19. Moen P A Life Course Perspective on Retirement, Gender and Well-Being J Occup Health Psychology 1996 1 2 131 144 10.1037/1076-8998.1.2.131 20. Doyle YG McKee M Sherriff M A model of successful ageing in British populations Eur J Public Health 2012 22 1 77 76 10.1093/eurpub/ckq132 21. Muto T Sugisawa H Kim H Kobayashi E Fukaya T Sugihara Y Shibata H Kumashiro M Health status and lifestyles of elderly Japanese workers Aging and Work 2003 London Taylor & Francis 72 84 22. Rennemark M Wellbeing in old age. Life history evolutions, sense of coherence and social networks in relation to health 1999 Lund Department of Psychology Lund University 23. Vahtera J Westerlund H Hall M Sjösten N Kivimäki M Salo P Effect of retirement on sleep disturbances: the GAZEL prospective cohort study Sleep 2009 32 1459 1466 19928385 24. Laaksonen M Metsä-Simola N Martikainen P Pietiläinen O Rahkonen O Gould R Partonen T Lahelma E Trajectories of mental health before and after old-age and disability retirement: a register-based study on purchases of psychotropic drugs Scand J Work Environ Health 2012 38 5 409 417 10.5271/sjweh.3290 22411588 25. Sjösten N Nabi H Westerlund H Singh-Manoux A Dartigues J-F Goldberg M Zin M Oksanen T Salo P Pentti J Kivimäki M Vahtera J Influence of retirement and work stress on headache prevalence: A longitudinal modelling study from GAZEL Cohort Study Cephalagia 2010 31 6 696 705 10.1177/0333102410394677 26. Westerlund H Kivimäki K Sing-Manoux A Melchior M Ferrie JE Pentti J Jokela M Lein-eweber C Goldberg M Zins M Vahtera J Self-rated health before and after retirement in France (GAZEL): a cohort study Lancet 2009 374 1889 1896 10.1016/S0140-6736(09)61570-1 19897238 27. Bailis D Segall A Chipperfield JG Soc Sci Med 2003 56 203 217 10.1016/S0277-9536(02)00020-5 12473308 28. Layes A Asada Y Kephart G Whiners and deniers - What does self-rated health measure? Soc Sci Med 2012 75 1 9 10.1016/j.socscimed.2011.10.030 22265085 29. Jylhä M What is self-rated health and why does it predict mortality? Towards a unified conceptual model Soc Sci Med 2009 69 307 316 10.1016/j.socscimed.2009.05.013 19520474 30. Nordenfelt L Livskvalitet och Hälsa - Teori och kritik. Quality of Life and Health - Theory and Criticism 1991 Stockholm Almquist & Wiksell 31. Karlsson NE Carstens JM Gjesdal S Alexandersson KAE Work and Health. Risk factors for disability pension in a population-based cohort of men and women on long-term sick leave in Sweden Eur J of Public Health 2008 18 3 224 231 10.1093/eurpub/ckm128 18245150 32. Karpansalo M Manninen P Kauhanen J Lakka T Salonen JT Perceived health as a predictor of early retirement Scand J Work Environ Health 2004 30 4 287 292 10.5271/sjweh.796 15458011 33. Pietiläinen O Laaksonen M Rahkonen O Lahelma E Self-rated Health as a Predictor of Disability Retirement - The Contribution of Ill-Health and Working Conditions PLoS One 2011 6 9 e25004 10.1371/journal.pone.0025004 21949830 34. Idler EL Benyamini Y Self-rated Health and Mortality: A review of Twenty-Seven Community Studies J of Health and Social Behaviour 1997 38 21 37 10.2307/2955359 35. Mossey JM Shapir E Self-Rated Health: A Predictor of Mortality Among the Elderly Am J of Public Health 1982 72 8 800 808 10.2105/AJPH.72.8.800 7091475 36. Singh-Manoux A Gueguen A Martikainen P Ferrie J Marmot M Shipley M Self-Rated Health and Mortality: Short- and Long-Term Associations in the Whitehall II Study Psychosom Med 2007 69 138 143 10.1097/PSY.0b013e318030483a 17289825 37. Strawbridge WJ Wallhagen MI Self-rated health and Mortality over three decades Res Aging 1999 21 3 402 416 10.1177/0164027599213003 38. Seitsamo J Klockars M Aging and changes in health Scand J Work Environ Health 1997 23 1 27 35 9247993 39. Beehr TA The process of retirement: a review and recommendations for future investigation Per Psychol 1986 39 31 55 10.1111/j.1744-6570.1986.tb00573.x 40. Beehr TA Glazer S Nielson NL Farmer SJ Work and Nonwork Predictors of Employees’ Retirement Age J of Vocational Behav 2000 57 206 225 10.1006/jvbe.1999.1736 41. Nilsson K Pension eller arbetsliv? Pension or working life? 2005 Malmö Swedish National Institute in Working Life 42. Nilsson K Rignell-Hydbom A Rylander L Factors influencing the decision to extend working life or to retire Scand J Work Environ Health 2011 37 6 473 480 10.5271/sjweh.3181 21725583 43. Nilsson K To work or not to work in an extended working life? Factors in working and retirement decision 2013 Lund Faculty of Medicine, Lund University 44. Nilsson K. Conceptualisation of ageing in relation to factors of importance for extending working life - a review. Scand J Public Health 2016, online first, doi:1403494816636265 45. Cobb-Clark DA Stillman S The Retirement Expectations of Middle-aged Australians Econ Rec 2009 85 269 146 163 10.1111/j.1475-4932.2009.00543.x 46. Nilsson K. Äldre medarbetares attityder till ett långt arbetsliv. Skillnader mellan olika yrkesgrupper inom hälso- och sjukvården. [Older workers’ attitude to an extended working life. Differences between occupations in health and medical care]. Arbetsliv i omvandling 2006:10. Stockholm: Swedish National Institute of Working Life; 2006. (In Swedish). 47. Ilmarinen J Kumashiro M Promotion of Work Ability During Aging Aging and Work 2003 London Taylor & Francis 23 39 48. Li C-Y Sung F-C A review of the healthy worker effect in occupational epidemiology Occup Med 1999 49 225 229 10.1093/occmed/49.4.225

PMC005xxxxxx/PMC5000416.txt

==== Front BMC Res NotesBMC Res NotesBMC Research Notes1756-0500BioMed Central London 220910.1186/s13104-016-2209-xTechnical NoteTranscriptomic SNP discovery for custom genotyping arrays: impacts of sequence data, SNP calling method and genotyping technology on the probability of validation success http://orcid.org/0000-0001-6841-1269Humble Emily emily.humble@uni-bielefeld.de 12Thorne Michael A. S. mast3@cam.ac.uk 2Forcada Jaume jfor@bas.ac.uk 2Hoffman Joseph I. joseph.hoffman@uni-bielefeld.de 11 Department of Animal Behaviour, University of Bielefeld, Postfach 100131, 33501 Bielefeld, Germany 2 British Antarctic Survey, High Cross, Madingley Road, Cambridge, CB3 OET UK 26 8 2016 26 8 2016 2016 9 1 41831 3 2016 6 8 2016 © The Author(s) 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Single nucleotide polymorphism (SNP) discovery is an important goal of many studies. However, the number of ‘putative’ SNPs discovered from a sequence resource may not provide a reliable indication of the number that will successfully validate with a given genotyping technology. For this it may be necessary to account for factors such as the method used for SNP discovery and the type of sequence data from which it originates, suitability of the SNP flanking sequences for probe design, and genomic context. To explore the relative importance of these and other factors, we used Illumina sequencing to augment an existing Roche 454 transcriptome assembly for the Antarctic fur seal (Arctocephalus gazella). We then mapped the raw Illumina reads to the new hybrid transcriptome using BWA and BOWTIE2 before calling SNPs with GATK. The resulting markers were pooled with two existing sets of SNPs called from the original 454 assembly using NEWBLER and SWAP454. Finally, we explored the extent to which SNPs discovered using these four methods overlapped and predicted the corresponding validation outcomes for both Illumina Infinium iSelect HD and Affymetrix Axiom arrays. Results Collating markers across all discovery methods resulted in a global list of 34,718 SNPs. However, concordance between the methods was surprisingly poor, with only 51.0 % of SNPs being discovered by more than one method and 13.5 % being called from both the 454 and Illumina datasets. Using a predictive modeling approach, we could also show that SNPs called from the Illumina data were on average more likely to successfully validate, as were SNPs called by more than one method. Above and beyond this pattern, predicted validation outcomes were also consistently better for Affymetrix Axiom arrays. Conclusions Our results suggest that focusing on SNPs called by more than one method could potentially improve validation outcomes. They also highlight possible differences between alternative genotyping technologies that could be explored in future studies of non-model organisms. Electronic supplementary material The online version of this article (doi:10.1186/s13104-016-2209-x) contains supplementary material, which is available to authorized users. Keywords TranscriptomeRoche 454 sequencingIllumina HiSeq sequencingSingle nucleotide polymorphismValidation successMarine mammalAntarctic fur sealArctocephalus gazellahttp://dx.doi.org/10.13039/501100001659Deutsche ForschungsgemeinschaftHO 5122/3-1Marie Curie FP7 Reintegration GrantPCIG-GA-2011-303618Hoffman Joseph I. issue-copyright-statement© The Author(s) 2016 ==== Body Background High throughput sequencing and cost efficient genotyping technologies are revolutionising the study of wild organisms [1]. For example, many thousands of single nucleotide polymorphisms (SNPs) can now be genotyped in virtually any organism [2, 3]. Although individually less informative than multi-allelic markers, SNPs are appealing because they can be genotyped rapidly, in large numbers and with minimal error [4, 5]. Consequently, SNP datasets are being generated for an increasing number of wild animal populations, allowing researchers to address a variety of outstanding questions in evolutionary biology, conservation genetics and wildlife management [6–8]. In non-model species, SNPs are often mined from transcriptome assemblies, as these are smaller and simpler to generate than genomes. Nevertheless, there are a variety of alternative methods available for read mapping and variant discovery and it is not always straightforward to know which of these to use. Relatively few systematic comparisons of the available programs have been carried out and most have mainly been based on genomic data from humans [9, 10]. These studies suggest that in some cases the concordance between different methods can be poor [11, 12], yet it is still the norm to call SNPs with a single method [13–15]. By drawing upon vast numbers of previously known SNPs, human studies have also evaluated the relative success of different methods at discovering known variants [16]. However, less attention has been paid to non-model organisms, partly because for many species, SNPs are being discovered for the first time. SNP discovery can facilitate multiple genotyping approaches. Genotyping by sequencing approaches such as RAD, ddRAD and 2bRAD [17–19] allow simultaneous SNP discovery and genotyping. These approaches enable studies to scale up to much larger sample sizes of individuals and loci than what was possible with traditional markers such as microsatellites. However, large amounts of high quality DNA are required, library preparation can be costly and labour intensive, and downstream analyses are not straightforward [20]. High density SNP arrays, or ‘SNP chips’, have thus become increasingly popular for large-scale studies, as they are relatively cheap per sample, technically more straightforward, allow selected SNPs to be consistently genotyped across the majority of individuals, and enable candidate genes to be targeted [21, 22]. However, careful selection of SNPs is necessary as not all ‘putative’ SNPs will be suitable for genotyping. For example, SNPs must have sufficient flanking sequence that is compatible with a given genotyping technology. The two most widely used array platforms, Illumina Infinium iSelect HD [23] and Affymetrix Axiom [24], implement distinctive hybridization technologies and require probes of different lengths. Moreover, recent studies suggest that the genomic context of a SNP can have a significant impact on validation success [25, 26], defined as the propensity of a given SNP to be polymorphic and reliably scored in a sample of individuals. For example, transcripts representing paralogous genes can result in SNP probe sequences that map many times to a genome, whilst probe sequences inadvertently spanning intron–exon boundaries will result in failure of the probe to bind to the genomic DNA [25, 27–29]. By mapping SNP flanking sequences to reference genomes, both of these issues have been shown to have a significant impact on validation success [25, 26, 30–32]. An opportunity to quantify the extent of overlap between different SNP discovery methods and to explore the consequences for validation success is provided by a study of Antarctic fur seals (Arctocephalus gazella). A transcriptome assembly based on Roche 454 sequencing is already available for this species, from which two SNP datasets were generated using NEWBLER and SWAP454 respectively [33, 34]. Here, we supplement this transcriptome with short read Illumina sequencing, allowing a comparison of SNP discovery methods tailored to different types of sequence data. We also recently developed a predictive modeling framework to determine the likelihood of validation success by accounting for a variety of variables, from compatibility of the probe sequences with a given assay chemistry, through in silico features such as depth of coverage and minor allele frequency (MAF), to aspects of the genomic context [26]. This framework provides a basis by which we can evaluate the likely validation outcomes of the SNPs discovered by different methods. In this study, we first generated a ‘hybrid’ fur seal transcriptome from the 454 and Illumina data. We then mapped the Illumina reads to the hybrid transcriptome using BWA and BOWTIE2 before calling SNPs from each alignment with GATK. The two sets of resulting SNPs were then compared with the two sets of SNPs previously mined from the 454 transcriptome using NEWBLER and SWAP454 respectively. This allowed a direct comparison of a total of four methods for calling SNPs from two types of sequence data. Finally, we used predictive modeling to assess the suitability of the resulting SNPs for both an Affymetrix Axiom and an Illumina Infinium iSelect HD array. We hypothesized that SNPs with a high probability of validation success would be enriched for those called by more than one method. Due to the higher depth of coverage provided by Illumina relative to 454 sequencing, we also expected SNPs called from the former to have higher validation success probabilities. We provide an annotated workflow within the R programming language [35] for implementing the SNP filtering and assessment steps presented here (Additional file 1). Results Sequencing, assembly and annotation To improve upon the existing 454 transcriptome, which comprises 23,096 contigs of mean length 971 bp, we conducted an additional round of Illumina sequencing (see ‘Methods’ section for details). This generated a total of 17,894,042 101 bp paired-end reads (submitted to the sequence read archive, http://www.ncbi.nlm.nih.gov/sra; Study Accession SRP071273), which were assembled de novo to generate 26,266 contigs of mean length 904 bp. Blasting these contigs to the 454 backbone, we found that 15,520 (59.0 %) successfully mapped at an e-value threshold of 1e−10. After annotating the unmapped contigs, around 50 % were removed, either due to a lack of homology to known sequences or because the top BLAST hit revealed similarity to known bacterial or viral sequences. Most of the remaining 5452 annotated contigs showed sequence similarity to the Weddell seal (Leptonychotes weddellii) and the walrus (Odobenus rosmarus) and were thus concatenated to the original 454 transcriptome. This yielded a ‘hybrid transcriptome’ comprising a total of 28,548 contigs (Fig. 1, http://www.goo.gl/vj8VjD). To investigate homology to the dog (Canis familiaris), we mapped these contigs to the most recent and complete build of the dog transcriptome. 23,587 (82.6 %) of the seal contigs mapped to 35,724 (64.7 %) of the dog transcripts, suggesting that a reasonably large proportion of the fur seal transcriptome has been captured.Fig. 1 Circular plot showing the hybrid transcriptome assembly. The inner track represents the breakdown of the transcriptome into 454 (purple) and Illumina (blue) components. The middle and outer tracks show the depth of coverage of the 454 and Illumina reads plotted on a log scale. Transcripts are sorted in order of average Illumina coverage. As we required at least ten fold Illumina coverage of a given nucleotide to call a SNP, Illumina coverage of transcripts with less than tenfold average coverage has been truncated zero Overlap in SNP discovery The 454 transcriptome was previously mined for SNPs using NEWBLER and SWAP454, which identified 14,538 and 11,155 SNPs respectively [34]. To call SNPs from the Illumina data, we mapped the raw Illumina reads to the hybrid transcriptome using BWA and BOWTIE2 and parsed the resulting alignment files to GATK as described in the ‘Methods’ section. This resulted in a total of 18,353 SNPs from the BWA alignment and 15,109 from the BOWTIE2 alignment, of which 14,490 SNPs were called by both methods. Pooling SNPs across all four methods resulted in a dataset of 34,718 unique markers. To explore the extent of overlap between the SNP calling methods described above we generated a Venn diagram (Fig. 2). This shows that 49.1 % of the total 34,718 SNPs were called by a single method, 38.3 % were called by two methods, 4.6 % by three and 7.0 % by all four. Most of the SNPs identified by a single method (76.9 %) were called from the 454 transcriptome using NEWBLER or SWAP454. The overlap between SNPs discovered from the 454 and Illumina data was 13.5 %.Fig. 2 Venn diagram showing the extent of overlap among SNPs called using four different methods (see ‘Methods’ section for details) SNP parameter space The increased depth of coverage provided by Illumina sequencing should allow in silico minor allele frequency (MAF) to be estimated more accurately than for the 454 data. We therefore selected the subset of 4679 SNPs that were called from both the 454 and Illumina datasets and compared their respective parameter spaces. Two obvious differences emerge between the two datasets (Fig. 3). First, average log depth of coverage of the SNPs increases substantially, from around 1.2 (corresponding to 16× coverage) for the 454 data to 1.7 (corresponding to 50× coverage) for the Illumina data. Second, we find a marked difference in the respective MAF distributions, which are concentrated around 0.4 for the 454 data (Fig. 3a) but which are more evenly spread between around 0.2 and 0.5 for the Illumina data (Fig. 3b).Fig. 3 Variation in SNP minor allele frequency (MAF) and depth of sequence coverage. The upper panels correspond to 4679 SNPs that were called from both the 454 and Illumina datasets, with panel a showing the 454 parameter space and b showing the corresponding Illumina parameter space. The lower panels correspond to the total number of SNPs called from the 454 and Illumina data (20,426 and 18,971 respectively), with panel c showing the 454 parameter space and d showing the corresponding Illumina parameter space We also used the same approach to compare all of the SNPs called from the 454 data with all of the SNPs called from the Illumina data. Again we found marked differences between the two datasets (Fig. 3). For the 454 data, a clear relationship emerged between MAF and depth of coverage, SNPs with high MAF mainly being called at a relatively low depth of coverage, whereas SNPs with low MAF were mainly called at a relatively high depth of coverage (Fig. 3c). For the Illumina data, SNPs were predominantly called at a relatively low depth of coverage (Fig. 3d), which is probably a more accurate approximation of the underlying MAF distribution (see ‘Discussion’ section). SNP filtering and predicted assay success Although most studies present the total number of putative SNPs identified from transcriptome assemblies, when developing a custom SNP array it is important to consider the likelihood of each SNP successfully validating with a given genotyping technology. We therefore tested the total set of 34,718 SNPs for compatibility with both Illumina Infinium iSelect HD and Affymetrix Axiom high density SNP arrays (Fig. 4). In order to do this, we extracted the flanking sequences required for Infinium iSelect (121 bp) and Affymetrix Axiom (71 bp) probe design from the fur seal transcriptome. Complete 121 bp flanking sequences could be extracted for 31,192 of the SNPs (89.8 %) while the equivalent proportion was slightly higher for the 71 bp flanking sequences (n = 32,727, 94.3 %, Step 1 in Fig. 4). The Illumina and Affymetrix flanking sequences were then evaluated using Illumina’s Assay Design Tool (ADT) and Affymetrix’s SNP evaluation pipeline respectively. 26,110 SNPs (86.6 %) were assigned ADT scores of ≥0.8 and 24,778 (78.4 %) were classified as either ‘recommended’ or ‘neutral’ by Affymetrix (Step 2 in Fig. 4).Fig. 4 Flow diagram showing the number of SNPs remaining after each step of the SNP detection pipeline for both an Illumina Infinium iSelect HD array (blue circles) and an Affymetrix Axiom array (purple circles) Following this, we sought to remove SNPs with an undesirable genomic context by mapping their flanking sequences to the draft fur seal genome (Step 3 in Fig. 4). Blasting the Infinium and Affymetrix SNP sequences with an e-value threshold of 1e−12 recovered 24,247 and 22,368 hits respectively. For these SNPs, we evaluated the probability of successful validation using a predictive model incorporating MAF, depth of coverage, ADT/p-convert score plus values of the predictor variables generated from the genome BLAST (see ‘Methods’ section for details). Based on the 121 bp Infinium sequences, 19,773 (81.5 %) SNPs were predicted to successfully validate with a probability threshold of 0.7. Both the number (21,141) and proportion (94.5 %) of equivalent Affymetrix sequences were higher. Simply filtering the flanking sequences for those that mapped completely and uniquely to the reference genome resulted in fewer SNPs being retained (11,057 Illumina flanking sequences and 14,901 Affymetrix flanking sequences, Fig. 4). We next asked whether the probability of successful validation varied according to SNP calling method. Of the SNPs called from the 454 data using NEWBLER and SWAP454, only 46.8 and 57.0 % respectively were predicted to successfully validate when using an Illumina assay (Table 1). By contrast, 75.7 % of SNPs called by GATK from the BOWTIE2 alignment and 72.1 % from the BWA alignment were predicted to successfully validate. A similar pattern was obtained when considering SNPs that map completely and uniquely to the reference genome, as well as for predictive models based on the Affymetrix flanking sequences (Table 1).Table 1 Proportion of SNPs from each discovery method predicted to successfully validate on both an Illumina Infinium and an Affymetrix Axiom array using predictive modeling and simple filtering approaches Discovery method Predicted validation success (%) Infinium Axiom Predictive Filtering Predictive Filtering BOWTIE2 75.7 45.6 83.3 61.7 BWA 72.1 39.7 78.5 54.6 NEWBLER 46.8 27.3 48.9 35.5 SWAP454 57.0 34.6 61.8 45.9 Finally, we tested whether the probability of successful validation varied with the number of methods by which a given SNP was called. Table 2 shows that, when using an Illumina assay, regardless of whether a predictive modeling or simple filtering approach is taken, predicted validation success rates are around one-third to two times higher for SNPs called by two or more methods, with those called by two methods yielding the greatest predicted validation success. The same pattern is found for the Affymetrix flanking sequences, although the predicted outcomes are somewhat less dependent on the number of methods by which a SNP is called.Table 2 Proportion of those SNPs shared by one, two, three and four calling methods predicted to successfully validate on both an Illumina Infinium and an Affymetrix Axiom array using using predictive modeling and simple filtering approaches Share Predicted validation success (%) Infinium Axiom Predictive Filtering Predictive Filtering One 66.8 30.7 91.7 57.0 Two 92.9 57.2 96.7 72.6 Three 89.3 52.5 93.2 68.0 Four 89.9 54.0 93.3 68.2 Discussion We used Illumina sequencing to augment an existing fur seal transcriptome assembly generated from 454 sequence data. We then attempted to maximise successful SNP discovery both by exploring the overlap between SNPs called using four different methods and by evaluating predicted validation outcomes. We found that SNPs called from the Illumina data were on average more likely to successfully validate, as were SNPs called by more than one method. Predicted validation outcomes were also found to be slightly better for Affymetrix Axiom than Illumina Infinium iSelect HD arrays. The hybrid transcriptome assembly We de novo assembled the Illumina HiSeq data into contigs and then mapped these to the 454 backbone. Over 5000 additional contigs were generated that revealed homology to walrus and Weddell seal sequences, suggesting that the hybrid assembly is more complete than the 454 assembly (Fig. 1). To explore this further, we mapped the fur seal contigs to the most recent and complete build of the dog transcriptome. We found that 82.6 % of the contigs mapped to 64.7 % of the dog transcripts. This is in contrast to what was previously reported for the 454 transcriptome, where 62.5 % of seal contigs mapped to 77 % of dog transcripts [34]. Therefore, whilst a greater proportion of the transcriptome is mapping, a slightly smaller fraction of the dog transcriptome is represented. This is probably because the mapping was performed against a more recent and complete build of the dog transcriptome. SNP discovery The greater depth of coverage and improved representation of fur seal transcripts provided by Illumina sequencing provides the opportunity both to increase the total pool of SNPs discovered and to cross-check SNPs called from the 454 and Illumina data. In this study, we compared four different methods for mining SNPs from two different types of sequence data. Specifically, 454 data were mined for SNPs using NEWBLER and SWAP454, whilst GATK was used to mine SNPs from both a BWA and a BOWTIE2 Illumina read alignment. We found poor concordance between the SNPs discovered by all four methods, with only 51.0 % of SNPs being discovered by two or more methods. This is consistent with previous studies, mostly based on genomic data from humans, which have also found relatively little overlap between SNPs called by different tools [12, 16] although few of these studies attempted to explore validation outcomes as we have done here. There are several potential explanations for the limited overlap between SNPs called from the 454 and Illumina datasets. First, the hybrid transcriptome contains around 5000 contigs that are only represented by Illumina sequences and from which any called SNPs will therefore be unique. However, these only account for 5.7 % of the total number of Illumina-specific SNPs, suggesting that the majority are located within contigs that are also represented by 454 data. Thus, it seems likely that Illumina sequencing allowed many more SNPs to be called from the same contigs by virtue of the increased depth of coverage provided. This is supported by two lines of evidence. First, the median depth of coverage of SNPs called from the Illumina data was 28, whereas the equivalent was only 16 for the 454 data. Second, we observed a shift towards SNPs with relatively low minor allele frequencies being called from the Illumina data, suggesting that greater depth of coverage facilitates the discovery of such polymorphisms. A second reason for the limited overlap could be that the 454 transcriptome includes both skin and necropsy samples whereas for the current round of Illumina sequencing we were only able to use remaining cDNA from the skin samples. Thus, the 454-specific SNPs were called from both the skin and necropsy parts of the transcriptome, whereas the Illumina-specific SNPs were only called from the skin part. Indeed, for both BWA and BOWTIE2 alignments, not all of the 454 transcriptome was mapped to by the Illumina reads (Fig. 1); around 40 % was left with insufficient Illumina sequence coverage for SNP calling, presumably because it represented necropsy-specific transcripts. Another possibility is that not all of the SNPs called from the 454 data may be genuine. In support of this, only 25.6 % of the 454-specific SNPs were called by both NEWBLER and SWAP454, suggesting that the two programs differ considerably in their outputs even for the same sequence resource. Regardless of the differences between SNPs called from the 454 and Illumina data, it is noteworthy that we also found some degree of overlap. Almost 5000 SNPs in total were called from what are essentially independent sequence datasets. For this reason, we consider these SNPs more likely to be genuine, consistent with the finding that SNPs called by more than one method are more likely to be suitable for use in a high density SNP array (see below). Direct comparison of SNPs called from the 454 and Illumina data also revealed marked differences in their MAF distributions, the former being dominated by SNPs with a MAF of around 0.4 while the latter show a more even MAF distribution. While we cannot yet say which of these is the most accurate portrayal of the true underlying distribution, we suspect that the Illumina data are closer to the mark because, at least in theory, greater depth of coverage should allow in silico allele frequency distributions to be estimated more accurately. This finding could thus explain why studies often find no association between in silico and realised allele frequencies [36–38]. Exploring validation success SNP discovery is an important goal of many studies and features prominently in many publications describing transcriptomes [39–41]. However, the resulting SNPs may not provide a reliable indication of the number that are likely to successfully validate with a given genotyping technology. For this it is necessary to account for variables such as (i) the proportion of SNPs for which complete flanking sequences can be extracted; (ii) compatibility of the SNP flanking sequences with the chosen assay chemistry; (iii) variation in the likelihood of a SNP being genuine with MAF and depth of coverage; and (iv) aspects of the genomic context including sequence uniqueness and proximity to intron–exon boundaries. We therefore incorporated the above factors into the predictive framework of Humble et al. [26] to evaluate the probability of each SNP successfully validating on both Illumina Infinium iSelect HD and Affymetrix Axiom genotyping arrays. A number of patterns emerged. First, the proportion of SNPs for which complete flanking sequences could be extracted was lower for Illumina than Affymetrix (86.8 versus 91.0 % respectively) reflecting Illumina’s requirement for substantially longer flanking sequences (121 versus 71 bp respectively) for probe design. Second, a larger proportion of SNPs was deemed suitable for assay design based on Illumina ADT scores than Affymetrix p-convert scores (86.6 versus 78.4 % respectively). This pattern is reflected in the proportion of SNPs predicted to successfully validate with each technology, which was over ten percent higher for Affymetrix (94.5 %) than Illumina (81.5 %). Although Illumina require longer flanking sequences for assay design, the probes themselves are only 60 bp long (plus a one base terminal SNP site). Therefore, the difference in predicted validation rates seems unlikely to be related to probe length. Instead, it could be possible that Affymetrix’s evaluation pipeline is more stringent, potentially in this case because it utilized the fur seal genome to determine strand specificity. Regardless of the exact reasons, our findings suggest that under certain circumstances Affymetrix Axiom genotyping arrays might be preferable in some respects to Illumina Infinium iSelect HD arrays, particularly when genotyping non-model organisms with SNPs that have not been experimentally validated in advance. We also tested whether the probability of successful validation varied according to the method by which a given SNP was called. Above and beyond the pattern described above, we found that SNPs called only from the 454 data (using either NEWBLER or SWAP454) were less likely on average to successfully validate than SNPs called only from the Illumina data (using BOWTIE2 or BWA in combination with GATK). This suggests that a larger proportion of SNPs called from the 454 data may be spurious, in line with the lower depth of coverage of the 454 data, the fact that only around a quarter of these SNPs were called by both NEWBLER and SWAP454, and the limited overlap between SNPs called from the 454 and Illumina data. This finding would also be consistent with our previous work on fur seals in which we experimentally validated a panel of putative SNPs derived from the 454 transcriptome using Illumina’s GoldenGate assay [37]. This study found a positive relationship between in silico MAF and validation success, which suggests that some of the assays may have been designed from paralogous loci. Finally, we found that the probability of successful validation was greater for SNPs detected using more than one method than for SNPs flagged by a single method. The highest overall validation success rate was obtained for SNPs called by two methods while a marginal reduction was found for SNPs called by three or four methods. To explore this further, we calculated the proportion of the total number of SNPs called by each of the four methods separately for SNPs called by one, two, three or four methods respectively. We found that the peak in validation success corresponding to SNPs called by two methods can be explained by a greater proportion of those SNPs having been called by GATK after mapping with either BOWTIE2 or BWA (Additional file 1: Figure S1). By contrast, SNPs called by three or four methods were more likely to have been called by NEWBLER or SWAP454. As previously discussed, the latter may be of lower average quality and therefore appear to contribute towards a slight deterioration in predicted validation success rates for SNPs called by three and four methods. Despite the above, a general tendency for SNPs called by more than one method to be more likely to successfully validate makes good sense because the more methods that are used to call a given SNP, the more robust that SNP should be to the peculiarities of any single computer program. Thus, we would advocate the use of more than one SNP calling method as a means of identifying the most robust SNPs, particularly when resources are limited and a high rate of validation success is an important outcome. Overall, our results also highlight how Illumina sequencing is preferable for SNP discovery given the substantially greater depth of coverage that it provides. Conclusions We used Illumina sequencing to improve upon an existing fur seal transcriptome assembly. We then attempted to maximise successful SNP discovery both by exploring the overlap between SNPs called using four different methods and by evaluating predicted validation outcomes. We found that SNPs called from the Illumina data had higher likelihoods of successful validation, as did SNPs called by more than one method. Predicted validation outcomes were also found to be consistently better for Affymetrix Axiom than Illumina Infinium iSelect HD arrays. One possible means of exploring the relative merits of these two genotyping technologies would be to genotype a set of individuals and SNPs using both technologies. Methods Initial transcriptome This study partly builds upon a previously published fur seal transcriptome assembly. This was constructed using 454 sequence reads generated from two different cDNA libraries, one comprising skin samples from 12 individuals [33] and the other comprising necropsy samples from nine individuals [34]. Assembly of these data using NEWBLER generated a total of 23,096 contigs [34], which we refer to as the ‘454 transcriptome’. Library preparation and Illumina sequencing Using RNA from the same 12 individuals used for the skin transcriptome, we generated cDNA libraries using Illumina’s TruSeq® Stranded protocol. Briefly, poly-A containing mRNA molecules were purified from the pooled total RNA using oligo-dT beads. The mRNA was subsequently fragmented and reverse transcribed into cDNA with strand specificity. Adaptors and a single ‘A’ base were attached to each fragment before purificiation and PCR enrichment in order to generate the final cDNA library. This was sequenced on one lane of an Illumina HiSeq 2000. Sequence assembly Raw sequencing reads with a Phred quality score of less than 20 were removed and primer and adaptor sequences were trimmed prior to assembly. Cleaned reads were assembled together using SOAPdenovo. After running a range of kmer sizes to determine the optimal k value for contig length and number, the kmer run of 23 was chosen. Only transcripts of length greater than 500 bp were retained in the final assembly. Mapping and sequence annotation All newly generated Illumina contigs were mapped to the previously assembled 454 transcriptome using blastn in BLAST at an e-value threshold of 1e−10. Contigs that did not result in a significant BLAST match were annotated using the non-redundant sequence database at an e-value threshold of 1e−10. Transcripts with putative gene products of bacterial and viral origin were removed whilst all remaining annotated contigs were concatenated to the 454 transcriptome, which we refer to as the ‘hybrid transcriptome’. To determine the completeness of the improved transcriptome, we mapped the assembled fur seal contigs against the most recent set of annotated dog transcripts [http://www.ncbi.nih.gov/genomes/Canis_familiaris/RNA/] using blastn in BLAST at an e-value threshold of 1e−10. SNP discovery To mine SNPs from the hybrid transcriptome, we generated two bam files by mapping the raw Illumina paired-end reads to the hybrid transcriptome using both BWA and BOWTIE2 with the default parameters. Each mapping file was then parsed to GATK for SNP detection using the UnifiedGenotyper tool (-stand_call_conf 30, -stand_emit_conf 10). Each set of SNP calls was then hard-filtered using GATK’s VariantFiltration tool based on the following criteria: fisher strand bias <30, quality by depth >2, unfiltered read depth ≥10, read mapping quality ≥40. SNPs consequently flagged with anything other than ‘PASS’ were removed from the datasets. We also removed SNPs if read support for the minor allele was less than three. In order to determine the extent of overlap between SNPs called by different methods, we revisited two sets of SNPs called from the 454 transcriptome using NEWBLER and SWAP454 respectively [33, 34]. A small number of SNPs within these datasets were duplicated or had an alternative allele frequency of one. These were therefore removed, leaving a total of 14,536 NEWBLER SNPs and 11,135 SWAP454 SNPs. SNP filtering and predicted assay success We generated a global list of SNPs representing all of those called from (i) the 454 transcriptome using NEWBLER and SWAP454 and (ii) the hybrid transcriptome using BWA and BOWTIE2 in combination with GATK. We then implemented the steps outlined below to obtain subsets of SNPs suitable for designing Illumina Infinium iSelect HD and Affymetrix Axiom SNP assays respectively. Firstly, we used the BEDTOOLS command getfasta to extract the 121 bp SNP flanking sequences required for Illumina assays and the 71 bp flanking sequences required for Affymetrix assays. Loci with insufficient flanking sequence were discarded, as were a small number of SNPs that did not match the corresponding base in the genome sequence. The suitability of the resulting flanking sequences for each assay’s hybridization technology was then determined by generating Illumina Assay Design Tool (ADT) scores for the 121 bp SNP flanking sequences and Affymetrix p-convert scores for the 71 bp SNP flanking sequences. These were obtained from both Illumina and Affymetrix directly. SNPs assigned an ADT score of <0.8 were discarded from the Infinium dataset. For the Affymetrix dataset, SNPs with forward and/or reverse sequences designated ‘not recommended’ or ‘not possible’ were discarded. For each SNP, we recorded the depth of coverage, minor allele frequency (MAF), ADT score (for Illumina assays) or p-convert score (for Affymetrix assays). We then mapped the corresponding Illumina and Affymetrix flanking sequences to the Antarctic fur seal reference genome [26] using blastn in BLAST with an e-value threshold of 1e−12. From this, we determined the alignment length of the top blast hit (a full and continuous mapping indicates that a SNP and its flanking sequences lie fully within an exon) and the total number of mappings (a proxy for sequence uniqueness). Given the above information, we used two approaches to identify SNPs with high likelihoods of validation success for each SNP. First we simply filtered for SNPs whose flanking sequences match completely and uniquely to the genome, as these two characteristics have been shown to have a major affect on validation success [26]. Second, we used a predictive modeling approach based on the outcome of a pilot assay in which 144 putative fur seal SNPs were genotyped in 480 individuals [37]. Here, the known genotyping outcomes were used together with the genomic characteristics of the 144 SNP flanking sequences to construct a model of SNP validation success using k-fold cross validation. This approach splits the 144 observations into k = 5 non-overlapping subsets of approximately equal size, uses one subset as a validation sample and the remaining four subsets as training data in order to generate the best predictive model. This best model was then used to output the probability of each SNP successfully validating given values of the predictor variables using the predict function in the bestglm package in R [26]. A given SNP was predicted to validate successfully if its associated probability value was above 0.7. Additional file 10.1186/s13104-016-2209-x SNPs called by one, two, three or four methods, broken down by calling method, averaged across technology and filtering approach. Authors’ contributions JIH and EH conceived and designed the study. JF contributed materials. EH and MAST analysed the data. EH and JIH wrote the first version of the manuscript. All authors read and approved the final manuscript. Acknowledgements We thank the British Antarctic Survey field assistants on Bird Island for collection of tissue samples. We thank Shilo Dickens at the University of Cambridge for cDNA library preparation and sequencing. We would also like to thank three anonymous reviewers for their comments. Competing interests The authors declare that they have no competing interests. Availability of data and material The Illumina reads have been submitted to the sequence read archive (http://www.ncbi.nlm.nih.gov/sra) under Accession Number SRP071273. The hybrid transcriptome assembly and the dataset of all unique SNPs are available at http://www.goo.gl/vj8VjD. Computer code and documentation are available as an HTML file written in Rmarkdown (Additional file 1). A GitHub repository containing the data and scripts for the analysis is available at https://www.goo.gl/57jRgu. Funding This work contributes to the Ecosystems project of the British Antarctic Survey, Natural Environmental Research Council, and is part of the Polar Science for Planet Earth Programme. It was supported by a Deutsche Forschungsgemeinschaft standard Grant (HO 5122/3-1), a Marie Curie FP7-Reintegration-Grant within the 7th European Community Framework Programme (PCIG-GA-2011-303618) and core funding from the Natural Environment Research Council to the British Antarctic Survey’s Ecosystems Program. ==== Refs References 1. Morin PA Luikart G Wayne RK The SNP workshop group SNPs in ecology, evolution and conservation Trends Ecol Evol 2004 19 208 216 10.1016/j.tree.2004.01.009 2. Senn H Ogden R Cezard T Gharbi K Iqbal Z Johnson E Reference-free SNP discovery for the Eurasian beaver from restriction site-associated DNA paired-end data Mol Ecol 2013 22 3141 3150 10.1111/mec.12242 23432348 3. Johnston SE Lindqvist M Niemelä E Orell P Erkinaro J Kent MP Fish scales and SNP chips: SNP genotyping and allele frequency estimation in individual and pooled DNA from historical samples of Atlantic salmon (Salmo salar) BMC Genomics 2013 14 439 10.1186/1471-2164-14-439 23819691 4. Chen X Sullivan PF Single nucleotide polymorphism genotyping: biochemistry, protocol, cost and throughput Pharmacogenomics J 2003 3 77 96 10.1038/sj.tpj.6500167 12746733 5. Vignal A Milan D SanCristobal M Eggen A A review on SNP and other types of molecular markers and their use in animal genetics Genet Sel Evol 2002 34 275 305 10.1186/1297-9686-34-3-275 12081799 6. Allendorf FW Hohenlohe PA Luikart G Genomics and the future of conservation genetics Nat Rev Genet 2010 11 697 709 10.1038/nrg2844 20847747 7. Ogden R Gharbi K Mugue N Martinsohn J Senn H Davey JW Sturgeon conservation genomics: SNP discovery and validation using RAD sequencing Mol Ecol 2013 22 3112 3123 10.1111/mec.12234 23473098 8. Hoffman JI Simpson F David P Rijks JM Kuiken T Thorne MAS High-throughput sequencing reveals inbreeding depression in a natural population Proc Natl Acad Sci USA 2014 111 3775 3780 10.1073/pnas.1318945111 24586051 9. Yu X Sun S Comparing a few SNP calling algorithms using low-coverage sequencing data BMC Bioinformatics 2013 14 274 10.1186/1471-2105-14-274 24044377 10. Yi M Zhao Y Jia L He M Kebebew E Stephens RM Performance comparison of SNP detection tools with illumina exome sequencing data–an assessment using both family pedigree information and sample-matched SNP array data Nucleic Acids Res 2014 42 e101 10.1093/nar/gku392 24831545 11. O’Rawe J Jiang T Sun G Wu Y Wang W Hu J Low concordance of multiple variant-calling pipelines: practical implications for exome and genome sequencing Genome Med 2013 5 28 10.1186/gm432 23537139 12. Greminger MP Stoelting KN Nater A Goossens B Arora N Bruggmann R Generation of SNP datasets for orangutan population genomics using improved reduced-representation sequencing and direct comparisons of SNP calling algorithms BMC Genomics 2014 15 16 10.1186/1471-2164-15-16 24405840 13. Du L Li W Fan Z Shen F Yang M Wang Z First insights into the giant panda (Ailuropoda melanoleuca ) blood transcriptome: a resource for novel gene loci and immunogenetics Mol Ecol Resour 2015 15 1001 1013 10.1111/1755-0998.12367 25556892 14. Pratlong M Haguenauer A Chabrol O Klopp C Pontarotti P Aurelle D The red coral (Corallium rubrum ) transcriptome: a new resource for population genetics and local adaptation studies Mol Ecol Resour 2015 15 1205 1215 10.1111/1755-0998.12383 25648864 15. Tepolt CK Palumbi SR Transcriptome sequencing reveals both neutral and adaptive genome dynamics in a marine invader Mol Ecol 2015 24 4145 4158 10.1111/mec.13294 26118396 16. Hwang S Kim E Lee I Marcotte EM Systematic comparison of variant calling pipelines using gold standard personal exome variants Sci Rep 2015 5 17875 10.1038/srep17875 26639839 17. Baird NA Etter PD Atwood TS Currey MC Shiver AL Lewis ZA Rapid SNP discovery and genetic mapping using sequenced RAD markers PLoS One 2008 3 e3376 10.1371/journal.pone.0003376 18852878 18. Peterson BK Weber JN Kay EH Fisher HS Hoekstra HE Double digest RADseq: an inexpensive method for de novo SNP Discovery and genotyping in model and non-model species PLoS One 2012 7 e37135 10.1371/journal.pone.0037135 22675423 19. Wang S Meyer E McKay JK Matz MV 2b-RAD: a simple and flexible method for genome-wide genotyping Nat Methods 2012 9 808 810 10.1038/nmeth.2023 22609625 20. Garvin MR Saitoh K Gharrett AJ Application of single nucleotide polymorphisms to non-model species: a technical review Mol Ecol Resour 2010 10 915 934 10.1111/j.1755-0998.2010.02891.x 21565101 21. Syvänen AC Toward genome-wide SNP genotyping Nat Genet 2005 37 Suppl S5 S10 10.1038/ng1558 15920530 22. LaFramboise T Single nucleotide polymorphism arrays: a decade of biological, computational and technological advances Nucleic Acids Res 2009 37 4181 4193 10.1093/nar/gkp552 19570852 23. Gunderson KL Steemers FJ Lee G Mendoza LG Chee MS A genome-wide scalable SNP genotyping assay using microarray technology Nat Genet 2005 37 549 554 10.1038/ng1547 15838508 24. Kennedy GC Matsuzaki H Dong S Liu WM Huang J Liu G Large-scale genotyping of complex DNA Nat Biotechnol 2003 21 1233 1237 10.1038/nbt869 12960966 25. Milano I Babbucci M Panitz F Ogden R Nielsen RO Taylor MI Novel tools for conservation genomics: comparing two high-throughput approaches for SNP discovery in the transcriptome of the European hake PLoS One 2011 6 e28008 10.1371/journal.pone.0028008 22132191 26. Humble E Barrio AM Forcada J A draft fur seal genome provides insights into factors affecting SNP validation and how to mitigate them Mol Ecol 2016 16 909 921 10.1111/1755-0998.12502 27. Wang S Sha Z Sonstegard TS Liu H Xu P Somridhivej B Quality assessment parameters for EST-derived SNPs from catfish BMC Genomics 2008 9 450 10.1186/1471-2164-9-450 18826589 28. Helyar SJ Hemmer-Hansen J Bekkevold D Taylor MI Ogden R Limborg MT Application of SNPs for population genetics of nonmodel organisms: new opportunities and challenges Mol Ecol Resour 2011 11 123 136 10.1111/j.1755-0998.2010.02943.x 21429169 29. De Wit P Pespeni MH Palumbi SR SNP genotyping and population genomics from expressed sequences—current advances and future possibilities Mol Ecol 2015 24 2310 2323 10.1111/mec.13165 25808983 30. van Bers NEM Santure AW van Oers K de Cauwer I Dibbits BW Mateman C The design and cross-population application of a genome-wide SNP chip for the great tit Parus major Mol Ecol Resour 2012 12 753 770 10.1111/j.1755-0998.2012.03141.x 22487530 31. Hagen IJ Billing AM Rønning B Pedersen SA Pärn H Slate J The easy road to genome-wide medium density SNP screening in a non-model species: development and application of a 10K SNP-chip for the house sparrow (Passer domesticus ) Mol Ecol Resour 2013 13 429 439 10.1111/1755-0998.12088 23480404 32. Montes I Conklin D Albaina A Creer S Carvalho GR Santos M SNP discovery in European anchovy (Engraulis encrasicolus L.) by high-throughput transcriptome and genome sequencing PLoS One 2013 8 e70051 10.1371/journal.pone.0070051 23936375 33. Hoffman JI Gene discovery in the Antarctic fur seal (Arctocephalus gazella ) skin transcriptome Mol Ecol Resour 2011 11 703 710 10.1111/j.1755-0998.2011.02999.x 21466659 34. Hoffman JI Thorne MAS Trathan PN Forcada J Transcriptome of the dead: characterisation of immune genes and marker development from necropsy samples in a free-ranging marine mammal BMC Genomics 2013 14 52 10.1186/1471-2164-14-52 23347513 35. R Core Team R: a language and environment for statistical computing 2015 Vienna R Core Team 36. Hellmann I Mang Y Gu Z Li P de la Vega FM Clark AG Population genetic analysis of shotgun assemblies of genomic sequences from multiple individuals Genome Res 2008 18 1020 1029 10.1101/gr.074187.107 18411405 37. Hoffman JI Tucker R Bridgett SJ Clark MS Forcada J Slate J Rates of assay success and genotyping error when single nucleotide polymorphism genotyping in non-model organisms: a case study in the Antarctic fur seal Mol Ecol Resour 2012 12 861 872 10.1111/j.1755-0998.2012.03158.x 22727236 38. Nielsen R Korneliussen T Albrechtsen A Li Y Wang J SNP calling, genotype calling, and sample allele frequency estimation from new-generation sequencing data PLoS One 2012 7 e37558 10.1371/journal.pone.0037558 22911679 39. Van Belleghem SM Roelofs D Van Houdt J Hendrickx F De novo transcriptome assembly and SNP discovery in the wing polymorphic salt marsh beetle Pogonus chalceus (Coleoptera, Carabidae) PLoS One 2012 7 e42605 10.1371/journal.pone.0042605 22870338 40. Zhou ZC Dong Y Sun HJ Yang AF Chen Z Gao S Transcriptome sequencing of sea cucumber (Apostichopus japonicus ) and the identification of gene-associated markers Mol Ecol Resour 2014 14 127 138 10.1111/1755-0998.12147 23855518 41. Yu Y Wei J Zhang X Liu J Liu C Li F SNP discovery in the transcriptome of white Pacific shrimp Litopenaeus vannamei by next generation sequencing PLoS One 2014 9 e87218 10.1371/journal.pone.0087218 24498047

PMC005xxxxxx/PMC5000417.txt

==== Front BMC Complement Altern MedBMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 130310.1186/s12906-016-1303-xResearch ArticleComparative evaluation of the sexual functions and NF-κB and Nrf2 pathways of some aphrodisiac herbal extracts in male rats Sahin Kazim +904242370000/3938nsahinkm@yahoo.com 12Orhan Cemal 1Akdemir Fatih 2Tuzcu Mehmet 3Gencoglu Hasan 3Sahin Nurhan 1Turk Gaffari 4Yilmaz Ismet 5Ozercan Ibrahim H. 6Juturu Vijaya 71 Department of Animal Nutrition, Faculty of Veterinary Science, Firat University, 23119 Elazig, Turkey 2 Department of Nutrition, Faculty of Fisheries, Inonu University, Malatya, 44280 Turkey 3 Division of Biology, Faculty of Science, Firat University, 23119 Elazig, Turkey 4 Department of Reproduction and Artificial Insemination, Faculty of Veterinary, Firat University, 23119 Elazig, Turkey 5 Department of Pharmacology, Faculty of Pharmacy, Inonu University, Malatya, 44100 Turkey 6 Department of Pathology, Faculty of Medicine, Firat University, 23119 Elazig, Turkey 7 Research and Development, OmniActive Health Technologies Inc., Morristown, NJ USA 26 8 2016 26 8 2016 2016 16 1 31811 5 2016 18 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Mucuna pruriens, Tribulus terrestris and Ashwagandha (Withania somnifera) are widely known as antioxidant effective herbals and have been reported to possess aphrodisiac activities in traditional usages. In this study, we determined the effects of these herbals on sexual functions, serum biochemical parameters, oxidative stress and levels of NF-κB, Nrf2, and HO-1 in reproductive tissues. Methods Thirty-five male rats were divided into five groups: the control group, sildenafil-treated group (5 mg/kg/d), Mucuna, Tribulus and Ashwagandha groups. The extract groups were treated orally either with Mucuna, Tribulus or Ashwagandha (300 mg/kg b.w.) for 8 weeks. Results All of the extracts were found to be significantly effective in sexual functioning and antioxidant capacity and Tribulus showed the highest effectiveness. Serum testosterone levels significantly increased in Tribulus and Ashwagandha groups in comparison to control group. Tribulus was able to reduce the levels of NF-κB and increase the levels of Nrf2 and HO-1 to a much greater extent than Mucuna and Ashwagandha. Conclusions These results demonstrate for the first time that Mucuna, Tribulus and Ashwagandha supplementation improves sexual function in male rats via activating Nrf2/ HO-1 pathway while inhibiting the NF-κB levels. Moreover, Tribulus terrestris extract was found to be more bioavailable from Ashwagandha extract followed by Mucuna extract. Graphical abstract Schematic representation of the mode of action of some aphrodisiac herbal extracts to improve sexual functions Keywords Mucuna pruriensTribulus terrestrisWithania somniferaSexual enhancerFertilityReproductive organsOmniactive Health Technologies (NJ, USA) Turkish Academy of Sciences (Ankara, Turkey). issue-copyright-statement© The Author(s) 2016 ==== Body Background Male infertility and reproductive dysfunctions are serious widespread health problems and almost half of the human infertility is considered to be male, moreover, the etiology is not obvious in 40 to 50 % of infertile males as well [1, 2]. It has also been observed that oligospermia is the single most common reason for the depressed male fertility [3]. Metabolic generation of reactive oxygen species (ROS) is required for male sexual function, whilst high levels of ROS may be the reason for low sperm quality and male infertility [4]. Animal experiments suggest that a better treatment for sexual dysfunction or infertility may not only improve sexual relationships but also the overall quality of life [5]. Free radicals are extremely reactive molecules which include reactive oxygen species (ROS) and also nitrogen species. These radicals are normally generated by subcellular compartments of testes, particularly mitochondria; however excessive free radical production can cause tissue injury and cell death and result as depleting the antioxidant status [6]. In many countries, different varieties of plants have been used as sexual stimulants in traditional medicine including Mucuna pruriens, Tribulus terrestris, and Ashwagandha [7–9]. Mucuna pruriens (commonly known as Cowitch), a leguminous plant grown in Africa, South America, and South Asia is identified as an herbal medicine for improving the antioxidant, antidiabetic, antidiabetes-induced erectile dysfunction, anti-inflammatory, neuroprotective, aphrodisiac, anticataleptic and antiepileptic, antibacterial, cardioprotective properties [10–13]. Tribulus terrestris (belonging to family Zygophyllaceae), known as Gokshur or Gokharu, has been used for a long time for treatment of various kinds of diseases by anti-inflammatory, antidiabetic, hypolipidemic, cardiotonic, hepatoprotective, analgesic, antispasmodic, anticancer, and antibacterial activities [14, 15]. Withania somnifera (Ashwagandha) is one of the most valuable herbs in the traditional Indian systems of medicine [16] and it possesses immunomodulatory, anticancer, antimetastatic, antistress, antioxidant, hemopoietic, rejuvenating properties and besides its positive influence on the endocrine, cardiopulmonary and central nervous system [17, 18]. Additionally, Mucuna pruriens also is shown to improve sperm density, motility, serum testosterone levels, male sexual behavior and even androgenic in hyperglycemic male rats [19, 20]. Tribulus terrestris has the potential to increase hormone levels of testosterone and enhance premature ejaculation [21, 22]. Ashwagandha has the capacity to improve blood circulation in the body, thus naturally enhance sperm quality [8, 23]. Diminishing the excessive ROS levels is suggested for being a penetrating way to overcome infertility problems during the aging [24] and even enhancing the testosterone alone were shown to reduce ROS production and increase p-eNOS/eNOS ratio in the castrated rats, therefore, ameliorate erectile dysfunction as well [25]. The signaling pathway of erythroid 2-related factor 2 (Nrf2) antioxidant response element (ARE) plays a key role in oxidative stress response and Nrf2 was suggested to be a transcription factor that promotes the expression of many crucial antioxidant genes [26, 27]. Heme oxygenase (HO) pathway has a major role in male reproductive system and sexual dysfunctions and heat shock protein 32 or heme oxygenase-1 protein (HO-1) is known to be an inducible isoform protein and it is induced by many different conditions and agents including hypoxia, cytokines, oxidative stress, heat shock, and reactive oxygen species [28, 29]. The nuclear factor kappa-B (NF-κB), is an enhancer-binding transcription factor located in the immune response and participates to cell proliferation and apoptosis, meanwhile, NF-κB binds to DNA at 50 kDa subunit (p50) and involves the transcriptional activation with the 65 kDa subunit (p65) [30]. The molecular mechanism including NF-κB and Nrf2 pathway of Tribulus terrestris, Mucuna pruriens and Ashwagandha extracts as an antioxidant in reproductive tissues of male rats has not been investigated. Based on the current knowledge, the aim of the present study was to determine the effects of the Tribulus terrestris, Mucuna pruriens and Ashwagandha on levels of serum biochemical parameters, hormones, MDA (malondialdehyde) and the expression of Nrf2 and NF-κB in reproductive tissues of male rats. Methods Animals Thirty-five Sprague-Dawley male rats (age: 8 weeks, weight: 180 ± 20 g) were provided by the Laboratory of Experimental Animals of Inonu University, Malatya, Turkey and the Animal Care Committee of the Inonu University approved this study (2014/A-53). The rats were housed in standard plastic cages (10.25″w × 18.75″1 × 4″d), in a controlled environment with humidity (55 ± 5 %) under a 12:12-h light-dark cycle at 22 °C. Rats were provided with standard diet and tap water ad libitum. The animals were housed in plastic wire meshed cages in the animal house. The wooden material used as bedding was replaced every 3 days. All the experiments were conducted between 09.30 and 17.00 h to minimize the effects of environmental changes. Experimental design The rats were housed in plastic metabolic cages and were randomly divided into 5 groups of seven. These groups were as follows: (i) Group I – Negative control: Rats only fed with standard diet and tap water as vehicle, (ii) Group II – Positive control: Rats were fed with standard diet and treated with sildenafil citrate (Viagra, Pfizer) orally via an intragastric tube (5 mg/kg/d), (iii) Group III: Rats were fed with standard diet and treated with Mucuna pruriens (iv) Group IV: Rats were fed with standard diet treated with Tribulus terrestris Group V: Rats were fed with standard diet and treated with Ashwagandha. The animals were treated with dried seed powder of Mucuna pruriens extracted with an ethanol-water mixture (70:30) (Item code: 51002; active ingredient: min.20 % L-DOPA), dried fruit extract powder of Tribulus terrestris extracted with a methanol-water mixture (70:30) (Item code: 35002; active ingredient: min. 40 % saponis) and dried roots of the plant Withania somnifera extracted with a methanol-water mixture (70:30) (Item code: 13002; active ingredients: min: 2.5 % total withanoloids). The purity of all the extracts was 98 %. All the extracts were dissolved in distilled water and administered orally via an intragastric tube daily with an optimal dose of 300 mg/kg body weight for 8 weeks in parallel with sildenafil citrate treated positive control group which was also administered in the same way like the other extract treated groups based on preliminary and published data [31]. All of the herbal extracts were supplied by Omni Active Health Technologies Pvt. Ltd. (Mumbai, India). Animals were sacrificed under the ether anesthesia and the testes, epididymis, vas deferens, and ventral prostate were removed and cleared from adhering connective tissue and weighed. One of the testis samples was fixed in Bouin’s solution for histopathological examination. The other testes samples were stored at 20 ° C for biochemical analyses. Testes were taken from 20 °C freezer and immediately transferred to the cold glass tubes. Then, the testes were diluted with a nine-fold volume of phosphate-buffered saline (PBS; pH 7.4). For the enzymatic analyses, testes were minced in a glass and homogenized by a Teflon–glass homogenizer for 3 min in cold physiological saline on ice [32]. The cauda epididymidis was cut longitudinally with a pair of fine-pointed scissors and compressed with forceps. The sperm was released by mincing the cauda epididymis into pieces on the Petri dishes that contained phosphate buffer saline (PBS) for sperm characteristics analysis. Since epididymis came in pairs, one cauda epididymis was put in a Petri dish containing 10 ml of 0.1 M PBS specifically for sperm count and sperm motility analysis while the other cauda epididymis put in another Petri dish containing 1 ml 0.1 M PBS for sperm viability and sperm morphology. The spermatozoa were allowed to flow out from cauda epididymis into the buffer. Then, the sperm suspensions were left at room temperature for 10 min for the suspension to allow sperm to swim out of the lumen of the cauda epididymidis for sperm characteristics analysis. Sperm quality Sperm analyses were performed using the methods previously reported in the study by Turk et al. [33]. Sperm count was determined using the hemocytometer under a light microscope. A cover slip was placed on the hemocytometer before a drop with 10 μl of caudal epididymal sperm solution loaded under the cover slip. Haemocytometer was again used for sperm motility analysis. A cover slip was placed on the hemocytometer before a drop with 10 μL of caudal epididymal sperm solution was loaded under the cover slip. Sperm viability analysis will be used the sperm from the other cauda epididymis that will be put in a Petri dish with 1 ml 0.1 M PBS. On a clean glass slide, 1 drop of sperm suspension will be gently mixed with 3 drops of eosin using the sharp glass slide end. After 30 s, 1 drop of nigrosin was mixed together with the solution and a smear will be made. The smear will be then air-dried and observed under x200 magnification of imaging microscope. The sperm will be counted based on the degree of membrane permeability. The dead sperm will be showed pink coloration of the head whereas the viable sperm will be showed whitish or colorless head. Sperm morphology analysis will be used the same sperm smear made for sperm viability analysis. This time, the sperm will be observed under x400 magnification of imaging microscope to clearly evaluate the morphology of the sperm head, neck, and tail. The sperm will be generally classified as normal or abnormal without further characterized the types of abnormality found on the sperm. The normal sperm will be given a score of 100 and the abnormal one will be given a score of “0” to enable statistical analysis by using Statistical Analysis System (SAS) to be carried out easily. Laboratory analyses Hematology parameters in whole blood samples were analyzed by automated analyzer Exigo EOS Vet (Boule Medical AB, Spanga, Sweden). Serum aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (ALP), creatine kinase (CK) concentrations were analyzed by the biochemical analyzer (Samsung LABGEO-PT10). Serum testosterone, luteinizing hormone (LH), follicle stimulating hormone (FSH) levels were measured by using an ELISA kit (Cayman Chemical Company, Ann Arbor, Michigan, USA; Elx-800; Bio-Tek Instruments Inc, Vermont, USA). A 10 % (w/v) tissue homogenate was prepared in 10 mM phosphate buffer (pH 7.4). The homogenate was centrifuged at 13,000 g for 10 min at 4 °C. The supernatant was collected and stored at -80 °C. The concentration of MDA, an index of lipid peroxidation and oxidative stress, was measured using the fully automatic HPLC (Shimadzu, Kyoto, Japan) equipped with a pump (LC-20 AD), an ultraviolet-visible detector (SPD-20A), an inertsil ODS-3 C18 column (250 × 4.6 mm, 5 m), a column oven (CTO-10ASVP), an autosampler (SIL-20A), a degasser unit (DGU-20A5) and a computer system with LC solution Software (Shimadzu) [34]. Western blot analysis for NF-κB p65, Nrf-2 and HO-1 Protein (NF-κB, Nrf2, and HO-1) levels were analyzed by western blotting technique. For western blotting; epididymis, prostate, testes and vas deferens tissues of the rats were removed after sacrification to analyze the target protein expressions among the groups. Briefly, accurately weighed each tissue sample was homogenized in 1:10 (w/v) in 10 mM Tris-HCl buffer at pH 7.4, containing 0.1 mM NaCl, 0.1 mM phenylmethylsulfonyl fluoride, and 5 μM soybean (soluble powder; Sigma, St. Louis, MO, USA) as trypsin inhibitor. Tissue homogenate was centrifuged at 15,000 g at 4 °C for 30 min, and the supernatant was transferred into fresh tubes. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis sample buffer containing 2 % β-mercaptoethanol was added to the supernatant. Equal amounts of protein (20 μg) were electrophoresed and subsequently transferred to nitrocellulose membrane (Schleicher and Schuell Inc., Keene, NH, USA). Nitrocellulose blots were washed twice for 5 min in phosphate buffered saline (PBS) and blocked with 1 % bovine serum albumin in PBS for 1 h prior to the application of primary antibody. Rat antibodies against NF-κB 65, Nrf-2 and HO-1 were purchased from Abcam (Cambridge, UK). Primary antibody was diluted (1:1000) in the same buffer containing 0.05 % Tween-20. The nitrocellulose membrane was incubated overnight at 4 °C with protein antibody. The blots were washed and incubated with horseradish peroxidase-conjugated goat anti-mouse IgG (Abcam, Cambridge, UK). Specific binding was detected using diaminobenzidine and hydrogen peroxide as substrates. Protein loading was controlled using a monoclonal mouse antibody against β-actin antibody (A5316; Sigma). Band intensities of the proteins were quantified by densitometric analysis using an image analysis system (Image J; National Institute of Health, Bethesda, USA). Samples were analyzed in quadruplicate, and a representative blot is shown in the respective figures. Results were normalized to the β-actin expression in each group as percent of control. Statistical analysis Data analysis was performed using Statistical Analysis System (SAS) version 9.2. Data of body weight, serum testosterone, mounting latency and mounting frequency were subjected to analysis of variance (ANOVA) to analyze the significant treatment effect and the mean between the groups was compared using Duncan Multiple Range Test if F value was significant at P < 0.05. Results No significant change of the extracts was observed on final body weight, absolute and relative reproductive organ weights of the animals among the groups (P > 0.05) (Table 1). Sexual behavior changes were presented in Table 2. All the treatment groups showed significant decreases in mounting latency and intromission latency values (P < 0.0001), and significant increases in mounting frequency and intromission frequency values when compared to the standard control group (P < 0.0001).Table 1 The effects of extracts on final body weight, absolute and relative reproductive organ weights Item Groups --P-- Control Sildenafil Mucuna pruriens Tribulus terrestris Ashwagandha Final body weight, g 324.00 ± 24.84 316.57 ± 17.99 312.00 ± 29.73 307.71 ± 52.45 331.80 ± 28.31 >0.05 Testis, g 1.39 ± 0.07 1.35 ± 0.11 1.35 ± 0.09 1.36 ± 0.14 1.33 ± 0.08 >0.05 Whole epididymis, g 0.58 ± 0.05 0.61 ± 0.04 0.56 ± 0.05 0.58 ± 0.07 0.59 ± 0.04 >0.05 Right cauda epididymis, g 0.25 ± 0.03 0.24 ± 0.03 0.24 ± 0.03 0.23 ± 0.05 0.23 ± 0.02 >0.05 Vas deferens, g 0.12 ± 0.01 0.13 ± 0.01 0.12 ± 0.01 0.13 ± 0.01 0.13 ± 0.01 >0.05 Seminal vesicles, g 1.16 ± 0.30 1.10 ± 0.25 1.09 ± 0.18 1.17 ± 0.24 1.39 ± 0.31 >0.05 Ventral prostate, g 0.43 ± 0.07 0.49 ± 0.07 0.41 ± 0.07 0.50 ± 0.17 0.48 ± 0.18 >0.05 Testis*, % 0.43 ± 0.04 0.43 ± 0.04 0.43 ± 0.05 0.45 ± 0.05 0.40 ± 0.03 >0.05 Whole epididymis*, % 0.18 ± 0.02 0.19 ± 0.02 0.18 ± 0.02 0.19 ± 0.02 0.18 ± 0.01 >0.05 Right cauda epididymis*, % 0.08 ± 0.01 0.08 ± 0.01 0.08 ± 0.01 0.08 ± 0.01 0.07 ± 0.004 >0.05 Vas deferens*, % 0.04 ± 0.005 0.04 ± 0.008 0.04 ± 0.003 0.04 ± 0.007 0.04 ± 0.004 >0.05 Seminal vesicles*, % 0.36 ± 0.09 0.35 ± 0.09 0.35 ± 0.05 0.38 ± 0.06 0.42 ± 0.9 >0.05 Ventral prostate*, % 0.13 ± 0.02 0.15 ± 0.02 0.13 ± 0.02 0.16 ± 0.04 0.14 ± 0.05 >0.05 *Relative reproductive organ weights [organ weight (g) / final body weight (g) X 100] Control, no treatment; Sildenafil, rats treated with Sildenafil (5 mg/kg/d); Mucuna, rats treated with Mucuna pruriens (300 mg/kg bw); Tribulus; rats treated with Tribulus terrestris (300 mg/kg bw); Ashwagandha; rats treated with Ashwagandha (300 mg/kg bw). Data are LS means ± SE (n = 7). Different superscripts in the same row (a–c) indicate group mean differences (p < 0.05) Table 2 The effects of extracts on sexual behaviors in rats Item Groups --P-- Control Sildenafil Mucuna pruriens Tribulus terrestris Ashwagandha Mounting Latency, sec 9.48 ± 1.20a 2.10 ± 0.34d 5.27 ± 0.62b 2.86 ± 0.62cd 3.15 ± 0.60c 0.0001 Mounting Frequency* 65.65 ± 8.19d 214.27 ± 17.76a 118.99 ± 7.51c 170.86 ± 10.42b 167.72 ± 14.91b 0.0001 Intromission Latency, sec 9.40 ± 1.07a 0.98 ± 0.22d 4.87 ± 0.50b 2.13 ± 0.15c 2.73 ± 0.40c 0.0001 Intromission Frequency* 59.51 ± 4.70d 203.90 ± 15.78a 132.03 ± 11.32c 173.71 ± 14.02b 169.70 ± 18.41b 0.0001 Control, no treatment; Sildenafil, rats treated with Sildenafil (5 mg/kg/d); Mucuna, rats treated with Mucuna pruriens (300 mg/kg bw); Tribulus; rats treated with Tribulus terrestris (300 mg/kg bw); Ashwagandha; rats treated with Ashwagandha (300 mg/kg bw). Data are LS means ± SE (n = 7). Different superscripts in the same row (a–c) indicate group mean differences (p < 0.05). * The number of intromissions in an ejaculatory series The effects of extracts on sperm motility, sperm count, and abnormal sperm rate were shown in Table 3. Sperm motility was significantly different between the treatment groups (P < 0.05). Rats that were supplemented with Mucuna and Tribulus showed higher mean value for sperm motility compared with the control groups as shown in Table 3. However, Tribulus group showed the highest total sperm motility percentage than the other groups with a considerably high mean level of 84.29 % (P < 0.05). Mucuna, Tribulus, and Ashwagandha supplementation also produced a significant increase in sperm counts compared to the control group (P < 0.05). Tribulus group indicated the highest number of sperm count with a mean level of 161.42 million/right cauda epididymis (P < 0.05). However, there was no statistically significant difference in head, tail and total abnormal sperm rates (P > 0.05).Table 3 The effects of extracts on sperm characteristics and abnormal sperm rate (%) in rats Item Groups -- P -- Control Sildenafil Mucuna pruriens Tribulus terrestris Ashwagandha Total Motility,% 75.00 ± 5.49bc 80.00 ± 4.16b 80.00 ± 4.16b 84.29 ± 5.34a 78.00 ± 5.37b <0.05 Count * 110.33 ± 37.78b 146.29 ± 21.55a 148.57 ± 31.45a 161.42 ± 36.11a 154.80 ± 10.55a <0.05 Abnormal sperm rate, %  Head 5.00 ± 3.85 3.71 ± 2.56 4.43 ± 2.76 3.00 ± 1.53 5.40 ± 2.88 >0.05  Tail 5.17 ± 2.14 4.86 ± 1.35 5.00 ± 2.45 3.86 ± 1.68 7.40 ± 3.21 >0.05  Total 10.17 ± 4.71 8.57 ± 2.37 9.43 ± 3.82 6.86 ± 2.11 12.80 ± 5.31 >0.05 Control, no treatment; Sildenafil, rats treated with Sildenafil (5 mg/kg/d); Mucuna, rats treated with Mucuna pruriens (300 mg/kg bw); Tribulus; rats treated with Tribulus terrestris (300 mg/kg bw); Ashwagandha; rats treated with Ashwagandha (300 mg/kg bw). Data are LS means ± SE (n = 7). Different superscripts in the same row (a–c) indicate group mean differences (p < 0.05). *Million/right cauda epididymis The hematological data of the groups was presented in Table 4. No statistical significance was detected among all blood parameters of the Ashwagandha, Tribulus and Mucuna plant extract groups when compared to positive control sildenafil group and negative control standard group (P > 0.05). The effects of the extracts on serum biochemical parameters were shown in Table 5. There was no significant difference among the AST, ALT, CK, urea and creatine levels between the groups (P > 0.05), despite serum ALP level of Mucuna group was lower than the other groups (P < 0.05).Table 4 The effects of extracts on blood parameters in rats Item Groups --P-- Control Sildenafil Mucuna pruriens Tribulus terrestris Ashwagandha L WBC 8.59 ± 0.31 8.73 ± 0.34 8.66 ± 0.45 8.98 ± 0.68 8.59 ± 0.22 >0.05 LYM, % 62.03 ± 7.18 62.53 ± 5.90 60.98 ± 12.69 61.19 ± 9.99 64.58 ± 4.47 >0.05 MID, % 14.07 ± 4.37 12.31 ± 3.85 12.20 ± 2.63 16.30 ± 6.67 11.82 ± 4.39 >0.05 GRAN, % 23.90 ± 5.26 25.16 ± 6.01 26.81 ± 13.67 22.51 ± 3.99 23.60 ± 3.10 >0.05 MID, # 10*3/μl 6.62 ± 0.53 6.60 ± 0.52 6.53 ± 0.39 7.11 ± 0.94 6.35 ± 0.65 >0.05 GRAN, # 10*3/μl 7.13 ± 0.42 7.32 ± 0.39 7.26 ± 0.79 7.46 ± 0.87 7.16 ± 0.23 >0.05 RBC, 10*6/μl 15.96 ± 0.05 15.96 ± 0.5 16.00 ± 0.06 15.96 ± 0.07 15.99 ± 0.04 >0.05 HGB, g/dL 16.27 ± 0.96 16.14 ± 0.65 16.57 ± 0.81 16.14 ± 1.06 16.60 ± 0.67 >0.05 HCT, % 81.27 ± 5.63 79.93 ± 3.51 81.63 ± 3.62 80.64 ± 5.79 81.16 ± 2.78 >0.05 MCV, fL 94.75 ± 2.56 94.13 ± 5.61 91.94 ± 4.42 94.69 ± 2.81 91.90 ± 1.76 >0.05 MCH, pg 18.90 ± 0.37 18.94 ± 0.88 18.29 ± 0.56 18.89 ± 0.38 18.72 ± 0.50 >0.05 MCHC, d/dL 19.97 ± 0.60 20.14 ± 0.40 20.24 ± 0.51 19.56 ± 1.38 20.42 ± 0.33 >0.05 RDW-CV, % 17.32 ± 0.58 16.47 ± 1.36 17.24 ± 1.19 17.90 ± 1.09 17.76 ± 0.83 >0.05 PLT, 10*3/μl 13.79 ± 0.11 13.60 ± 0.17 13.82 ± 0.20 13.76 ± 0.19 13.85 ± 0.20 >0.05 MPV, fL 11.62 ± 0.35 11.36 ± 0.46 11.86 ± 0.16 11.44 ± 0.40 11.68 ± 0.22 >0.05 PDW, % 17.53 ± 1.31 16.99 ± 1.09 16.74 ± 1.36 16.83 ± 0.94 16.20 ± 0.91 >0.05 PCT, % 1.14 ± 0.13 0.92 ± 0.15 1.22 ± 0.28 1.09 ± 0.24 1.22 ± 0.21 >0.05 LWBC leukocyte white blood cells, LYM lymphocyte, MID minimum inhibitory dilution, GRAN granulocyte, RBC red blood cell, HGB hemoglobin, HCT hematocrit, MCV mean corpuscular volume, MCH mean corpuscular hemoglobin, MCHC mean corpuscular hemoglobin concentration, RDW-CV red cell distribution width- coefficient variation, PLT platelet, MPV mean platelet volume, PDW platelet distribution width, PCT platelet crit; Control, no treatment; Sildenafil, rats treated with Sildenafil (5 mg/kg/d); Mucuna, rats treated with Mucuna pruriens (300 mg/kg bw); Tribulus; rats treated with Tribulus terrestris (300 mg/kg bw); Ashwagandha; rats treated with Ashwagandha (300 mg/kg bw). Data are LS means ± SE (n = 7). Different superscripts in the same row (a–c) indicate group mean differences (p < 0.05) Table 5 The effects of extracts on serum biochemical parameters in rats Item Groups --P-- Control Sildenafil Mucuna pruriens Tribulus terrestris Ashwagandha AST, U/L 97.50 ± 8.92 118.86 ± 33.73 99.29 ± 15.85 104.29 ± 21.86 108.51 ± 23.25 >0.05 ALT, U/L 46.67 ± 6.50 59.86 ± 10.48 48.71 ± 6.78 47.71 ± 10.24 50.00 ± 10.05 >0.05 ALP, U/L 204.00 ± 43.32a 202.14 ± 31.02a 161.86 ± 29.29b 215.86 ± 25.58a 202.00 ± 22.03a 0.0349 CK, U/L 531.00 ± 224.85 399.29 ± 292.28 405.86 ± 189.07 226.86 ± 102.12 331.40 ± 44.57 >0.05 Urea, mg/dl 33.00 ± 2.61 36.43 ± 2.07 36.14 ± 5.08 36.00 ± .3.74 36.80 ± 6.18 >0.05 Creatine, mg/dl 0.30 ± 0.03 0.28 ± 0.03 0.28 ± 0.04 0.29 ± 0.02 0.30 ± 0.03 >0.05 AST aspartate aminotransferase, ALT alanine aminotransferase, ALP alkaline phosphatase, CK creatine kinase control, no treatment; Sildenafil, rats treated with Sildenafil (5 mg/kg/d); Mucuna, rats treated with Mucuna pruriens (300 mg/kg bw); Tribulus; rats treated with Tribulus terrestris (300 mg/kg bw); Ashwagandha; rats treated with Ashwagandha (300 mg/kg bw). Data are LS means ± SE (n = 7). Different superscripts in the same row (a–c) indicate group mean differences (p < 0.05) When we examined the effects of extracts on serum hormones and MDA (Table 6), FSH and LH hormone levels showed no significance (P > 0.05) but serum testosterone levels were 1.84, 1.76 and 1.58 fold higher than the normal control group in Sildenafil, Tribulus and Ashwagandha groups respectively (P < 0.0001). Mucuna pruriens administration did not show significance on serum testosterone levels when compared to standard control group of rats (P > 0.05). All of the extract treatments to the groups lowered the serum and testis tissue MDA levels significantly, in comparison to sildenafil group of rats (P < 0.0001). The most prominent ameliorating decrease was in Tribulus group with a 32 % in serum and 14 % in testis MDA levels (P < 0.0001).Table 6 The effects of extracts on serum hormones and MDA levels in rats Item Groups --P-- Control Sildenafil Mucuna pruriens Tribulus terrestris Ashwagandha FSH, mIU/ml 0.33 ± 0.22 0.40 ± 0.37 0.31 ± 0.25 0.34 ± 0.39 0.31 ± 0.24 >0.05 LH, mIU/ml 0.22 ± 0.29 0.30 ± 0.23 0.25 ± 0.17 0.24 ± 0.19 0.27 ± 0.19 >0.05 Testosterone, ng/ml 2.27 ± 0.17c 4.14 ± 0.71a 2.57 ± 0.44c 3.99 ± 0.16ab 3.58 ± 0.18b 0.0001 Serum MDA, μmol/L 0.54 ± 0.05a 0.56 ± 0.06a 0.49 ± 0.04b 0.38 ± 0.02c 0.44 ± 0.02b 0.0001 Testis MDA, nmol/g 1.64 ± 0.09ab 1.68 ± 0.09a 1.55 ± 0.06cb 1.45 ± 0.09d 1.51 ± 0.07cd 0.0001 FSH follicle-stimulating hormone, LH luteinizing hormone, MDA malonaldeyhde; Control, no treatment; Sildenafil, rats treated with Sildenafil (5 mg/kg/d); Mucuna, rats treated with Mucuna pruriens (300 mg/kg bw); Tribulus; rats treated with Tribulus terrestris (300 mg/kg bw); Ashwagandha; rats treated with Ashwagandha (300 mg/kg bw). Data are LS means ± SE (n = 7). Different superscripts in the same row (a–c) indicate group mean differences (p < 0.05) Figure 1(a–d), shows the western blot bands of the Control, Sildenafil, Mucuna, Tribulus and Ashwagandha plant extract groups on the expression levels of NF-κB and HO-1/Nrf2 in reproductive tissues which included epididymis, prostate, testes and vas deferens. In the epididymis tissue (Fig. 2a–c), higher levels of HO-1/Nrf2 were found in Tribulus, Mucuna and Ashwagandha groups compared to the other groups, in contrast, NF-κB levels were significantly lower in Tribulus, Mucuna and Ashwagandha administered groups (P < 0.0001). In the prostate tissue, HO-1/Nrf2 levels of Tribulus, Mucuna, and Ashwagandha extract treated groups significantly increased compared to the negative and positive controls while NF-κB levels decreased considerably in the same groups compared to the both controls (Fig. 3a–c), (P < 0.0001). Tribulus and Mucuna extract treated groups showed higher HO-1 levels in testis tissue than the other groups and also Tribulus, Mucuna and Ashwagandha groups Nrf2 levels were significantly higher than Sildenafil and Standard control (Fig. 4a,c), (P < 0.0001). Furthermore, the least NF-κB expression level was found in Tribulus group in the testes and also Mucuna and Ashwagandha groups suggested a significant decrease in comparison to positive control Sildenafil group (Fig. 4b) (P < 0.0001). In vas deferens tissue HO-1 expression, Tribulus extract treated rats showed the highest levels among the groups (Fig. 5a) (P < 0.0001), while significantly higher Nrf2 levels, were again found in Tribulus and moreover Ashwagandha groups (Fig. 5c) (P < 0.0001). All herbal extract treated groups showed significantly lower NF-κB levels when compared to both positive and negative controls (Fig. 5b) (P < 0.0001).Fig. 1 Epididymis tissue NF-κB, Nrf-2 and HO-1 levels (a); prostate tissue NF-κB, Nrf-2 and HO-1 levels (b); testes tissue NF-κB, Nrf-2 and HO-1 levels (c); vas deferens tissue NF-κB, Nrf-2 and HO-1 levels (d) western blot bands. Data are expressed as a ratio of normal control value (set to 100 %). Blots were repeated at least 4 times (n = 4) and a representative blot is shown. Actin was included to ensure equal protein loading. The bars represent the standard error of the mean. Data points with different superscripts are significantly different at the level of P < 0.05 by Fisher’s multiple comparison test Fig. 2 HO-1 (a), NF-κB (b) and Nrf-2 (c) Protein levels of epididymis tissue in male rats (P < 0.0001). Data are expressed as a ratio of normal control value (set to 100 %). Blots were repeated at least 4 times (n = 4) and a representative blot is shown. Actin was included to ensure equal protein loading. The bars represent the standard error of the mean. Data points with different superscripts are significantly different at the level of P < 0.05 by Fisher’s multiple comparison test Fig. 3 HO-1 (a), NF-κB (b) and Nrf-2 (c) Protein levels of prostate tissue in male rats (P < 0.0001). Data are expressed as a ratio of normal control value (set to 100 %). Blots were repeated at least 4 times (n = 4) and a representative blot is shown. Actin was included to ensure equal protein loading. The bars represent the standard error of the mean. Data points with different superscripts are significantly Fig. 4 HO-1 (a), NF-κB (b) and Nrf-2 (c) levels of testes tissue in male rats (P < 0.0001). Data are expressed as a ratio of normal control value (set to 100 %). Blots were repeated at least 4 times (n = 4) and a representative blot is shown. Actin was included to ensure equal protein loading. The bars represent the standard error of the mean. Data points with different superscripts are significantly different at the level of P < 0.05 by Fisher’s multiple comparison test Fig. 5 HO-1 (a), NF-κB (b) and Nrf-2 (c) Protein levels of vas deferens tissue in male rats (P < 0.0001). Data are expressed as a ratio of normal control value (set to 100 %). Blots were repeated at least 4 times (n = 4) and a representative blot is shown. Actin was included to ensure equal protein loading. The bars represent the standard error of the mean. Data points with different superscripts are significantly different at the level of P < 0.05 by Fisher’s multiple comparison test No significant change was observed neither in testes histopathology nor the sperm morphologies between the groups as shown in (Fig. 6a–e).Fig. 6 Representative photomicrographs of histopathological structure of testes and below sperm morphologies for each in different treatment groups. a: Control, no treatment; b: Sildenafil, rats treated with sildenafil (5 mg/kg/d); c: Mucuna, rats treated with Mucuna pruriens (300 mg/kg bw); d: Tribulus; rats treated with Tribulus terrestris (300 mg/kg bw); e: Ashwagandha; rats treated with Ashwagandha (300 mg/kg bw) Discussion Male infertility, erectile dysfunction, and reproductive system problems are common public health disorders besides a stressed out way of life has been increasing the sexual dysfunction suffering subjects around the world [2, 35]. A better understanding of erectile and sexual functions at molecular levels in the male reproductive system shall be a great achievement for the aim of precise aphrodisiac substance choice [35]. In this study, we tried to figure out the sexual enhancing capacity, blood and serum biochemical parameters and hormones, antioxidant capacity by assessing the serum and testis MDA and Nrf2 pathway of reproductive organ parts of the male rats by feeding the animals with Tribulus terrestris, Mucuna pruriens and Ashwagandha extracts and comparing them with standard control group and sildenafil-treated group as a positive control. Mucuna pruriens is a traditional Ayurvedic Indian medicinal plant and has been used in Indian medicine for a long time. The total alkaloids from Mucuna seeds were reported to increase sperm production and weight of some reproductive organ parts in the albino rats [36]. In our study, we also found out an increase of the sperm production by the effect of the Mucuna and the other extracts but there was no significant change observed in any reproductive tissue weights according to our results. In many studies, Mucuna pruriens seed powder were observed to significantly and sustainably improve sexual behaviors such as; increased mounting frequency and intromission frequency and decreased mounting latency and intromission latency as parallel to our results [37–39]. However, Mucuna pruriens hydrolysates were shown to be hypocholesterolemic and hypolipidemic effective and rich in protein content [40]. Diminishing the ROS level, MMP renewal, and apoptosis regulation through Mucuna pruriens improved the spermatogenesis mainly via its major component L-DOPA was shown in previous studies [41, 42], similarly as our study showed decreasing MDA levels, NF-κB protein expression, and increasing HO-1/NrF-2 levels correspondingly. Mucuna administration increased testosterone and LH levels and decreased lipid peroxidation and FSH in infertile men also stimulated the antioxidant enzymes and hormones such as catalase, SOD and GSH via reactivating the antioxidant defense system and recovered sperm count and motility [43–45]. In parallel to previous results, Mucuna pruriens treatment improved testosterone levels and significantly recovered sperm count and motility also lowered lipid peroxidation significantly but did not alter the FSH and LH levels in our study. Tribulus terrestris is regarded to be an aphrodisiac herb and has been used in traditional Far East medicine for ages because of its reputation to improve sexual functions along with its beneficial effects on various diseases [21, 22, 35, 46]. It was shown that Tribulus terrestris administration improved LH also sperm production and testosterone levels in rams [47] as we consistently showed its enhancing effect on sperm counts and testosterone levels except LH levels were determined indifferent than the control. According to our study, Tribulus group showed significant improvements in sexual behaviors, including; increased mounting and intromission frequencies and decreased mounting and intromission latencies similar to previously reported studies [7, 21, 48]. Tribulus terrestris were shown to increase the nitric oxide release and this is considered for its aphrodisiac capacity character [48, 49]. Our results suggested major increases in Nrf-2 and HO-1 levels and decreases in NF-κB and MDA levels of the various reproductive tissue parts and serum when administered with Tribulus terrestris similarly to a study which the extract was restored antioxidant enzyme activity and their expression profile in kidney tissue [50] and another study that extract blocked proliferation and induced apoptosis in cancer cells through the inhibition of NF-κB signaling [51]. Ashwagandha (Withania somnifera) has been known as Indian Ginseng and regarded as adaptogen, tonic with its aphrodisiac properties [35]. Even though an old study reported that it was shown as antifertility effective and mating behavior reducer in mice [52], recent studies reported its capability of combating stress-induced infertility and its protective effect to some reproductive endocrine dysfunctions in male rats were shown [53, 54]. Our study has been suggested Ashwagandha be effective as a sexual enhancer not as satisfactory as Tribulus terrestris but reasonably adequate. Conclusions The present study provides evidence that the extracts of Tribulus, Ashwagandha and Mucuna are potent enhancers of sexual function and behavior by the increasing testosterone levels and regulation of NF-κB and Nrf2/HO–1 pathways in male rats. The results of the present study have also indicated that Tribulus extract was comparatively more potent than the corresponding Ashwagandha and Mucuna extracts for the sexual functions. Moreover, further studies should be carried out to check the molecular markers related to the sexual function in male rats. Abbreviations ALPAlkaline phosphatase ALTAlanine aminotransferase AREAntioxidant response element ASTAspartate aminotransferase CKCreatine kinase FSHFollicle-stimulating hormone GRANGranulocyte HCTHematocrit HGBHemoglobin HO-1Heme oxygenase-1 LHLuteinizing hormone LWBCLeukocyte white blood cells LYMLymphocyte MCHMean corpuscular hemoglobin MCHCMean corpuscular hemoglobin concentration MCVMean corpuscular volume MDAMalondialdehyde MIDMinimum inhibitory dilution MPVMean platelet volume NF-κBNuclear factor kappa B NRF2Erythroid 2-related factor 2 PCTPlatelet crit PDWPlatelet distribution width PLTPlatelet RBCRed blood cell RDW-CVRed cell distribution width-coefficient variation Acknowledgements We would like to acknowledge to the Omniactive Health Technologies (NJ, USA) and Turkish Academy of Sciences (Ankara, Turkey) for supporting this study. Funding The study was financially supported by the Omniactive Health Technologies (NJ, USA) and Turkish Academy of Sciences (Ankara, Turkey). Availability of data and materials The data sets supporting the conclusions of this article are presented in this main paper. Animals were obtained from the animal house of the Laboratory of Experimental Animals of Inonu University, Malatya, Turkey. Authors’ contributions CO, FA, GT, HG, NS, IHO, IY, and MT conducted research, data collection and analyses. KS and VJ planned and designed the research, wrote the manuscript, and have responsibility for the final content. All authors read and gave final approval of the version to be published. Competing interests The KS, CO, FA, MT, HG, NS, GT, IY and IHO declare that there is no competing interest. The study was funded by Omniactive Health Technologies (NJ, USA). Vijaya Juturu is an employee of Omniactive Health Technologies (NJ, USA), the distributors of extracts used in this study. Consent for publication Not applicable. Ethics approval and consent to participate All procedures on animals in this study complied with the standards for the care and use of experimental animals and were approved by the Animal Ethics Committee of Inonu University, Malatya, Turkey (2014/A-53). ==== Refs References 1. Collins JA Mishell DR Lobo RA Paulsen CA Male infertility: the interpretation of the diagnostic assessment The Year Book of Infertility vol. 45 1989 Chicago Year Book Medical Publishers 271 2. Cooper TG Noonan E von Eckardstein S Auger J Baker HW Behre HM World health organization reference values for human semen characteristics Hum Reprod Update 2010 16 231 45 10.1093/humupd/dmp048 19934213 3. Biswas TK Pandit S Mondal S Biswas SK Jana U Ghosh T Clinical evaluation of spermatogenic activity of processed shilajit in oligospermia Andrologia 2010 42 48 56 10.1111/j.1439-0272.2009.00956.x 20078516 4. Agarwal A Virk G Ong C du-Plessis SS Effect of oxidative stress on male reproduction World J Mens Health 2014 32 1 17 10.5534/wjmh.2014.32.1.1 24872947 5. Shin BC Lee MS Yang EJ Lim HS Ernst E Maca (L. meyenii) for improving sexual function: a systematic review BMC Complement Altern Med 2010 10 44 6 10.1186/1472-6882-10-44 20691074 6. Dare BJ Oyeniyi F Olaniyan OT Role of antioxidant in testicular integrity Annu Res Rev Biol 2014 4 998 1023 10.9734/ARRB/2014/4453 7. Singh S Nair V Gupta YK Evaluation of the aphrodisiac activity of tribulus terrestris linn. in sexually sluggish male albino rats J Pharmacol Pharmacother 2012 3 43 7 10.4103/0976-500X.92512 22368416 8. Ambiye VR Langade D Dongre S Aptikar P Kulkarni M Dongre A Clinical evaluation of the spermatogenic activity of the root extract of ashwagandha (withania somnifera) in oligospermic males: A Pilot Study Evid Based Complement Alternat Med. 2013 201 571420 24371462 9. Suresh S, Prithiviraj E, Prakash S. Dose- and time-dependent effects of ethanolic extract of Mucuna pruriens Linn. seed on sexual behaviour of normal male rats. J Ethnopharmacol. 2009;122:497–501. 10. Misra L Wagner H Alkaloidal constituents of Mucuna pruriens seeds Phytochemistry 2004 65 2565 7 10.1016/j.phytochem.2004.08.045 15451318 11. Majekodunmi SO Oyagbemi AA Umukoro S Odeku OA Evaluation of the anti-diabetic properties of Mucuna pruriens seed extract Asian Pac J Trop Med 2011 4 632 6 10.1016/S1995-7645(11)60161-2 21914541 12. Uma S Gurumoorthi P Dietary antioxidant activities in different germplasms of Mucuna J Med Food 2013 16 618 24 10.1089/jmf.2012.2697 23875901 13. Obogwu MB Akindele AJ Adeyemi OO Hepatoprotective and in vivo antioxidant activities of the hydroethanolic leaf extract of Mucuna pruriens (Fabaceae) in antitubercular drugs and alcohol models Chin J Nat Med 2014 12 273 83 24863352 14. Chhatre S Nesari T Somani G Kanchan D Sathaye S Phytopharmacological overview of Tribulus terrestris Pharmacogn Rev 2014 8 45 51 10.4103/0973-7847.125530 24600195 15. Samani NB, Jokar A, Soveid M, Heydari M, Mosavat SH. Efficacy of the Hydroalcoholic Extract of Tribulus terrestris on the Serum Glucose and Lipid Profile of Women With Diabetes Mellitus: A Double-Blind Randomized Placebo-Controlled Clinical Trial. J Evid Based Complementary Altern Med. 2016. 16. Alam N Hossain M Mottalib MA Sulaiman SA Gan SH Khalil MI Methanolic extracts of Withania somnifera leaves, fruits and roots possess antioxidant properties and antibacterial activities BMC Complement Altern Med 2012 12 175 10.1186/1472-6882-12-175 23039061 17. Dar PA Singh LR Kamal MA Dar TA Unique Medicinal Properties of Withania somnifera: Phytochemical Constituents and Protein Component Curr Pharm Des 2016 22 535 40 10.2174/1381612822666151125001751 26601969 18. Durg S Dhadde SB Vandal R Shivakumar BS Charan CS Withania somnifera (Ashwagandha) in neurobehavioural disorders induced by brain oxidative stress in rodents: a systematic review and meta-analysis J Pharm Pharmacol 2015 67 879 99 10.1111/jphp.12398 25828061 19. Mahajan GK Mahajan AY Mahajan RT Efficacy of aphrodisiac plants towards improvement in semen quality and motility in infertile males J Complement Integr Med 2012 17 9:6 20. Suresh S Prithiviraj E Lakshmi NV Ganesh MK Ganesh L Prakash S Effect of Mucuna pruriens (Linn.) on mitochondrial dysfunction and DNA damage in epididymal sperm of streptozotocin induced diabetic rat J Ethnopharmacol 2013 145 32 41 10.1016/j.jep.2012.10.030 23103904 21. Gauthaman K Adaikan PG Prasad RN Aphrodisiac properties of Tribulus Terrestris extract (Protodioscin) in normal and castrated rats Life Sci 2002 71 1385 96 10.1016/S0024-3205(02)01858-1 12127159 22. Ghosian Moghaddam MH Khalili M Maleki M Ahmad Abadi ME The Effect of oral feeding of Tribulus terrestris L. on sex hormone and gonadotropin levels in addicted male rats Int J Fertil Steril 2013 7 57 62 24520465 23. Subramanian S Ashwagandha-an ancient Ayurvedic drug Arogya-Journal Health Sciences 1982 8 135 9 24. Gunes S Hekim GN Arslan MA Asci R Effects of aging on the male reproductive system J Assist Reprod Genet 2016 33 441 54 10.1007/s10815-016-0663-y 26867640 25. Li R Meng X Zhang Y Wang T Yang J Niu Y Cui K Wang S Liu J Rao K Testosterone improves erectile function through inhibition of reactive oxygen species generation in castrated rats Peer J 2016 4 e2000 10.7717/peerj.2000 27168996 26. Kensler TW Wakabayashi N Biswal S Cell survival responses to environmental stresses via the Keap1-Nrf2-ARE pathway Annu Rev Pharmacol Toxicol 2007 47 89 116 10.1146/annurev.pharmtox.46.120604.141046 16968214 27. Yu M Xu M Liu Y Yang W Rong Y Yao P Nrf2/ARE is the potential pathway to protect Sprague-Dawley rats against oxidative stress induced by quinocetone Regul Toxicol Pharmacol 2013 66 279 85 10.1016/j.yrtph.2013.04.005 23624349 28. Abraham NG Kappas A Pharmacological and clinical aspects of heme oxygenase Pharmacol Rev 2008 60 79 127 10.1124/pr.107.07104 18323402 29. Shamloul R The potential role of the heme oxygenase/carbon monoxide system in male sexual dysfunctions J Sex Med 2009 6 324 33 10.1111/j.1743-6109.2008.01068.x 19215614 30. Shishodia S Aggarwal BB Nuclear Factor-kB activation: A question of life and death J Biochem Mol Biol 2002 35 28 40 10.5483/BMBRep.2002.35.1.028 16248967 31. Sahoo HB Nandy S Senapati AK Sarangi SP Sahoo SK Aphrodisiac activity of polyherbal formulation in experimental models on male rats Pharmacognosy Res 2014 6 120 6 10.4103/0974-8490.129029 24761115 32. Turk G Ceribasi AO Sahna E Atessahin A Lycopene and ellagic acid prevent testicular apoptosis induced by cisplatin Phytomedicine 2011 18 356 61 10.1016/j.phymed.2010.07.008 20739166 33. Turk G Atessahin A Sonmez M Ceribasi AO Yuce A Improvement of cisplatin-induced injuries to sperm quality, the oxidant-antioxidant system, and the histologic structure of the rat testis by ellagic acid Fertil Steril 2008 89 1474 81 10.1016/j.fertnstert.2007.04.059 17681317 34. Gencoglu H Tuzcu M Hayirli A Sahin K Protective effects of resveratrol against streptozotocin-induced diabetes in rats by modulation of visfatin/sirtuin-1 pathway and glucose transporters Int J Food Sci Nutr 2015 66 314 20 10.3109/09637486.2014.1003534 25666415 35. Chauhan NS Sharma V Dixit VK Thakur M A review on plants used for improvement of sexual performance and virility Biomed Res Int. 2014 2014 868062 25215296 36. Saksena S Dixit VK Role of total alkaloids of Mucuna pruriens Baker in spermatogenesis in Albino rats Indian J Nat Prod 1987 3 3 7 37. Amin YMN Rehman ZS Khan NA Sexual function improving effect of M. pruriens in sexually normal male rats Fitoterapial 1996 67 53 8 38. Kumar AKV Srinivasan KK Aphrodisiac activity of the seeds of Mucuna pruriens Indian Drugs 1994 31 321 7 39. Suresh S Prakash S Effect of Mucuna pruriens (Linn.) on sexual behavior and sperm parameters in streptozotocin-induced diabetic male rat J Sex Med 2012 12 3066 78 10.1111/j.1743-6109.2010.01831.x 20456630 40. Herrera Chalé F Ruiz Ruiz JC Betancur Ancona D Acevedo Fernández JJ Segura Campos MR The hypolipidemic effect and antithrombotic activity of Mucuna pruriens protein hydrolysates Food Funct 2015 20 434 44 41. Muthu K Krishnamoorthy P Evaluation of androgenic activity of Mucuna pruriens in male rats Afr J Biotechnol 2011 10 15017 9 10.5897/AJB09.896 42. Singh AP Sarkar S Tripathi M Rajender S Mucuna pruriens and its major constituent L-DOPA recover spermatogenic loss by combating ROS, loss of mitochondrial membrane potential and apoptosis PLoS One 2013 8 e54655 10.1371/journal.pone.0054655 23349947 43. Ahmad MK Mahdi AA Shukla KK Islam N Jaiswar SP Ahmad S Effect of Mucuna pruriens on semen profile and biochemical parameters in seminal plasma of infertile men Fertil Steril 2008 90 627 35 10.1016/j.fertnstert.2007.07.1314 18001713 44. Shukla KK Mahdi AA Shankwar SN Ahmad MK Effect of Mucuna pruriens on hormonal status and semen quality in infertile males Contraception 2008 78 194 10.1016/j.contraception.2008.04.110 45. Shukla KK Mahdi AA Ahmad MK Shankhwar SN Rajender S Jaiswar SP Mucuna pruriens improves male fertility by its action on the hypothalamus-pituitary-gonadal axis Fertil Steril 2009 92 1934 40 10.1016/j.fertnstert.2008.09.045 18973898 46. Amin A Lotfy M Shafiullah M Adeghate E The protective effect of Tribulus terrestris in diabetes Ann N Y Acad Sci 2006 1084 391 401 10.1196/annals.1372.005 17151317 47. Georgiev P Dimitrov M Vitanov S Effect of Tribestan (from Tribulus terrestris) on plasma testosterone and spermato-genesis inmale lambs and rams Veterinarna Sbirka 1988 86 20 2 48. Gauthaman K Ganesan AP Prasad RNV Sexual effects of puncturevine (Tribulus terrestris) extract (protodios-cin): an evaluation using a ratmodel J Altern Complement Med 2003 9 257 65 10.1089/10755530360623374 12804079 49. Adaikan PG Gauthaman K Prasad RNV Ng SC Proerectile pharmacological effects of Tribulus terrestris extract on the rabbit corpus cavernosum Ann Acad Med Singapore 2000 29 22 6 10748960 50. Kamboj P Aggarwal M Puri S Singla SK Effect of aqueous extract of Tribulus terrestris on oxalate-induced oxidative stress in rats Indian J Nephrol 2011 21 154 9 10.4103/0971-4065.83727 21886973 51. Kim HJ Kim JC Min JS Kim MJ Kim JA Kor MH Aqueous extract of Tribulus terrestris Linn induces cell growth arrest and apoptosis by down-regulating NF-κB signaling in liver cancer cells J Ethnopharmacol 2011 14 197 203 10.1016/j.jep.2011.04.060 21549825 52. Garg LC Parasar GC Effect of Withania somnifera on reproduction in mice Planta Med 1965 13 46 7 10.1055/s-0028-1100096 53. Misra DS Maiti RK Bera S Das K Ghosh D Protective effect of composite extract of Withania somnifera, Ocimum sanctum and Zingiber officinale on swimming-induced reproductive Endocrine dysfunctions in male rat IJPT 2005 4 110 7 54. Kyathanahalli CN Manjunath MJ Muralidhara Oral supplementation of standardized extract of Withania somnifera protects against diabetes-induced testicular oxidative impairments in prepubertal rats Protoplasma 2014 251 1021 9 10.1007/s00709-014-0612-5 24488064

PMC005xxxxxx/PMC5000418.txt

==== Front Orphanet J Rare DisOrphanet J Rare DisOrphanet Journal of Rare Diseases1750-1172BioMed Central London 50310.1186/s13023-016-0503-2ReviewHealth-related quality of life in mucopolysaccharidosis: looking beyond biomedical issues Hendriksz Christian J. 0161 206 4365cfya@sky.com 110Berger Kenneth I. Kenneth.Berger@nyumc.org 2Lampe Christina Christina.lampe@helios-kliniken.de 3Kircher Susanne G. susanne.kircher@meduniwien.ac.at 4Orchard Paul J. orcha001@umn.edu 5Southall Rebecca rebeccasouthall@nhs.net 6Long Sarah S.A.Long@bath.ac.uk 7Sande Stephen stephen.sande@bmrn.com 8Gold Jeffrey I. jgold@chla.usc.edu 91 Adult Inherited Metabolic Disorders, Consultant Transitional Metabolic Medicine, The Mark Holland Metabolic Unit, Salford Royal NHS Foundation Trust, Ladywell NW2- 2nd Floor Room 112, Salford Manchester, M6 8HD UK 2 Division of Pulmonary, Critical Care and Sleep Medicine, New York University School of Medicine and André Cournand Pulmonary Physiology Laboratory, Bellevue Hospital, New York, USA 3 Centre for Rare Diseases, Clinic for children and adolescents, Helios Dr. Horst Schmidt Kliniken, Wiesbaden, Germany 4 Institute of Medical Chemistry and Medical Genetics, Medical University of Vienna, Vienna, Austria 5 Department of Pediatrics, Division of Blood & Marrow Transplantation, University of Minnesota, Minneapolis, MN USA 6 GB Prohealth Ltd, Lichfield, UK 7 School of Sociology and Social Policy, University of Bath, Bath, UK 8 BioMarin Pharmaceutical Inc., Novato, CA USA 9 Keck School of Medicine, Departments of Anesthesiology, Pediatrics, and Psychiatry & Behavioral Sciences, Children’s Hospital Los Angeles, Anesthesiology Critical Care Medicine, Pediatric Pain Management Clinic, University of Southern California, California, USA 10 Paediatrics and Child Health, University of Pretoria, Steve Biko Academic Unit, Pretoria, South Africa 26 8 2016 26 8 2016 2016 11 1 11922 4 2016 17 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.The mucopolysaccharidoses (MPS) comprise a heterogeneous family of rare, genetic lysosomal storage disorders that result in severe morbidity and reduced life expectancy. Emerging treatments for several of these disorders have triggered the search for clinically relevant biomarkers and clinical markers associated with treatment efficacy in populations and individuals. However, biomedical measures do not tell the whole story when characterizing a complex chronic disorder such as MPS. Health-related quality of life (HRQoL) tools that utilize patient reported outcomes to address patient parameters such as symptoms (pain, fatigue, psychological health), functioning (activity and limitations), or quality of life, have been used to supplement traditional biomedical endpoints. Many of these HRQoL tools have demonstrated that quality of life is negatively impacted in patients with MPS. There is both the opportunity and need to formally standardize and validate HRQoL tools for the different MPS disorders. Electronic supplementary material The online version of this article (doi:10.1186/s13023-016-0503-2) contains supplementary material, which is available to authorized users. Keywords MucopolysaccharidosesQuality of lifeEnzyme replacement therapyClinical trialPain measurementEQ-5DMPS HAQHRQoLADLhttp://dx.doi.org/10.13039/100008484BioMarin Pharmaceuticalissue-copyright-statement© The Author(s) 2016 ==== Body Background The mucopolysaccharidoses (MPS) are a group of lysosomal storage disorders associated with accumulation of glycosaminoglycans (GAGs) in tissues and organs due to enzyme deficiencies required for degradation of cellular GAGs (Additional file 1). Shared clinical features of the MPS disorders include skeletal deformities such as kyphosis, scoliosis, pectus carinatum, valgus deformities of the knees, and carpal tunnel syndrome, joint abnormalities, spinal cord compression, reduced growth, coarse facial features, vision and hearing damage, and cardiorespiratory manifestations [1, 2]. Intellectual and neurological impairment occurs in some MPS subtypes (I, II, III, and VII) due to GAG accumulation in the brain [1, 2]. MPS patients generally appear healthy at birth with clinical manifestations gradually worsening with age. There is a wide variety of clinical presentations and progression rates among the different MPS subtypes, and within the diseases themselves. Many of these disorders lead to severe morbidity and premature death [3]. Therapies including enzyme replacement therapy (ERT) and hematopoietic stem cell transplantation (HSCT) have altered the course of morbidity and mortality for some of the MPS disorders [4, 5]. Biochemical and clinical measures assessing clinical benefit for regulatory purposes have been well characterized for several of the MPS disorders and include urine GAG metabolites, 6-min walk test and pulmonary function tests. What these biomedical endpoints mean in daily life to patients and their caregivers is less well understood; and payers are increasingly asking for evidence that these treatments are having an effect that is “meaningful” to both patients and families. The term “quality of life” is an expansive multi-dimensional concept that typically includes subjective assessments of both positive and negative aspects of life [6]. While health is an important facet of overall quality of life, it is not the only one. Other aspects, including occupation, environs, school, ethos, beliefs, and spirituality are important domains of quality of life and add to the inherent difficulty of its measurement. The concept of “health-related quality of life (HRQoL),” specifically comprises those areas of quality of life that can clearly be shown to affect health – physical, mental, emotional, and social functioning. In fact, the U.S. Centers for Disease Control has defined HRQoL as “an individual’s or group’s perceived physical and mental health over time” [7]. It should be noted that HRQoL is usually measured through self-assessment; however, if the patient is too ill or too young, a caregiver/parent assessment can serve as proxy. In many disease states, HRQoL tools that attempt to measure patient or caregiver outcomes, are used to supplement traditional measures of morbidity, mortality and the effects of treatment. The purpose of this review is to survey the HRQoL tools that have been used to study MPS disorders, and to examine the impact of treatments on patient reported outcomes (PROs). Methodology Relevant literature was obtained from clinical trial publications and PubMed searches for MeSH terms “(quality of life[MeSH Terms]) AND mucopolysaccharidoses[MeSH Terms]” (30 articles) and free text “(mucopolysaccharidosis) AND [(quality of life) or (pain) or (fatigue)]” (151 articles). Additional publications were identified from reference lists within the most relevant MPS-related papers focusing on PROs, fatigue, pain, and HRQoL. The literature search was completed in June 2015. How does MPS affect HRQoL? The multi-organ clinical manifestations of MPS can lead to poor endurance and mobility, often associated with pain, restricted range of motion (ROM), low energy levels, and fatigue which negatively affect HRQoL and activities of daily living (ADL) (Fig. 1). MPS patients may experience increased physical and emotional dependence on family and friends, reduced participation in school, work and social life, low self-esteem, and psychological, behavioral and mental health conditions such as anxiety and depression (Fig. 1) [8]. Impaired vision and hearing and frequent surgeries may further reduce physical activity, while negatively affecting interpersonal functioning, social life, educational engagement, employment, and the ability to live independently [9–12] (Fig. 1).Fig. 1 Important factors affecting HRQoL in patients with MPS. Some of the manifestations may also have a direct impact on ADL, participation in school/employment or social life (due to e.g. surgery, cognitive impairment) Impaired mobility is prevalent in MPS patients, with many individuals requiring walking aids or a wheelchair [11, 13–15] (Fig. 2). Mobility problems may be due to skeletal and joint abnormalities, spinal cord compression, pain in the lower extremities, and reduced energy levels caused by cardiorespiratory issues [2]. Joint abnormalities can result in poor shoulder ROM, wrist weakness, stiffness or changes in mobility which in turn affect simple ADL tasks such as dressing, washing and eating [16]. Pain may arise from joint defects, infections including otitis media, neurological involvement and neuropathic signals arising in the brain, increased intracranial pressure, spinal cord compression, or carpal tunnel syndrome [13, 15, 17]. Fatigue, the result of impaired cardiopulmonary function, can produce stress, anger, frustration, and potentially depression [18].Fig. 2 Mobility impairment in the International Morquio A registry (including 326 patients with MPS IVA) [14], the Morquio A Clinical Assessment Program (MorCAP) (including 325 patients with MPS IVA) [11], the MPS VI Survey (including 121 patients with MPS VI) [13] and the Dutch MPS Survey (including 55 patients with MPS I, II, III, IV, and VI) [15] Patient Reported Outcome (PRO) measures in MPS PRO measures are collected by standardized questionnaires designed to measure explicit concepts such as symptoms (pain, fatigue, psychological health), functioning (activity limitations), HRQoL, or quality of life (QoL) [19]. Thousands of PRO instruments have been described including both generic and disease-specific questionnaires [20, 21]. The advantage of generic questionnaires lies in their broad applicability across different disease types, severities and medical interventions, and among diverse demographic and cultural groups, allowing comparison across studies and diseases [22]. Disease-specific questionnaires are intended for a particular patient population with questions designed to be relevant, meaningful and acceptable for that affected population, and may be used to measure the efficacy of interventions and treatments. PROs used in clinical trials with MPS patients are summarized in Table 1 and in Additional file 2, including information regarding age ranges, outcomes, and type of respondent.Table 1 PROs used in patients with MPS Name questionnaire Acronym Age range (yrs) Assessment of Completed bya Reference Symptom PROs  Pain Visual Analog Scalesb VAS ≥8 Pain intensity Patient  Adolescent Pediatric Pain Tool APPT 8–17 Pain location, description and intensity Patient [60, 61]  Brief Pain Inventory Short Form BPI-SF Adults Severity of pain, impact of pain on daily function, location of pain, use of pain medications, amount of pain relief Patient [30, 62]  Six-face Faces Pain Scale-Revised FPS-R ≥8 Pain intensity Patient [15]  Non-communicating Children’s pain Checklist-Revised NCCPC-R 3–18c Pain-associated behavior Observer [15, 63]  Achenbach System of Empirically Based Assessment Adult Self Report ASEBA ASR 18–59 Social-adaptive and psychological symptoms Patient [30]  Achenbach System of Empirically Based Assessment Older Adult Self Report ASEBA OASR ≥60 Social-adaptive and psychological symptoms Patient [30]  Yatabe-Guilford Personality testd Y-G test NA personality and psychiatric aspects Patient [31]  Tree-drawing test (Baum test) TDT NA personality and psychiatric aspects Patient [31]  General Health Questionnaire 60 GHQ-60 Adolescents and adults Mental health Patient [31]  State-Trait Anxiety Inventory STAI Adolescents and adults Anxiety Patient [31] Functioning PROs  Health Assessment Questionnaire HAQ >18 Functional capacity and independence in activities of daily living, pain, overall well-being Patient [13, 64]  Childhood Health Assessment Questionnaire CHAQ ≤18 Patient [13, 65]  Mucopolysaccharidosis Health Assessment Questionnaire MPS HAQ Children and adults Self-care, mobility skills, extent of caregiver assistance in performing activities Patient [32]  Hunter Syndrome-Functional Outcomes for Clinical Understanding Scale HS-FOCUS Children (>12) and adults Impact of MPS II on function Patient [27]  Modified version of the Functional Independence Measuree FIM Children and adults Physical and cognitive disability Observer [31, 34]  Pediatric Evaluation of Disability inventory PEDI 0.5–7.5 Capability and performance in self-care, mobility and social function Patient or parentf [66, 67]  Vineland Adaptive Behavior Scales VABS Children and adults Adaptive behavior Parent/caregiver [36, 37, 39, 40]  Behavior Assessment System for Children BASC Children Emotional adjustment and adaptive behavior Patient and/or parent [37]  Scales of Independent Behavior-Revised SIB-R Infancy-80+ Adaptive behavior Patient [68] Health-related quality of life  EuroQol 5D EQ-5D Versions for children (≥8) and adults (≥16) Physical and mental health Patient [28, 69]  Short form-36 SF-36 ≥16 Physical and mental health Patient [15, 30, 70]  Health Utilities Index HUI ≥5 Impact of disease and therapy Patient or parent [27]  Pediatric Quality of Life inventory PedsQL Children physical, emotional, social, and school functioning Patient or parent [15, 43, 71]  TNO-AZL Preschool children Quality of Life TAPQOL 0.5–5 Physical, social, cognitive, and emotional functioning Parent [72]  TNO-AZL Children Quality of Life TACQOL 6–15 Health status and children’s subjective emotional appraisal of their health Patient or parent [73]  Childhood Health Questionnaire CHQ 5–18 functional capacity and independence in activities of daily life Patient and parent [27, 37, 39] Impact on family/caregivers  Pediatric Quality of Life inventory Family Impact Module PedsQL Family Impact Module Children Parent’s problems in physical, emotional, social, and cognitive functioning, communication, worry, and problems specific to the family’s daily activities and family relationships Parent [74]  Zarit Burden Interview ZBI Adult patients Burden of caring on relationship, emotional well-being, social and family life, finances, control over one’s life Parent/caregiver [8] aIn MPS studies bPain VAS scores that have been used in MPS patients are included in the HAQ, CHAQ, EQ-5D, APPT cPatients who are unable to speak because of intellectual impairments or disabilities dJapanese version of the Guilford test eAdapted for patients with MPS fNormally completed by parent or observer NA not available Disease impact on PROs in MPS Impact of MPS on self-reported symptoms: pain, fatigue and psychological health Pain has been assessed as an exploratory endpoint in several clinical trials evaluating ERT, mostly using the (Childhood) Health Assessment Questionnaire ((C)HAQ) Pain Scale or a modified version [23–26]. Baseline pain measurements from these trials recorded before patients were treated with ERT, and data from a number of other studies using (C)HAQ (Tables 2 and 3), indicate that MPS patients can experience considerable pain [23]. A score of 0.93 on a scale from 0 (no pain) to 3 (severe pain) has been reported for untreated patients with MPS I (N = 30) [23]. Mean Pain Scale scores reported for patients with MPS VI in the phase II study and the MPS VI Survey Study varied between 30 and 40 on a scale from 0 to 100, corresponding with mild to moderate pain, while scores were somewhat higher in older patients (>18 years) [13, 25]. A score of 28 has been reported for patients with attenuated MPS II [27].Table 2 Clinical studies assessing PROs in patients with MPS, excluding ERT trials Reference MPS type N Age (yrs) PRO instrument [40] MPS IH 41 NAa VABS [39] MPS IH 47 Mean 10.5 VABS II CHQ [34] MPS II 27 5–41 FIM [31] MPS II 10 Mean 23.2 FIM Personality tests: Y-G test, Tree-drawing test Psychological tests: GHQ-60, STAI [38] MPS II 50 Mean 6.0 Different standardized tests for cognitive, adaptive, language, and motor functions [35] MPS II 29 Mean 11.5 MPS HAQ [27] MPS II 96 patients & caregivers Mean 14.2 CHAQ HS-FOCUSb CHQ HUI3 [36, 43] MPS II 73 patients & parents Mean 12.5 PedsQL Peds QL Family Impact Module VABS II [37] MPS II 15 10.8 VABS II CHQ BASC-2 [14] MPS IVA 326 1–73 (C)HAQ [11] MPS IVA 325 Mean 14.5 MPS HAQ [8, 28] MPS IVA 63 patients 56 caregivers 5–17 years (N = 36) ≥18 years (N = 27) Patients: EQ-5D, APPT (<18 years)/BPI-SF (≥18 years), fatigue question Caregivers : caregiver questionnaire, ZBI [30] MPS IVA 20 NA ASEBA ASR/ OASR SF-36 BPI [41] MPS VIA 24 10–17 (N = 10) 18–54 (N = 14) EQ-5D [13] MPS VI 121c 4–56 (C)HAQ [15] MPS I, II, III, IV, VI 55 Median 11.3 MPS-specific questionnaire NCCPC-Rd FPS-Rd Pain VASd SF-36d PedsQL [42] MPS I, II, IVA, IVB, VI 81 ≥18 EQ-5D amean age at transplant was 21.7 months; mean years of follow-up from transplant was 67.2 months bHS-FOCUS completed by 53 patients aged ≥12 years cDisability, Pain and Arthritis scores for 91, 90, and 81 patients ≤18 years, respectively and Disability and Pain scores for 29 and 28 patients >18 years, respectively dNCCPC-R was completed by parents of patients <8 years or with intellectual disability (N = 35) and the FPS-R by patients 8–18 years with no intellectual disability (N = 11); eight patients completed the Pain VAS, 16 patients over 18 years completed the SF-36; 35 participants (patients or parents) completed the PedsQL NA not available Table 3 Clinical studies assessing the impact of ERT on PROs in MPS patients Reference MPS type Treatment Comparator N Mean age (years)a Study duration PRO instrument [33] I 83 % Hurler-Scheie, 13 % Scheie iv laronidase (0.58 mg/kg/week) Placebo 45 15.6 26 weeks (C)HAQ [23] I iv laronidase (0.58 mg/kg/week) / 45 15.7 3.5 year (extension of [66] (C)HAQ [52] I iv laronidase (0.58 mg/kg/week) / 5 12.0 6 year Modified MPS HAQ [53] I Scheie, Hurler- Scheie iv laronidase (0.58 mg/kg/week) / 7 16.3 52–208 weeks MPS HAQ [55] II iv idursulfase (0.5 mg/kg/week) / 94 14.5 2 years extension (C)HAQ [32, 45] IVA iv elosulfase alfa (2.0 mg/kg every other week or weekly) Placebo 176 15.3 and 13.1 24 weeks MPS HAQ [29] IVA iv elosulfase alfa (2.0 or 4.0 mg/kg/week) / 25 13.7 27 weeks APPT [24] VI iv galsulfase (1.0 or 2.0 mg/kg/week) / 5 11.0 48 weeks (C)HAQ [25] VI iv galsulfase (1.0 mg/kg/week) / 10 12.7 48 weeks (C)HAQ [26] VI iv galsulfase (1.0 mg/kg/week) Placebo 39 13.7 24 weeks Joint pain and stiffness, physical energy level [56] VI iv galsulfase (1.0 mg/kg/week) / 9 NA 2 years (C)HAQ [5] VI iv galsulfase (1.0 mg/kg/week) / 55 12.0 6.8 ± 2.2 years (C)HAQ [44] VI iv galsulfase (1.0 mg/kg/week) / 8 6.8 1.0–4.5 years TAPQOL/TACQOLb aMean age at baseline from all patients or from ERT group bThe TAPQOL was completed by the parents of four patients <6 years, TACQOL was completed by seven parents of patients ≥6 years NA not available Several studies have evaluated pain in MPS in more detail using other questionnaires [15, 28, 29]. In the Dutch National MPS Survey of 55 patients with different types of MPS, joint pain was evaluated with the Non-communicating Children’s Pain Checklist-Revised (NCCPC-R), the Six-face Faces Pain Scale-Revised (FPS-R) and an MPS-specific questionnaire. Overall, 69 % of patients reported pain, mainly hip and back pain (27.8 and 25.9 %, respectively), with a pain score above the critical cut-off value for significant pain in 40 % of cases [15]. Somewhat surprisingly, pain was most frequently reported for patients with cognitive impairment, particularly for MPS III, while patients with MPS IV (which is not associated with cognitive impairment) appeared to experience the most severe pain [15]. The unexpected high prevalence of pain in cognitively impaired patients suggests that pain may be underestimated in this group, but may also reflect difficulties with parents distinguishing between MPS- and pain-related behavior (as assessed in the NCCPC-R) in these patients. The finding from the Dutch survey that pain was most severe in MPS IV patients is not unexpected given the severe skeletal and joint abnormalities in these individuals. Consistent with this finding, a phase II MPS IVA study reported a pain intensity score of 4.6 on the Adolescent Pediatric Pain Tool (APPT) at baseline, indicating medium pain [29]. In addition, an international MPS IVA PRO survey reported joint pain in 74 % of adults (N = 27) and 64 % of children (N = 36), as documented using the Brief Pain Inventory Short Form (BPI-SF) and APPT [28]. In both studies, pain was described most often in the lower extremities [28, 29]. The PRO survey also demonstrated an association between pain and mobility as measured by wheelchair use. Adult patients who sometimes used a wheelchair tended to report more severe and widespread pain than those always using a wheelchair, while pain interference with daily activities was highest in the latter group [28]. This suggests adult MPS IVA patients may tolerate considerable pain if mobility and wheelchair independence are retained. Our literature search revealed only a single study assessing fatigue. The aforementioned MPS IVA PRO study, assessed fatigue/low stamina by querying patients on the number of evenings per week that they reported feeling extremely tired. Using this definition, 63 % of adults and 69 % of children reported feeling fatigued, a high prevalence warranting further investigation. Possible contributions from pulmonary or cardiac causes would be difficult to distinguish in many patients. To date, two studies have evaluated the psychological health of patients with MPS [30, 31]. A study in ten MPS II patients showed that many had difficulties establishing relationships and that patients and their parents had increased levels of anxiety [31]. A correlation was found between psychological status and ADL, suggesting that reduced ADL negatively affects psychological status [31]. Another study of 20 MPS IVA patients showed psychological symptoms (at least one or more ASEBA [Achenbach System of Empirically Based Assessment, which assesses social-adaptive function deficits and psychological symptoms] problem Scales within the symptomatic range) in 11 individuals [30]. Interestingly, these patients had higher pain severity scores and pain interference scores on the BPI, suggesting that pain and psychological issues, including depression, may be interdependent. Impact of MPS on patient functionality ADL, as assessed by the MPS Health Assessment Questionnaire (MPS HAQ), have been measured as an exploratory endpoint in some clinical trials [32, 33], as well as in a number of studies in patients with MPS II, IV and VI [11, 14, 34, 35] (Table 2). Overall difficulties with mobility and self-care, which tend to increase with age, have been reported. In patients with MPS II, cognitive decline negatively affects ADL. A study of 96 patients with attenuated MPS II (age 5.0–30.9 years) reported impairments in walking/standing and reach/grip domains of the Hunter Syndrome-Functional Outcomes for Clinical Understanding Scale (HS-FOCUS) and impairments in hygiene, reach and dressing, and grooming domains of the CHAQ [27]. HS-FOCUS function scores were lower in patients with better endurance in the 6MWT (r = −0.6) and better joint mobility (r = −0.3). Another smaller study (N = 29; age 2–29 years) suggested that difficulties with ADL (as assessed using the MPS HAQ) in MPS II patients mainly depend on the cognitive status and age of these patients [35]. Younger patients with normal mental development were generally independent with regard to self-care, mobility and walking, but assistance with daily activities increased with age [35]. Cognitively impaired MPS II patients required moderate or complete caregiver assistance in self-care within all categories [35]. Two studies used the Functional Independence Measure (FIM) to assess ADL in patients with MPS II [31, 34]. In patients with severe MPS II, cognitive scores decreased rapidly, reaching a minimum score at about 7 years of age, in contrast to motor scores, which decreased more slowly. In slowly progressing MPS II patients, total FIM scores increased with age, similar to increases in FIM scores seen in healthy children [34]. In patients with MPS II, daily living skills have also been assessed as part of adaptive behavior scales [36, 37]. In both mild and severe forms of MPS II (N = 73), the Vineland-II Adaptive Behavior Scales (VABS II) showed significantly reduced functioning in communication, daily living skills, socialization, and motor skills as compared to normative data [36], but scores were significantly lower (more severe) in severe than in mild MPS II. A study including 15 patients with slowly progressing MPS II showed adaptive skills within the average range on the VABS II, as well as the Behavior Assessment System for Children (BASC)-2 Parent Rating Scale [37]. Daily living skills domain scores of the VABS II decreased significantly with age across patients. Children aged ≥12 years showed an increasing sense of inadequacy and anxiety as well as decreasing self-esteem over time in the BASC-2. In a retrospective review of longitudinal data from 50 patients with MPS II, two groups of patients could be distinguished based on adaptive behavior data (obtained using the Scales of Independent Behavior, Revised [SIB-R] and the Pediatric Evaluation of Disability Inventory [PEDI]): one group reaching a plateau at around 48–60 months and then declining and one group maintaining relatively normal adaptive abilities over time [38]. In patients with MPS IH, the VABS has been used to evaluate the impact of HSCT on adaptive skills [39, 40]. These studies are discussed below under “Effects of therapy on HRQoL in MPS”. MPS IVA and MPS VI have also been shown to significantly interfere with patients’ ADL [14]. In the International Morquio A registry of 326 MPS IVA patients, only 40–60 % of patients were able to perform ADL independently [14]. In the MorCAP study with 325 MPS IVA patients, 20–40 % reported self-care ADL tasks (including the ability to wash or brush hair, tie shoelaces and cut fingernails) were affected by their disease (Fig. 3) [11]. In the Survey Study of 121 MPS VI patients, the (C)HAQ disability index indicated a mild level of disability in patients aged >18 years (mean 1.0) and moderate disability in those aged ≤18 years (mean 2.0) [13].Fig. 3 Impact of MPS on self-care ADL as measured by the MPS HAQ in the MorCAP study including 325 patients with MPS IVA (mean age 14.5 years) [11] Impact of MPS on health-related quality of life (HRQoL) Several studies report significant disease impact on HRQoL in patients with MPS disorders (Table 2). Overall, the greatest deviations from a healthy population were seen in domains of pain/discomfort and mobility. Problems with self-care or usual activities were also critical factors affecting HRQoL. In addition, wheelchair use, unemployment, poor endurance, and poor pulmonary function were also associated with worse HRQoL [28, 41]. Despite the deviation in pain domains, no differences in HRQoL could be found between patients with or without pain [15, 28]. Although pain has been identified as a significant issue for patients with MPS, other symptoms such as mobility appear to have greater impact on HRQoL. Two studies used the generic Short Form-36 (SF-36) to assess HRQoL in patients with MPS (Table 2). In 16 adults from the Dutch national MPS survey, including patients with MPS I, II, III, IV and VI, deviations from average were predominantly seen in the physical component score (29–30 vs. 50 in a reference population) [15]. The largest deviation was observed in the bodily pain domain (37–41 vs. 81), possibly due to bone pain reported in 68 % of patients. The Pediatric Quality of Life (PedsQL) was used in this study to assess HRQoL in patients <18 years, showing the largest deviations compared with healthy individuals in the PedsQL physical score (53–57 vs. 79–85 in a reference population). A study including 20 patients with MPS IV recently showed scores below the US mean in physical, but not mental, health on the SF-36 [30]. Although several patients had psychological symptoms on the ASEBA Adult Self-Report (ASR), these did not seem to affect HRQoL outcomes. Two studies used the generic Euroqol-5 dimensions (EQ-5D) questionnaire, assessing mobility, self-care, usual activities, pain/discomfort, and anxiety/depression in patients with MPS. Lavery et al. examined 81 adult patients from the UK and USA with various types of MPS and a mean utility value of 64.1 (with 100 indicating best health) [42]. HRQoL in these patients was mainly affected by mobility impairment and pain/discomfort, and to a lesser degree by problems with self-care or performing usual activities [42]. In the international PRO survey using the EQ-5D, Hendriksz et al. [28] reported impairment in mobility, self-care, usual activities, pain/discomfort, and anxiety/depression for both adults (N = 27) and children (N = 36) with MPS IVA (Fig. 4). Results from this study indicated that HRQoL diminished with increasing wheelchair use. Adult patients who used a wheelchair sometimes (when needed) had a mean utility value of 0.582, a score comparable to that for patients with chronic ischemic heart disease or non-insulin dependent diabetes mellitus (Table 4). The utility value of adults using a wheelchair all the time (0.057) was only slightly better than that of bed ridden or completely immobile multiple sclerosis patients (Table 4). EQ-5D utility values were also considerably lower in unemployed (0.275) than in employed (0.640) MPS patients [28]. A study in a subset of German MPS IVA patients (N = 24) from the international PRO survey showed strong correlations of EQ-5D utility values with endurance in the 6-min walk test and 3-min stair climb test (R = 0.884 and R = 0.852, respectively) and with pulmonary function (forced vital capacity: R = 0.815; maximum voluntary ventilation: R = 0.825), suggesting that these measures might be used as surrogate measures for HRQoL in patients with MPS IVA [41].Fig. 4 Mean score for the five EQ-5D domains in 61 MPS IVA patients (adults and children) [28] Table 4 Comparison of HRQoL according to the EQ-5D utility value in MPS IVA (Morquio A) patients with different levels of mobility impairment and other serious chronic diseases Disease HRQoL mean utilitya MPS IVA: adults, no wheelchair [28] 0.846 MPS IVA: adults, sometimes wheelchair [28] 0.582 MPS IVA: adults, always wheelchair [28] 0.057 Multiple sclerosis, walking-aid [75] 0.460 Multiple sclerosis, bedridden [75] −0.195 Moderate to severe rheumatoid arthritis [76] 0.489 Chronic ischaemic heart disease [77] 0.640 Non-insulin dependent diabetes mellitus [77] 0.670 aEQ-5D utility scores were calculated using a time tradeoff method [78]. This generates scores ranging from −0.59 to 1, where 1 means full health and zero stands for death. Negative scores could be emotively interpreted as a health state “worse than death”. Tradeoff tariffs used differed depending on the ethnic background of patients Several MPS studies have used generic instruments to measure HRQoL in children. The Dutch national MPS survey of patients with MPS I, II, III, IV and VI [15] used the PedsQL to assess HRQoL in patients <18 years, with the largest deviations compared to healthy individuals seen in the PedsQL physical score (53–57 vs. 79–85 in a reference population). The PedsQL was also used in a study of 73 patients with MPS II and their parents, showing reduced scores in all domains (physical, emotional, social, and school functioning) versus healthy individuals and patients with several other chronic illnesses (cancer, maple syrup urine disease, galactosemia) [43]. In slowly progressing MPS II patients (N = 96), Raluy-Callado et al. [27] demonstrated significant distress and dysfunction in global health, physical functioning and role/social-limitations-physical and bodily pain, as measured by the generic Childhood Health Questionnaire (CHQ) (N = 96). Low scores were reported in the self-esteem and family cohesion domains, suggesting that MPS II has a severe psychological impact on patients and their parental caregivers. It is unclear how these domains were affected by cognitive function. A clear correlation between joint ROM and better physical functioning scores of the CHQ (r = 0.5) was observed. A smaller study in 15 slowly progressing MPS II patients showed a CHQ psychosocial summary score within the normal range. The physical summary score was 1.5 standard deviations below the normative average for the whole group, >2 standard deviations below average in children ≥12 years (N = 10), and tended to worsen with age [37]. Finally, Brands et al. [44] used the TNO-AZL Child Quality of Life (TACQOL) and TNO-AZL Preschool Children’s Quality of Life (TAPQOL) questionnaires to evaluate the impact of ERT on HRQoL in children with MPS VI aged 6–15 years (N = 7) and aged 6 months to 6 years (N = 4), respectively (Table 3). Baseline data reflected the greatest deviations from healthy peers in lung problems, social functioning, motor functioning and positive mood domains of the TAPQOL and in body and motor domains of the TACQOL [44]. Impact of MPS on caregivers Only a few studies, to date, have addressed the impact on caregivers looking after individuals with MPS. In a study including 73 caregivers of patients with MPS II (both mild and severe forms), the PedsQL Family Impact Module showed that the impact of the disease on the family is similar to that for other pediatric outpatients with chronic illnesses [36]. Domain scores for family HRQoL, family functioning summary, total scale score, physical functioning, social function, daily activities, and family relationships negatively correlated with the severity of illness. In the international PRO survey in MPS IVA patients [8], outcomes of the Zarit Burden Interview (ZBI), the MPS HAQ, and a caregiver questionnaire revealed that MPS IVA poses a large burden on caregivers, affecting their physical and emotional health, family life, social life and financial situation. Caregiver burden increased with disease progression and mobility problems. Wheelchair use by MPS IVA patients had a profound negative impact on caregiver’s support (Fig. 5) [8]. Because wheelchair-bound patients require much more caregiver support than those using a wheelchair occasionally, the investigators concluded that even small improvements in patient mobility might substantially reduce the level of caregiver support and the burden of caregiving.Fig. 5 Level of assistance required from caregivers for performing daily activities as measured by the MPS HAQ in adult MPS IVA patients, according to wheelchair use [8] Effects of therapy on HRQoL in MPS Therapies for MPS Two treatment options target the pathophysiology of MPS: HSCT and ERT. HSCT is primarily used in MPS conditions with a neurologic component since difficulties exist in delivery of intravenous enzyme products across the blood-brain barrier [9]. ERT is currently available to treat MPS I, II, IVA and VI. Several randomized, placebo-controlled phase II/III clinical studies have demonstrated favorable effects of ERT on urinary GAG levels, endurance, respiratory function, joint ROM, hepatomegaly, growth/height, and cardiac function [25, 45–49]. HSCT While HSCT has been considered the standard of care for the severe form of MPS I (Hurler syndrome) for decades [50], effects on patient reported HRQoL are not well studied. A few studies in patients with MPS IH and MPS II have been published. A study of 41 MPS IH children transplanted at a mean age of 21.7 months and followed for 2–21 years (mean follow-up 67.2 months) showed declining adaptive behavior scores on the VABS over time, indicating development of skills at a lower than average rate compared with unaffected peers [40]. VABS scores were significantly better in transplanted patients after the age of 2 years when compared to a cross-sectional non-transplanted MPS IH group. Cognitive ability, not age, at transplant correlated significantly with the ultimate adaptive level. Another study of 47 MPS IH patients transplanted between 6 and 44 months and evaluated 1–24 years post-HSCT showed no significant impact of the type of transplant, number of transplants, age at transplant, time since transplant, or total body irradiation treatment on adaptive functioning on the VABS. [39]. However, individuals undergoing HSCT at an older age reported poorer physical QoL on the CHQ. In addition, patients receiving unrelated bone marrow HSCT exhibited poorer psychosocial QoL compared with those receiving bone marrow HSCT from a relative. Finally, a retrospective study in 13 HSCT-treated, Japanese MPS II patients showed stable or improved ADL (school status, movement and daily activities, conversation, and toileting) versus baseline in most patients after a mean follow-up of 9.6 years [51]. ERT Several clinical trials evaluating ERT in patients with MPS I, II, IVA and VI have included PRO measures as exploratory efficacy endpoints using the (C)HAQ or MPS HAQ (Table 3). It should be noted that these studies were not powered to assess the true effect of ERT on PRO measures and results should be interpreted with caution. MPS I Studies in patients with MPS I describe improvements in ADL, pain and HRQoL after long-term ERT [23, 52, 53]. The 3.5-year extension of the laronidase phase III study (N = 45) showed a stable or improved (C)HAQ Disability Index in 77 % of patients (57 % improved) [23]. The mean decrease of 0.31 was considered clinically meaningful based on data from patients with rheumatoid arthritis [54]. In the 30 patients in this study with available pain data, the Pain Index decreased from 0.93 at baseline to 0.56 after long-term treatment. A smaller study of 7 patients with attenuated MPS I (Scheie or Hurler-Scheie) showed significant improvements in ADL (eating/drinking, dressing, tooth brushing, toileting, and walking) as documented by the MPS HAQ after 52–208 weeks of ERT [53]. The impact of ERT on HRQoL in patients with MPS I has been assessed in a 6-year re-evaluation of patients enrolled in the original laronidase phase I/II trial using an MPS-specific QoL questionnaire containing 100 questions. The largest effects of treatment were reported for energy, endurance, independence (personal hygiene, dressing, transfers), sleep quality, participation in daily activities, and self-esteem [52]. MPS II One long-term open-label study has investigated the effect of ERT on PRO measures in patients with MPS II. This study demonstrated statistically significant improvements from baseline in the CHAQ Disability Index in 48 patients aged ≥12 years as soon as 20 months after ERT initiation [55]. Disability was also evaluated by a parent-assessed Disability Index completed by 81 parents, showing significant improvements 8 months after onset of ERT. MPS IVA Studies in patients with MPS IVA have shown improvements in ADL and pain within a relatively short time after ERT initiation [29, 32, 45]. The MPS HAQ was used to evaluate the effect of ERT on ADL in 176 patients with MPS IVA in a phase III study. After 24 weeks, small to modest improvements in caregiver assistance and mobility domains were observed for ERT compared to placebo, though with wide confidence intervals [32, 45]. Interim results of an ongoing randomized, double-blind study in 25 MPS IVA patients with relatively good endurance showed clear (numerical) improvements in pain, as assessed using the APPT, after 24 weeks of ERT [29]. MPS VI Studies in patients with MPS VI have shown improvements or no change in ADL, pain and HRQoL in patients receiving ERT [5, 24–26, 44, 56]. The (C)HAQ, or a modified version, was used in all clinical trials evaluating the efficacy of ERT in MPS VI, showing improvements in pain and arthritis or joint stiffness and improvements in ADL (e.g. picking up coins, tying shoelaces, pulling shirt overhead) as measured by investigator observation in the phase I/II and phase II studies [24, 25]. In a small Taiwanese open-label study, nine patients showed an improvement in the (C)HAQ Disability Index with ERT [56]. However, these results could not be confirmed in the phase III study, which reported no changes in any of the tertiary efficacy measures [26]. Similarly, a 10-year Resurvey of the MPS VI Survey Study reported no change in the HAQ disability, pain and arthritis scores from baseline in patients receiving ERT for a mean period of 6.8 years despite the fact that most of the participants had rapidly progressing phenotypes [5]. A Dutch study of 11 patients with MPS VI showed improvements with ERT in lung problems, sleeping, liveliness, positive mood, social functioning, and communication domains of the TAPQOL in younger patients and in body and motor domains of the TACQOL in older patients [44]. Discussion and conclusions In recent years, more patient-centric studies have attempted to measure the burden of illness as it relates to individuals with MPS disorders. The varied clinical manifestations of MPS disorders, from skeletal, pulmonary and cardiac impairment to psychological, fatigue and pain management issues make this group of diseases unique and challenging from both the clinicians’ and patients’ perspective. The use of formal PRO tools to measure ADL and HRQoL has given researchers much insight into what patients living with a progressive, debilitating disease like MPS go through on a daily basis. While recent studies have specifically focused on PROs as clinically meaningful measures for functioning and life, PROs were previously mainly used as exploratory endpoints in clinical trials. Our review of the current literature demonstrates that HRQoL is strongly negatively affected in both MPS patients and their caregivers, with mobility, pain and psychological issues being significant problem areas. Two approved treatments for MPS are currently available, HSCT and ERT. The effects of HSCT on patient reported HRQoL has not been adequately addressed to date and future research is warranted. Multiple studies in patients with MPS I, IVA and VI receiving ERT report improvements in ADL, pain and HRQoL. Clinicians, industry, regulatory agencies and payers increasingly recognize the importance of PROs in the evaluation of new therapies. PROs should continue to be utilized in future ERT studies as primary or secondary endpoints to capture relevant data on HRQoL and there is both the opportunity and the need to validate and expand the routine use of customized MPS-specific HRQoL tools. While PRO measures often provide important information about the burden of illness in patients with MPS, they may be limited in their use, particularly in children or adolescents. HRQoL is a multi-dimensional concept that includes subjective evaluations, which can be challenging to measure due to heterogeneity in number and content of domain items included in questionnaires, discrepancies between patient and parent ratings, and lack of information regarding test–retest reliability, structural validity, or sensitivity to change [57, 58]. It is also important to keep in mind that while functional status may be related to HRQoL, functionality may not always be indicative of a patient’s subjective perception of his or her life [57, 59]. For example, children who have never experienced a healthy state and who have adapted to this condition may have a good HRQoL despite their functional limitations [57]. In addition, the impact of physical, social, and cognitive factors on HRQoL can change with age [59]. Social roles and independence may be more important for HRQoL in adolescents than for children [59]. Therefore, the terms HRQOL, health status, and functioning should not be used interchangeably, and different types of items and response formats should be used for different ages or developmental levels. Although several generic PRO tools used in MPS studies are validated and allow comparisons across diseases, disease-specific PRO measures may be more suitable as endpoints in clinical trials as they are more likely to detect clinically meaningful changes [58, 59]. The results of PRO assessments should always be seen in the socio-cultural and economic context in which they exist and should take into account the patient’s personality, cognitive ability, and community support network. In addition, ADL measures need to become broader in scope to capture technical advances of the modern world that increasingly impact, both positively and negatively, the lives of patients. Finally, it is important to keep in mind that a patient’s medical status is only a part of his or her personality. It is hoped that QoL assessments do not provoke alienation from a patient’s own personality and curtail his or her ability to flourish. To address these challenges, health professionals and patients should develop partnerships to exchange academic and life experiences. As the MPS disorders are rare diseases and patients are spread over many centers, it will be important to develop a unified approach for monitoring PROs in these patients. In closing, the impact of MPS on subjective symptoms, functionality, and HRQoL is a critical area of investigation in the field of lysosomal storage disorders. Development of valid and reliable assessment tools and implementation of routine evaluations could lead to early identification of areas of difficulty and subsequent intervention, minimizing the negative impact of MPS-related problems. Managing the negative effects of MPS through early identification and treatment will prove vital for both patients and caregivers. Ultimately, focusing on the primary medical disease alone is not enough when treating patients with a chronic illness. Considering the entire person’s biology, psychology, and social circumstances will ultimately lead to improved patient outcomes and a better understanding of the unique challenges they face. Additional files Additional file 1: Table S1. Classification of MPS. (DOCX 13 kb) Additional file 2: Patient-reported outcome (PRO) measures used in mucopolysaccharidosis (MPS) studies. (DOCX 66 kb) Abbreviations 3MSC3-minute stair climb 6MWT6-minute walk test ADLActivities of daily living APPTAdolescent pediatric pain tool ASEBAAchenbach system of empirically based assessment ASRAdult self-report BASCBehavior assessment system for children BPI-SFBrief pain inventory short form CHAQChildhood health assessment questionnaire CHQChildhood health questionnaire EQ-5DEuroqol-5 dimensions ERTEnzyme replacement therapy FIMFunctional independence measure FPS-RSix-face faces pain scale-revised FVCForced vital capacity GAGGlycosaminoglycan GHQ-60General health questionnaire 60 HRQoLHealth-related quality of life HSCTHematopoietic stem cell transplantation HS-FOCUSHunter syndrome-functional outcomes for clinical understanding scale HUIHealth utilities index MCSMental component score MorCAPMorquio A clinical assessment program MPSMucopolysaccharidoses MPS HAQMucopolysaccharidoses health assessment questionnaire MVVMaximum voluntary ventilation NCCPC-RNon-communicating children’s pain checklist-revised OASROlder adult self-report PCSPhysical component score PEDIPediatric evaluation of disability inventory PedsQLPediatric quality of life PROPatient-reported outcome QoLQuality of life rhASBRecombinant human arylsulfatase B ROMRange of motion SF-36Short form-36 (items) SIB-RScales of independent behavior, revised STAIState-trait anxiety inventory TACQOLTNO-AZL child quality of life TAPQOLTNO-AZL preschool children’s quality of life VABSVineland adaptive behavior scales VASVisual analogue scale Y-G testYatabe-Guilford personality test ZBIZarit Burden interview The authors are grateful to Ismar Healthcare NV who provided medical writing assistance on behalf of BioMarin Pharmaceuticals Ltd. Funding The writing of this manuscript was funded by BioMarin Pharmaceuticals Ltd. Availability of data and materials The datasets supporting the conclusions of this article are included within the article and its additional files. Authors’ contributions CJH, KIB, CL, SGK, PJO, RS, SL, SS and JIG made substantial contributions to the content and interpretation of data discussed in this review. SS helped to draft the manuscript. All authors critically revised the manuscript for important intellectual content and approved the final manuscript. Authors’ information CL is Vice-Director of the Center for Rare Diseases at the Clinic for Pediatric and Adolescent Medicine of the Helios Dr. Horst-Schmidt Kliniken in Wiesbaden, Germany. She completed her medical school at the Humboldt-University of Berlin, Charité and was trained as a surgeon at the Königin-Elisabeth-Herzberge Hospital in Berlin. Since 2008 she is working and following patients with rare neurometabolic diseases, especially patients with lysosomal storage disorders with focus on MPS. She is following patients clinically and is also principal investigator of several clinical trials. SGK is working as an assistant professor at the Medical University of Vienna, Austria. She is specialized in laboratory medicine and performs the selective screening for MPS and related disorders. As a clinical geneticist she counsels MPS families. Her long-standing experience with MPS started when confounding the Austrian MPS-Society 1984. She is honorary member of the Austrian and the German MPS-Society. PJO is a Professor of Pediatrics, and has extensive experience in the use of HSCT and other cellular therapies for patients with inherited, metabolic and storage diseases. A focus has been the use of combination therapies to achieve improved outcomes, and the design of clinical trials for these patient populations. JIG, PhD, is a Professor at the Keck School of Medicine, University of Southern California, in the Departments of Anesthesiology, Pediatrics, and Psychiatry & Behavioral Sciences. JIG, a licensed clinical psychologist, is director of the Pediatric Pain Management Clinic in the Department of Anesthesiology Critical Care Medicine, and director of the Children’s Outcomes, Research, and Evaluation (C.O.R.E.) program at Children’s Hospital Los Angeles. JIG completed a research fellowship at the National Center for Posttraumatic Stress in Boston and later a clinical post-doctoral fellowship in the Departments of Hematology/Oncology and Psychiatry at the UCSF Benioff Children’s Hospital Oakland. JIG has specialized and is actively engaged in the assessment, treatment, and the evaluation of integrative health (i.e., virtual reality, massage, acupuncture) therapies for reducing pain, anxiety, psychological distress and increasing comfort and satisfaction in children, adolescents, and adults with various chronic medical illnesses. Competing interests Dr. Hendriksz is a consultant for BioMarin, Shire, Genzyme; Dr. Berger is a consultant for and received travel and accommodation support from BioMarin, Genzyme, Teva, and Sarepta Vertex and received payments for lectures from BioMarin. Dr. Lampe received speaker fees, honoraria and travel support from BioMarin, Shire, Genzyme, Actelion and Alexion. Dr. Kircher has received several fees for lectures and travel costs from BioMarin Pharmaceutical Inc., from Shire Human Genetic Therapies, Inc. and Genzyme Corporation Inc. during the last 3 years; she was actively involved in discussions of working groups and the Expert Meeting about “MPS and adulthood” and contributed with her extensive experiences with Austrian MPS-patients. Dr. Orchard reports grants and personal fees from Genzyme, as well as grants from BioMarin, outside the submitted work; Dr. Southall received a speakers fee and travel support from BioMarin; Ms. Long receives ERT and has been involved in a clinical trial from BioMarin Pharmaceutical since 2012; Dr. Sande is an employee of BioMarin. Dr. Gold has received speaker fees and travel support from BioMarin. Consent for publication Not applicable. Ethics approval and consent to participate Not applicable. ==== Refs References 1. Muenzer J Overview of the mucopolysaccharidoses Rheumatology (Oxford) 2011 50 Suppl 5 v4 v12 10.1093/rheumatology/ker394 22210669 2. Neufeld EF, Muenzer J. The mucopolysaccharidoses. 2001;8:3421–3452. 3. Lavery C Hendriksz C Mortality in patients with Morquio syndrome A JIMD Rep 2015 15 59 66 24718838 4. Jones SA Almássy Z Beck M Burt K Clarke JT Giugliani R Mortality and cause of death in mucopolysaccharidosis type II - a historical review based on data from the Hunter Outcome Survey (HOS) J Inherit Metab Dis 2009 32 534 543 10.1007/s10545-009-1119-7 19597960 5. Giugliani R Lampe C Guffon N Ketteridge D Leão-Teles E Wraith JE Natural history and galsulfase treatment in mucopolysaccharidosis VI (MPS VI, Maroteaux-Lamy syndrome)-10-year follow-up of patients who previously participated in an MPS VI survey study Am J Med Genet A 2014 164A 1953 1964 10.1002/ajmg.a.36584 24764221 6. The WHOQOL Group The World Health Organization Quality of Life Assessment (WHOQOL). Development and psychometric properties Soc Sci Med 1998 46 1569 85 10.1016/S0277-9536(98)00009-4 9672396 7. Centers for Disease Control and Prevention Measuring healthy days: Population assessment of health-related quality of life 2000 Atlanta Centers for Disease Control and Prevention 8. Hendriksz CJ Lavery C Coker M Kalkan Ucar S Jain M Bell L The burden endured by caregivers of patients with morquio a syndrome: results from an international patient-reported outcomes survey JIEMS 2014 9 32 9. Hendriksz CJ Al-Jawad M Berger KI Hawley SM Lawrence R Mc Ardle C Clinical overview and treatment options for non-skeletal manifestations of mucopolysaccharidosis type IVA J Inherit Metab Dis 2013 36 309 322 10.1007/s10545-012-9459-0 22358740 10. Bergwerk KL Rabinowitz YS Falk RE Quality of life related to visual function in three young adults with mucopolysaccharidoses ScientificWorldJournal 2003 3 922 929 10.1100/tsw.2003.88 15241497 11. Harmatz P Mengel KE Giugliani R Valayannopoulos V Lin SP Parini R The Morquio A Clinical Assessment Program: baseline results illustrating progressive, multisystemic clinical impairments in Morquio A subjects Mol Genet Metab 2013 109 54 61 10.1016/j.ymgme.2013.01.021 23452954 12. Tomatsu S Montaño AM Oikawa H Rowan DJ Smith M Barrera L Mucopolysaccharidosis type IVA (Morquio A disease): clinical review and current treatment Curr Pharm Biotechnol 2011 12 931 945 10.2174/138920111795542615 21506915 13. Swiedler SJ Beck M Bajbouj M Giugliani R Schwartz I Harmatz P Threshold effect of urinary glycosaminoglycans and the walk test as indicators of disease progression in a survey of subjects with mucopolysaccharidosis VI (Maroteaux-Lamy syndrome) Am J Med Genet 2005 134A 144 150 10.1002/ajmg.a.30579 15690405 14. Montaño AM Tomatsu S Gottesman GS Smith M Orii T International Morquio A Registry: clinical manifestation and natural course of Morquio A disease J Inherit Metab Dis 2007 30 165 174 10.1007/s10545-007-0529-7 17347914 15. Brands MMG Güngör D van den Hout JMP Karstens FPJ Oussoren E Plug I Pain: a prevalent feature in patients with mucopolysaccharidosis. Results of a cross-sectional national survey J Inherit Metab Dis 2015 38 323 331 10.1007/s10545-014-9737-0 25048386 16. Aslam R van Bommel AC Hendriksz CJ Jester A Subjective and objective assessment of hand function in mucopolysaccharidosis IVa patients JIMD Rep 2013 9 59 65 10.1007/8904_2012_179 23430548 17. White K Kim T Neufeld JA Clinical assessment and treatment of carpal tunnel syndrome in the mucopolysaccharidoses J Pediatr Rehabil Med 2010 3 57 62 21791830 18. The self-care tool kit. http://www.nhs.uk/Planners/yourhealth/Documents/Self%20Care%20Toolkit%20Booklet%20-%20Oct%2010%20-%20READ.pdf. Accessed 21 Apr 2016. 19. Doward LC Gnanasakthy A Baker MG Patient reported outcomes: looking beyond the label claim Health Qual Life Outcomes 2010 8 89 10.1186/1477-7525-8-89 20727176 20. The Patient Reported Outcomes and Quality of Life Instrument database (PROQOLID). http://www.proqolid.org/. Accessed 10 July 2015. 21. The On-Line Guide to Quality-of-Life Assessment. http://www.OLGA-QoL.com. Accessed 21 Apr 2016. 22. Riazi A Patient-reported outcome measures in multiple sclerosis Int MS J 2006 13 92 99 17101077 23. Clarke LA Wraith JE Beck M Kolodny EH Pastores GM Muenzer J Long-term efficacy and safety of laronidase in the treatment of mucopolysaccharidosis I Pediatrics 2009 123 229 240 10.1542/peds.2007-3847 19117887 24. Harmatz P Whitley CB Waber L Pais R Steiner R Plecko B Enzyme replacement therapy in mucopolysaccharidosis VI (Maroteaux-Lamy syndrome) J Pediatr 2004 144 574 580 10.1016/j.jpeds.2004.03.018 15126989 25. Harmatz P Ketteridge D Giugliani R Guffon N Teles EL Miranda MC Direct comparison of measures of endurance, mobility, and joint function during enzyme-replacement therapy of mucopolysaccharidosis VI (Maroteaux-Lamy syndrome): results after 48 weeks in a phase 2 open-label clinical study of recombinant human N-acetylgalactosamine 4-sulfatase Pediatrics 2005 115 e681 e689 10.1542/peds.2004-1023 15930196 26. Harmatz P Giugliani R Schwartz I Guffon N Teles EL Sá Miranda MC Enzyme replacement therapy for mucopolysaccharidosis VI: a phase 3, randomized, double-blind, placebo-controlled, multinational study of recombinant human N-acetylgalactosamine 4-sulfatase (recombinant human arylsulfatase B or rhASB) and follow-on, open-label extension study J Pediatr 2006 148 533 539 10.1016/j.jpeds.2005.12.014 16647419 27. Raluy-Callado M Chen WH Whiteman DAH Fang J Wiklund I The impact of Hunter syndrome (mucopolysaccharidosis type II) on health-related quality of life Orphanet J Rare Dis 2013 8 101 10.1186/1750-1172-8-101 23837440 28. Hendriksz CJ Lavery C Coker M Ucar SK Jain M Bell L Burden of disease in patients with Morquio A syndrome: results from an international patient-reported outcomes survey Orphanet J Rare Dis 2014 9 32 10.1186/1750-1172-9-32 24602160 29. Harmatz P Treadwell M Burton BK Mitchell J Muschol N Jones S Pastores G Lau H Sparkes R Sutton VR Genter F Haller C Shaywitz A Impact of elosulfase alfa on pain in patients with Morquio syndrome type A Mol Genet Metab 2015 114 S51 S52 10.1016/j.ymgme.2014.12.104 30. Ali N Cagle S Psychological health in adults with morquio syndrome JIMD Rep 2015 20 87 93 10.1007/8904_2014_396 25614312 31. Kuratsubo I Suzuki Y Orii KO Kato T Orii T Kondo N Psychological status of patients with mucopolysaccharidosis type II and their parents Pediatr Int 2009 51 41 47 10.1111/j.1442-200X.2008.02652.x 19371276 32. Hendriksz CJ Giugliani R Harmatz P Mengel E Guffon N Valayannopoulos V Multi-domain impact of elosufase alfa in Morquio A syndrome in the pivotal phase III trial Mol Genet Metab 2015 114 178 185 10.1016/j.ymgme.2014.08.012 25284089 33. Wraith JE Clarke LA Beck M Kolodny EH Pastores GM Muenzer J Enzyme replacement therapy for mucopolysaccharidosis I: a randomized, double-blinded, placebo-controlled, multinational study of recombinant human a-L-iduronidase (laronidase) J Pediatr 2004 144 581 588 10.1016/j.jpeds.2004.01.046 15126990 34. Kato T Kato Z Kuratsubo I Ota T Orii T Kondo N Evaluation of ADL in patients with Hunter disease using FIM score Brain Dev 2007 29 298 305 10.1016/j.braindev.2006.08.015 17307320 35. Marucha J Jurecka A Syczewska M Rózdzynska-Swiatkowska A Tylki-Szymanska A Restricted joint range of motion in patients with MPS II: correlation with height, age and functional status Acta Paediatr 2012 101 e183 e188 10.1111/j.1651-2227.2011.02522.x 22077147 36. Needham M Packman W Rappoport M Quinn N Cordova M Macias S MPS II: adaptive behavior of patients and impact on the family system J Genet Couns 2014 23 330 338 10.1007/s10897-013-9665-4 24190099 37. Shapiro EG Rudser K Ahmed A Steiner RD Delaney KA Yund B A longitudinal study of emotional adjustment, quality of life and adaptive function in attenuated MPS II Mol Genet Metab Rep 2016 7 32 39 10.1016/j.ymgmr.2016.03.005 27114914 38. Holt JB Poe MD Escolar ML Natural progression of neurological disease in mucopolysaccharidosis type II Pediatrics 2011 127 e1258 1265 10.1542/peds.2010-1274 21518713 39. Kunin-Batson AS Shapiro EG Rudser KD Lavery CA Bjoraker KJ Jones SA Long-term cognitive and functional outcomes in children with mucopolysaccharidosis (MPS)-IH (hurler syndrome) treated with hematopoietic cell transplantation JIMD Rep 2016 40. Bjoraker KJ Delaney K Peters C Krivit W Shapiro EG Long-term outcomes of adaptive functions for children with mucopolysaccharidosis I (Hurler syndrome) treated with hematopoietic stem cell transplantation J Dev Behav Pediatr 2006 27 290 296 10.1097/00004703-200608000-00002 16906003 41. Lampe C Jain M Olaye A Meesen B Decker C Mengel E Relationship between patient-reported outcomes and clinical outcomes in patients with Morquio A syndrome J Inborn Error Metab Screen 2015 3 1 8 10.1177/2326409815576188 42. Lavery C, Wedehase B, Graham S, Harmatz P, Hendriksz CJ. Impact of Mucopolysaccharidosis on Daily living, Employment, General Health and Parenthood of Adult Patients. Poster presented at WORLDSymposium 2015; February 9–13, 2015; Orlando, FL. 43. Needham M Packman W Quinn N Rappoport M Aoki C Bostrom A Health-related quality of life in patients with MPS II J Genet Couns 2015 24 635 644 10.1007/s10897-014-9791-7 25395377 44. Brands MMMG Oussoren E Ruijter GJG Vollebregt AAM van den Hout HMP Joosten KFM Up to five years experience with 11 mucopolysaccharidosis type VI patients Mol Genet Metab 2013 109 70 76 10.1016/j.ymgme.2013.02.013 23523338 45. Hendriksz CJ Burton B Fleming TR Harmatz P Hughes D Jones SA Efficacy and safety of enzyme replacement therapy with BMN 110 (elosulfase alfa) for Morquio A syndrome (mucopolysaccharidosis IVA): a phase 3 randomised placebo-controlled study J Inherit Metab Dis 2014 37 979 990 10.1007/s10545-014-9715-6 24810369 46. Valayannopoulos V Wijburg FA Therapy for the mucopolysaccharidoses Rheumatology (Oxford) 2011 50 Suppl 5 v49 v59 10.1093/rheumatology/ker396 22210671 47. Decker C Yu Z Giugliani R Schwartz IVD Guffon N Teles EL Enzyme replacement therapy for mucopolysaccharidosis VI: growth and pubertal development in patients treated with recombinant human N -acetylgalactosamine 4-sulfatase J Pediatr Rehabil Med 2010 3 89 100 20634905 48. Braunlin E Rosenfeld H Kampmann C Johnson J Beck M Giugliani R Enzyme replacement therapy for mucopolysaccharidosis VI: long-term cardiac effects of galsulfase (Naglazyme®) therapy J Inherit Metab Dis 2013 36 385 394 10.1007/s10545-012-9481-2 22669363 49. Jones SA Bialer M Parini R Martin K Wang H Yang K Safety and clinical activity of elosulfase alfa in pediatric patients with Morquio A syndrome (mucopolysaccharidosis IVA) less than 5 years Pediatr Res 2015 78 717 22 10.1038/pr.2015.169 26331768 50. Beck M Arn P Giugliani R Muenzer J Okuyama T Taylor J The natural history of MPS I: global perspectives from the MPS I Registry Genet Med. 2014 16 759 65 10.1038/gim.2014.25 24675674 51. Tanaka A Okuyama T Suzuki Y Sakai N Takakura H Sawada T Long-term efficacy of hematopoietic stem cell transplantation on brain involvement in patients with mucopolysaccharidosis type II: a nationwide survey in Japan Mol Genet Metab 2012 107 513 520 10.1016/j.ymgme.2012.09.004 23022072 52. Sifuentes M Doroshow R Hoft R Mason G Walot I Diament M A follow-up study of MPS I patients treated with laronidase enzyme replacement therapy for 6 years Mol Genet Metab 2007 90 171 180 10.1016/j.ymgme.2006.08.007 17011223 53. Tylki-Szymanska A Marucha J Jurecka A Syczewska M Czartoryska B Efficacy of recombinant human a-L-iduronidase (laronidase) on restricted range of motion of upper extremities in mucopolysaccharidosis type I patients J Inherit Metab Dis 2010 33 151 157 10.1007/s10545-010-9059-9 20217237 54. Kosinski M Zhao SZ Dedhiya S Osterhaus JT Ware JE Determining minimally important changes in generic and disease-specific health-related quality of life questionnaires in clinical trials of rheumatoid arthritis Arthritis Rheum 2000 43 1478 1487 10.1002/1529-0131(200007)43:7<1478::AID-ANR10>3.0.CO;2-M 10902749 55. Muenzer J Beck M Eng CM Giugliani R Harmatz P Martin R Long-term, open-labeled extension study of idursulfase in the treatment of Hunter syndrome Genet Med 2011 13 95 101 10.1097/GIM.0b013e3181fea459 21150784 56. Lin HY Chen MR Chuang CK Chen CP Lin DS Chien YH Enzyme replacement therapy for mucopolysaccharidosis VI-experience in Taiwan J Inherit Metab Dis 2010 33 Suppl 3 S421 S427 10.1007/s10545-010-9212-5 20924685 57. Davis E Waters E Mackinnon A Reddihough D Graham HK Mehmet-Radji O Paediatric quality of life instruments: a review of the impact of the conceptual framework on outcomes Dev Med Child Neurol 2006 48 311 318 10.1017/S0012162206000673 16542522 58. Solans M Pane S Estrada MD Serra-Sutton V Berra S Herdman M Health-related quality of life measurement in children and adolescents: a systematic review of generic and disease-specific instruments Value Health 2008 11 742 764 10.1111/j.1524-4733.2007.00293.x 18179668 59. Palermo TM Long AC Lewandowski AS Drotar D Quittner AL Walker LS Evidence-based assessment of health-related quality of life and functional impairment in pediatric psychology J Pediatr Psychol 2008 33 983 996 10.1093/jpepsy/jsn038 18430762 60. Burton BK Berger KI Lewis GD Tarnopolsky M Treadwell M Mitchell JJ Safety and physiological effects of two different doses of elosulfase alfa in patients with morquio a syndrome: a randomized, double-blind, pilot study Am J Med Genet A 2015 167A 2272 81 10.1002/ajmg.a.37172 26069231 61. Jacob E Mack AK Savedra M Van Cleve L Wilkie DJ Adolescent pediatric pain tool for multidimensional measurement of pain in children and adolescents Pain Manag Nurs 2014 15 694 706 10.1016/j.pmn.2013.03.002 23870767 62. The Brief Pain Inventory. MD Anderson Cancer Society. http://www.mdanderson.org/education-and-research/departments-programs-and-labs/departments-and-divisions/symptom-research/symptom-assessment-tools/brief-pain-inventory.html. Accessed 21 Apr 2016. 63. Breau LM McGrath PJ Camfield CS Finley GA Psychometric properties of the non-communicating children’s pain checklist-revised Pain 2002 99 349 357 10.1016/S0304-3959(02)00179-3 12237214 64. Ramey DR Raynauld JP Fries JF The health assessment questionnaire 1992: status and review Arthritis Care Res 1992 5 9 29 10.1002/art.1790050303 65. Singh G Athreya BH Fries JF Goldsmith DP Measurement of health status in children with juvenile rheumatoid arthritis Arthritis Rheum 1994 37 761 9 10.1002/art.1780371209 66. Guarany NR Schwartz IVD Guarany FC Giugliani R Functional capacity evaluation of patients with mucopolysaccharidosis J Pediatr Rehabil Med 2012 5 37 46 22543891 67. Beck M Muenzer J Scarpa M Evaluation of disease severity in mucopolysaccharidoses J Pediatr Rehabil Med 2010 3 39 46 21791828 68. Scales of Independent Behavior-Revised. http://www.hmhco.com/hmh-assessments/other-clinical-assessments/sib-r#sthash.hZuzuTRo.dpuf. Accessed May 2016. 69. EuroQol Group. EQ-5D-5L User Guide. http://www.euroqol.org/about-eq-5d/publications/user-guide.html. Accessed 21 Apr 2016 & http://www.euroqol.org/about-eq-5d/valuation-of-eq-5d.html. Accessed 21 Apr 2016. 70. SF-63.org. http://www.sf-36.org/tools/sf36.shtml. Accessed 21 Apr 2016. 71. Varni JW Burwinkle TM The PedsQL™ as a patient-reported outcome in children and adolescents with attention-deficit/hyperactivity disorder: a population-based study Health Qual Life Outcomes 2006 4 26 10.1186/1477-7525-4-26 16630344 72. Bunge EM Essink-Bot ML Kobussen MPHM van Suijlekom-Smit LWA Moll HA Raat H Reliability and validity of health status measurement by the TAPQOL Arch Dis Child 2005 90 351 358 10.1136/adc.2003.048645 15781921 73. Verrips EGH Vogels TGC Koopman HM Theunissen NCM Kamphuis RP Fekkes M Measuring health-related quality of life in a child population Eur J Public Health 1999 9 188 193 10.1093/eurpub/9.3.188 74. Varni JW Sherman SA Burwinkle TM Dickinson PE Dixon P The PedsQL™ family impact module: preliminary reliability and validity Health Qual Life Outcomes 2004 2 55 10.1186/1477-7525-2-55 15450120 75. Orme M Kerrigan J Tyas D Russell N Nixon R The effect of disease, functional status, and relapses on the utility of people with multiple sclerosis in the UK Value Health 2007 10 54 60 10.1111/j.1524-4733.2006.00144.x 17261116 76. Kobelt G Lindgren P Lindroth Y Jacobson L Eberhardt K Modelling the effect of function and disease activity on costs and quality of life in rheumatoid arthritis Rheumatology 2005 44 1169 1175 10.1093/rheumatology/keh703 15956093 77. Kobelt G, Lekander I,Santesson Nicolae Y. Access to innovative treatments for rheumatoid arthritis in New Zealand. A comparison with Australia and the UK. European Health Economics. 2010;1–78. www.arthritis.org.nz/wp-content/uploads/2014/05/ACCESS-TO-INNOVATIVE-TREATMENTS-RA-FINAL.pdf. Accessed 23 Aug 2016. 78. Dolan P Modeling valuations for EuroQol health states Med Care 1997 35 1095 1108 10.1097/00005650-199711000-00002 9366889

PMC005xxxxxx/PMC5000419.txt

==== Front J NeuroinflammationJ NeuroinflammationJournal of Neuroinflammation1742-2094BioMed Central London 67910.1186/s12974-016-0679-3ResearchRed LED photobiomodulation reduces pain hypersensitivity and improves sensorimotor function following mild T10 hemicontusion spinal cord injury Hu Di di.hu@anu.edu.au 1Zhu Shuyu iriszhu062@gmail.com 1Potas Jason Robert jason.potas@anu.edu.auj.potas@unsw.edu.au 121 The John Curtin School of Medical Research, The Australian National University, Building 131, Garran Rd, Canberra, ACT 2601 Australia 2 ANU Medical School, The Australian National University, Canberra, ACT 2601 Australia 26 8 2016 26 8 2016 2016 13 1 20017 4 2016 17 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background The development of hypersensitivity following spinal cord injury can result in incurable persistent neuropathic pain. Our objective was to examine the effect of red light therapy on the development of hypersensitivity and sensorimotor function, as well as on microglia/macrophage subpopulations following spinal cord injury. Methods Wistar rats were treated (or sham treated) daily for 30 min with an LED red (670 nm) light source (35 mW/cm2), transcutaneously applied to the dorsal surface, following a mild T10 hemicontusion injury (or sham injury). The development of hypersensitivity was assessed and sensorimotor function established using locomotor recovery and electrophysiology of dorsal column pathways. Immunohistochemistry and TUNEL were performed to examine cellular changes in the spinal cord. Results We demonstrate that red light penetrates through the entire rat spinal cord and significantly reduces signs of hypersensitivity following a mild T10 hemicontusion spinal cord injury. This is accompanied with improved dorsal column pathway functional integrity and locomotor recovery. The functional improvements were preceded by a significant reduction of dying (TUNEL+) cells and activated microglia/macrophages (ED1+) in the spinal cord. The remaining activated microglia/macrophages were predominantly of the anti-inflammatory/wound-healing subpopulation (Arginase1+ED1+) which were expressed early, and up to sevenfold greater than that found in sham-treated animals. Conclusions These findings demonstrate that a simple yet inexpensive treatment regime of red light reduces the development of hypersensitivity along with sensorimotor improvements following spinal cord injury and may therefore offer new hope for a currently treatment-resistant pain condition. Keywords PhotobiomodulationLight therapyM2 macrophage polarizationAllodyniaNeuropathic pain670 nmGretel and Gordon Bootes Medical Research Foundationissue-copyright-statement© The Author(s) 2016 ==== Body Background The experience of pain serves as an essential survival mechanism that motivates us to protect ourselves from harm; however, following spinal cord injury, the development of treatment-resistant neuropathic pain often ensues, bringing no advantage to the sufferer but severely reducing the quality of life. Chronic pain affects a vast sector of the population for which the socioeconomic cost exceeds that of heart disease, cancer and diabetes [1]; thus, successfully treating neuropathic pain would bring significant benefits. The non-invasive application of light, at wavelengths that penetrate transcutaneously [2], has begun to emerge as a potential therapy for improving functional outcomes from a variety of neural injuries [3]. Photobiomodulation with wavelengths ranging from 630 to 1100 nm has demonstrated positive effects in animal models of neurodegenerative diseases such as Alzheimer’s [4] and Parkinson’s [5], genetic models of dementia [6], as well as acute nervous injuries to the retina [7–9], optic nerve [9, 10], sciatic nerve [11–15] and spinal cord [16]. In humans, photobiomodulation has been reported to be effective against a variety of pain conditions including mucositis [17], carpel tunnel syndrome [18–20], orthodontic pain [21], temporomandibular joint pain [22], neck pain [23] and neuropathic pain resulting from amputation [24]. Inflammatory mediators have long been implicated in the development and maintenance of pain [25–28]. These chemical mediators are controlled by a variety of immune cells including the balance of pro- and anti-inflammatory microglia/macrophage subpopulations [29–35]. As in non-neural tissues, macrophages can be activated by T helper cell type 1 (Th1) or type 2 (Th2) to generate opposing immune responses following spinal cord injury [30, 31]. Th1-activated microglia/macrophages (M1) have been considered potentially damaging to healthy tissues, as they induce a pro-inflammatory response and have been shown to inhibit axonal regeneration [30]. Conversely, Th2-activated microglia/macrophages (M2) have been considered protective, as they have a role in suppressing the pro-inflammatory response by producing anti-inflammatory mediators [30, 31]. Following spinal cord injury, there is evidence suggesting that the M1 response prevails over a more transient M2 response, and this observation has been proposed to contribute to the poor regenerative capacity of the spinal cord following injury [30, 31]. Consistent among various in vitro and in vivo studies, including spinal cord and peripheral nerve injury models, are reports of reduced levels of pro-inflammatory cell mediators, including as IL-6, iNOS, MCP-1, IL-1β and TNFα in response to treatment with various wavelengths including 633 nm [36], 660 nm, 780 nm [37], 810 nm [16] and 950 nm [14]. Coincidently, these pro-inflammatory cell mediators are secreted by M1 cells; thus, we were curious to examine the effect of light treatment on microglia/macrophage populations. There are various wavelengths used throughout the literature which demonstrate biological effects. In an attempt to find the better wavelength option for treating nervous system injuries, one study compared the effects of two wavelengths in a variety of CNS injury models, to find that 670 nm treatment resulted in better outcomes for a number of parameters when compared to 830 nm [9]. Our aim therefore was to evaluate the effect of the 670 nm wavelength following spinal cord injury on a variety of functional parameters, namely the development of hypersensitivity to innocuous stimuli (allodynia), as well as on (tactile) sensory pathway conduction and locomotor recovery, and to see if there were alterations to the M1/M2 sub-populations. We found that red light treatment significantly reduced the severity of hypersensitivity while improving sensorimotor function and that these improvements were preceded by an anti-inflammatory microglia/macrophage cell population in the injury zone. Methods Hemicontusion spinal cord injury All animal work was approved by the ANU Animal Experimentation Ethics Committee. Hemicontusion spinal cord injuries were performed on 7-week-old Wistar rats under isoflurane (1.7–2.3 % v/v) anaesthesia. Following hair removal, a laminectomy of T10 vertebral body and removal of dura and arachnoid was performed, followed by a spinal cord hemicontusion using a customized impactor system [38] comprising of a cylindrical 10 g weight with a 1-mm diameter tip that was guided onto the right dorsal horn and dropped from 25 to 50 mm above the spinal cord. Treatment and experimental groups Injured animals were divided into 670-nm-treated (SCI+670) and sham-treated (SCI) groups. SCI+670 rats received 30 min of 670 nm irradiation commencing 2 h after surgery and then every 24 h after locomotor assessment for the remainder of the recovery period. A commercially available 670 nm LED array (WARP 75A, Quantum Devices, Barneveld, WI; 75 mm2 treatment area) was used for treatment. Spectral characteristics and power output (Fig. 1) of the LED were measured using a spectrometer (CCS175, Thorlabs) and custom made power meter that was calibrated against a commercially available power meter (PM100D, ThorLabs). Treatment was delivered through a transparent treatment box which was used to confine the animal within its home cage. This resulted in a 7-mm distance between the dorsal surface of the animal and the LED array and delivery of 35 mW/cm2 (fluence = 63 J/cm2) of 670 nm at the contact surface of the animal’s dorsum. SCI rats (n = 29) were restrained in the identical way as the SCI+670 group (n = 29), but without the LED device switched on to control for 30 min restraint in the transparent treatment box. Three additional control groups were included: an intact uninjured group (control; n = 7) was untreated and did not receive any sham operations or sham treatment; a sham-injured group (shamSCI; n = 8) underwent the spinal surgery, but without the contusion, and was subjected to sham treatment; a sham-injured 670-nm-treated group (shamSCI+670; n = 10) underwent spinal surgery without the contusion and received daily 30 min treatments.Fig. 1 Externally applied red light penetrates through the entire rat spinal cord. a Photograph shows the ventral surface of the spinal cord following removal of the T10 vertebral body in a cadaver rat. Topography of the vertebral column is shown centred around the 10th vertebral body under normal light conditions. b The identical region as shown in a, with a 670 nm LED array light source (35 mW/cm2) placed directly on the dorsum of the animal and with ambient lights switched off. Note the visible red light illuminating from the ventral surface of the cord (exposed, arrow) indicating excess penetration through dorsal layers of the hair, skin, muscle, bone and spinal cord. c Intensities measured by a 670 nm power meter are shown for six freshly sacrificed cadaver rats (each dot represents the mean of triplicate readings). Readings shown are taken at the light source (through the Perspex restraining box, intensity at dorsal surface) and at the ventral surface of the spinal cord as shown by the white arrow in b (intensity at ventral surface). Black arrow indicates proportion of light absorbed and/or scattered by intervening tissues. d Spectral analysis of the light source indicating central frequency of 675 nm Light penetration Uninjured, unshaven animals (n = 6) were euthanised with sodium pentobarbital solution (325 mg/ml; Virbac; dosage, 100 mg/kg). The overlaying heart, great vessels and muscles were detached from the anchoring connective tissues and retracted to the side to expose the underlying vertebral column. The T10 vertebral body was eroded with a dental drill to expose the spinal cord from the ventral surface. The cadaver was placed on its back in an inverted transparent treatment box so that the dorsum of the cadaver could be positioned over the 670 nm LED array and the ventral surface of the rat was accessible to enable placement of a custom-built light measuring device. This device comprised of a photodiode chip (surface area, 0.62 mm2; maximal response (>95 %) to 630–685 nm; ODD-660W, Opto Diode Corp.) that was fixed to the bottom of an aluminium cylinder (height, 7.0 mm; external diameter, 8.7 mm). The top of the cylinder was sealed with a glass coverslip, and the entire probe was painted with black paint but leaving a small circular window (4.0 mm diameter) centred over the chip sensor. This left a ~2.4-mm lip between the external edge of the glass window and the external circumference of the cylinder. When pressed onto the ventral surface of the spinal cord, no light could penetrate from the side because the chip was located 7.0 mm behind the 4.0-mm aperture; thus, only light rays between 71° and 90° are able to reach the surface of the sensor; angles deviating from 90° do not hit the entire surface of the photosensitive diode and therefore contribute less to the total power reading. The signal from the probe was amplified by a custom amplifier built for purpose. The key component was the logarithmic converter amplifier (AD8304, Analog Devices). The readings were then calibrated against a commercially available light power meter tuned at 670 nm (PM100D ThorLabs) by producing a calibration table for different radiant power (controlled by distance from the light source) and subsequently converted into intensity (power/unit area). The probe was used to determine light intensity from the 670 nm array through (i) the treatment box, (ii) the spinal cord and dorsal overlying structures and (iii) the equivalent space through the air to provide a measure of attenuation over the distance of the light path. Prior to activating the LED, ambient lights were switched off; however, we also confirmed that no photons were detected by the light meter with the ambient lights on. Three repeat readings were acquired for each measurement. Example images were obtained with a D1X Nikon (5.3 megapixels) camera and 120-mm lens (Medical NIKKOR) with a ×2 adaptor and built in ring flash. Images were captured with both the ambient lights and LED array on and then repeated in the same position with the ambient lights off. Temperature measurement A temperature probe (ML309/MLT422, ADInstruments) connected to a data acquisition system (PowerLab 26T, LabChart v7.3.7, ADInstruments) was attached to the dorsum of the animals prior to, and 2 min after sham or light treatment on consecutive days from four sham- and four light-treated rats. Sensitivity assessment Sensitivity assessment was carried out on day 7 post-injury prior to locomotor and electrophysiological assessments. To assess hypersensitivity, a nylon filament (OD: 1.22 mm) was used to deliver innocuous tactile stimuli over six defined regions over the animals’ dorsum: Above-Level (dermatomes C6-T3), At-Level (dermatomes T9-T12) and Below-Level (dermatomes L2-L5) on ipsi- and contralateral sides relative to the injury. The boundary for each of the six regions was marked on the animals’ back, and 10 consecutive innocuous “pokes” were delivered in each boundary at an inter-poke-interval of approximately 1–2 s, or until the animal recovered from movement evoked from the previous poke if longer than 2 s. Prior to testing, the operator practiced the stimulus procedure. This ensured that each poke was as brief as possible, that the filament landed normal (90°) to the skin surface and that the final position of the filament handle was approximately half the distance to that of the distance at initial contact of the filament. This protocol ensured pokes of consistent duration and maximum force which was confirmed using a weighing balance (maximum bending force: 2.86 ± 0.09 g; n = 10 pokes). During sensitivity testing, animals were “semi-restrained” in a V-shaped plastic box. This restricted the animal’s ability to avoid the testing procedure and thereby facilitated the operator’s accuracy of each poke but enabled sufficient movement for the animal to display behavioural responses of interest. Testing was recorded using a webcam (Logitec HD Pro C920). Videos were assessed blind to the observer in slow motion play back by evaluating the response to each innocuous poke that was graded into one of four categories as (I) no response; (II) mild response characterised by acknowledgment of the stimulus, head turns, brief shuddering of the contacted skin, but no obvious pain avoidance behaviours; (III) medium response, characterised by moderate signs of pain perception, including moderate avoidance attempts by moving away from the stimulus and (IV) severe response, characterised by severe signs of pain perception, including attacking the stimulus and “desperate” avoidance attempts and escape behaviours including jumping, running, writhing or audible vocalization. The four categories, I–IV, were chosen because these behaviours are easily distinguishable. The frequency of each response category was multiplied by a weight; categories I–IV were multiplied by 0, 1, √2 and 2, respectively, to provide greater separation between ordinal pain behaviours between non-painful and painful [39], as well as to help minimise heteroscedasticity of the data. The sum of the 10 weighted responses provided a regional sensitivity score (RSS) for each region. This paradigm enables high-resolution measures of sensitivity to 10 innocuous pokes with each possible RSS ranging between 0 and 20. Scores from ipsi- and contralateral regions were pooled to determine level sensitivity scores (LSS) above, at and below the level of injury. An cumulative sensitivity score (CSS) was derived for each animal by summing the RSS from all six regions; the maximum CSS possible is therefore 120. The hypersensitivity threshold was defined by the mean + 2 standard deviations (confidence interval of 95.5 %) of CSSs calculated from uninjured intact rats (control group). Somatosensory assessment Animals were anaesthetised with urethane (12.5 % w/v; 1.4 g/kg; i.p.) and maintained at 37 °C on a heating mat. A tracheotomy was performed, and animals were placed in a stereotaxic frame. The gracile nuclei were exposed through the foramen magnum by head flection and removal of overlying muscles and meninges. Both left and right sciatic and sural nerves were exposed by the removal of the overlying skin followed by a splitting incision of the gluteus maximum and semimembranosus muscles, respectively. The exposed nerves were isolated from adjacent connective tissues and bathed in paraffin oil. Silver wire bipolar hook electrodes were used to stimulate sural nerves, and a single hook silver wire electrode was used to record from sciatic nerves to ensure complete recruitment of all sural nerve fibres upon electrical stimulation (square wave pulse, 0.5–1.1 mA, 0.05 ms). A platinum wire electrode was used to record from a single midline position on the brainstem at a location that was established to provide evoked potentials of equal magnitude and latency from left and right sural nerve stimulation. Thirty-three individual evoked potentials were recorded and averaged from the sciatic nerve and the brainstem in response to repeated sural nerve stimulations. Signals recorded from the brainstem were then processed offline (MATLAB, MathWorks). The averaged signal was band-pass filtered (500–3350 Hz) and response magnitudes calculated from the integral of rectified signals (integral limits: 5.00 ms before and 8.75 ms after the primary peak) after subtraction from baseline levels obtained prior to the stimulus. Latency was measured from the filtered signal where it first exceeded 3 standard deviations (confidence interval 99.7 %) of background levels. Locomotor assessment Prior to surgery, animals were trained to run along an 80-cm custom build transparent walking-track with mirrors that reflected left and right sides and underneath of the animal. This enabled exquisite locomotor detail from all sides of interest to be video captured simultaneously from a single viewpoint. 2 h following surgery, initial recordings of animals running three consecutive times down the walking-track were acquired with a digital camera (Sony, NEX-VG20EH) at 50 frames per second, which provided adequate data for detailed gait analysis. Recordings were repeated every 24 h post-surgery for 7 consecutive days. Each video file was coded and assessed blind by one assessor. The BBB locomotor scale [40] for the left and right hind-limbs was used to generate locomotor scores from video files assessed in slow motion. Immunohistochemistry and TUNEL Animals from both groups (SCI, n = 15; SCI+670, n = 15) were divided into three recovery time points and sacrificed at 1, 3 and 7 days post-injury. At the end of designated recovery periods, animals were transcardially perfused with saline and 4 % buffered paraformaldehyde (w/v). Harvested spinal cords were cryoprotected in 30 % sucrose (w/v), cryosectioned at 20 μm in the longitudinal plane using a Leica CM1850 cryostat, and dorsal sections labelled with primary antibodies (1:200) against rat CD68 (ED-1 clone, MAB1435, Millipore), and Arginase-1 (AB60176, Abcam) or CD80 (AB53003, Abcam) to quantify microglia/macrophages (ED1+) and polarized subtypes M1 (CD80+ED1+) and M2 (Arginase1+ED1+), respectively. Tissue was subsequently incubated with the appropriate secondary antibodies (1:1000, Invitrogen, Alexa 594 conjugated chicken anti-goat #A21468, Alexa 488 conjugated goat anti-mouse #A31619, Alexa 594 conjugated goat anti-mouse #A31623, Alexa 488 conjugated donkey anti-rabbit #A21206). Slides were then incubated in Hoechst solution (2 μg/ml Sigma-Aldrich). Standard immunohistochemical controls were included. To detect cells undergoing apoptosis/necrosis, a TUNEL assay was performed. Slides were incubated with 1:10 Terminal Deoxynucleotidyl Transferase (TdT) buffer (125 mM Tris-HCl, 1 M sodium cacodylate, 1.25 mg/ml BSA, pH 6.6) for 10 min and then 1-h incubation at 37 °C with reaction mixture [0.5 enzyme unit/μl TdT (Roche Applied Science) and 2.52 μM Biotin-16-dUTP (Roche Applied Science) diluted in 1:10 TdT buffer]. This was followed by 15 min incubation in 1:10 saline sodium citrate (SSC) buffer (175.3 mg/ml sodium chloride, 88.2 mg/ml sodium citrate, pH 7.0) and blocked with 10 % normal goat serum in 0.1 M PBS for 10 min before incubating with secondary antibody in 0.1 M PBS (1:1000 dilution, Invitrogen, Alexa 488 conjugated streptavidin S11223) at 37 °C for 30 min. All image analysis was performed blind to the experimental group. 2D images were constructed from three colour channel (red, green and blue) images acquired from a LED fluorescent microscope (Carl Zeiss Colibri) with a ×20 objective and digital camera (AxioCam MRc 5) with all settings kept constant for each channel. Cells with co-labelling were quantified with ImageJ (v1.46r) using the Cell Counter plugin that enables the placement of different classes of markers onto an image. Cytoplasmic markers, a class for each channel, were used to tag positive label in a single focal plane for all green and red channels that were examined independently. To define ED1+ cells, the accompanying DAPI+ nucleus (blue channel) was tagged for cells where ED1 staining was clearly complementing the DAPI surface profile. Double-labelled cells (i.e., ED1+Arginase1+ or CD80+) were evaluated by scrutinising all tagged DAPI+ cells for co-labelling in red and green channels. These cells were tagged again with another marker class. All markers were automatically quantified for each class by the software. Cells out of focus were not included. Cell counts were obtained from dorsal horn regions with viable tissue and quantified as the mean of duplicate images, each covering a minimum area 0.05 mm2. The areas of interest were defined and quantified prior to cell quantification and included the dorsal horn grey matter region as well as the white matter in the surrounding dorsal columns and lateral funiculus. Cell quantification is expressed as the number of cells per unit area (mm2). Statistics All data expressed as boxplots with individual data points in figures or as mean ± SEM in the main text, unless otherwise stated. Boxplots indicate the median (thicker line), upper and lower quartiles with whiskers extending to maximum and minimum values excluding outliers (more than 1.5 times respective quartiles). Statistical analysis was carried out using R or MATLAB, and a criterion alpha level of 0.05 was adopted as statistically significant. Data sets were tested for normality and homoscedasticity, and t tests and linear mixed models (multi-factor ANOVA) were applied for normally distributed data (indicated by *) or Wilcoxon rank-sum (indicated by †) where data was not normally distributed. Results Red light penetrates the spinal cord We first set out to demonstrate that red light can pass through superficial and deep structures underlying the dorsal exterior surface and penetrate the entire spinal cord (Fig. 1). The penetrating light could be seen with the naked eye (example, Fig. 1a, b). The dorsal surface of uninjured rats (n = 6) was exposed to the LED array and 670 nm light intensity measured at the light source surface through the transparent treatment box which directly contacts the rat dorsum during treatment (Fig. 1c, intensity at dorsal surface; 35.4 ± 0.05 mW/cm2) and the ventral surface of the spinal cord, where light had to pass through an additional ~10 mm of the animals’ tissues from dorsal surface (Fig. 1c, intensity at ventral surface; 3.2 ± 0.6 mW/cm2). These data show that 91.1 ± 1.8 % of the light from the LED array was absorbed/dispersed by the tissues between the dorsal surface of the animal and the ventral surface of the spinal cord (Fig. 1c, black arrow). To indicate the approximate attenuation over the distance of light travelling from the light source through to the ventral spinal cord surface, we measured the intensity at the approximate distance (10 mm) through the air (33.0 ± 0.5 mW/cm2). This demonstrated that the expected attenuation (~7 %) of light is negligible over the distance required to travel to the ventral surface of the cord. Surface temperature changes following light treatment We measured the surface temperature of rats directly before and 2 min after treatment. Twenty-seven readings from sham-treated and 25 readings from light-treated animals were acquired from four animals in each group over consecutive days of treatment. While there was no significant difference in the surface temperature of sham-treated animals (before, 33.6 ± 0.23 °C; after, 33.6 ± 0.25 °C), there was a small but significant increase 2 min after light treatment (before, 32.8 ± 0.36 °C; after, 33.5 ± 0.22 °C; p = 0.038, paired t test). Red light reduces allodynia following spinal cord injury To examine the effect of red light on the development of neuropathic pain, we assessed sensitivity on six regions over the rat dorsum using a T10 hemicontusion spinal cord injury model that results in clear development of hypersensitivity in most animals within 7 days. The T10 spinal hemicontusion resulted in 63 % of animals (n = 12) developing hypersensitivity in both sham-treated (SCI, n = 19) and light-treated (SCI+670, n = 19) groups at 7 days post-injury. The hypersensitive subpopulation of rats from the SCI group had a mean CSS (SCI, CSS: 25.3 ± 4.5) that was 3.7 × the hypersensitive threshold (Fig. 2a). The mean CSS was significantly reduced by 40 % (SCI+670, CSS: 14.5 ± 1.6; 2.1 × the hypersensitivity threshold) in the hypersensitive subpopulation of rats from the SCI+670 group. Light treatment significantly reduced At- (T9-T12 dermatomes) and Below- (L2-L5 dermatomes) LSSs, which arose from contralateral At-Level and both ipsi-and contralateral Below-Level regions (Fig. 2b). Compared to the uninjured control group (control, Fig. 2c), sham injury without light treatment (shamSCI, n = 8) had no significant effect on LSS or RSS despite two sham-injured animals developing At-Level hypersensitivity. Light treatment of sham-injured animals (shamSCI+670, n = 10) resulted in significant reductions of At- and Below-LSS compared to the shamSCI group (Fig. 2c). Thus, while the incidence of hypersensitivity was not altered by red light, the level of hypersensitivity was markedly reduced At- and Below-levels in T10 contused light-treated allodynic animals. Red light also caused a significant reduction in sensitivity in 670-treated sham-injured animals (shamSCI+670, CSS: 0.8 ± 0.5) compared to uninjured control animals (control, CSS: 2.8 ± 0.8) as well as normosensitive spinal cord injured animals (SCI, CSS: 3.5 ± 0.9), even though these animals were not hypersensitive.Fig. 2 Hypersensitivity is reduced by red light treatment at 7 days post-T10 hemicontusion spinal cord injury. a CSSs (see the “Methods” section) for all groups are separated by the hypersensitivity threshold (6.9; indicated by dotted green line) into normosensitive (CSS < hypersensitivity threshold) and hypersensitive (CSSs > hypersensitivity threshold) subpopulations. b RSSs in hypersensitive sham-treated (SCI, dark blue) and 670-nm-treated (SCI+670, dark red) spinal cord injured animals (location of injury indicated). RSSs are represented as the mean ± SEM (colour-coded according to the insert: mean + SEM, mean, and mean − SEM concentrically represented) for the six tested regions (left and right sides; “Above-Level”, “At-Level” and “Below-Level” relative to the injury). RSSs are overlayed on schematic representations of the rat dorsum, with C2, T1, L1 and S2 dermatomes, and the midline, indicated (grey). Individual RSSs and LSSs are compared between hypersensitive subpopulation of the two groups. c RSSs shown for normal uninjured rats (control, green), sham-injury + sham-treatment (shamSCI, light blue, data includes both normo- and hypersensitive subpopulations), and sham-injury + 670 nm treatment (shamSCI+670, light red). Pairwise statistical comparisons are indicated for RSSs and LSSs by respective group colours. Note: statistical comparisons of CSSs from shamSCI+670 group in (a) is to the normosensitive subpopulation of SCI (indicated in dark blue) and to control groups (indicated in green); Statistical comparisons of RSSs from control group in (c) is to SCI (indicated in dark blue) or to SCI+670 (indicated in dark red) in b. *p < 0.05 (Student’s t test); † p < 0.05, †† p < 0.01, ††† p < 0.001, †††† p < 0.0001, (Wilcoxon rank-sum); ns, p > 0.05; n values indicated Red light improves sensory conduction through dorsal column pathways Could red light cause an anaesthetic-like effect on somatosensation that resulted in reduced sensitivity scores? To rule out the possibility that red light causes a reduced responsiveness to innocuous stimuli by bringing about a generalized inhibitory effect on somatic neural pathway conduction, we quantified the functional integrity of the sensory dorsal column pathway, at 7 days post-injury. The dorsal column pathways were activated by electrical stimulation of the left and right sural nerves, and a recording electrode was placed on the midline of the gracile nuclei (Fig. 3a). Stimulation of left and right nerves from control animals (n = 7) evoke responses of equal magnitude (Fig. 3b; right side: 101 ± 8 % of left side) and latency (Fig. 3c; left-right side latency difference: 0.09 ± 0.03 ms) on both sides when recorded from the same midline-positioned recording electrode, while sham-treated T10 hemicontusion spinal cord injury (n = 7) resulted in a 37 % reduction in magnitude (right side: 63 ± 16 % of left side) and a 0.48 ± 0.09 ms delay of the injured (right) pathway, when comparing the intact (left) side. Red light treatment (n = 7) rescued both the magnitude (Fig. 3b; right side: 93 ± 17 % of left side) and latency (Fig. 3c; left-right side latency difference: −0.05 ± 0.35 ms) deficits otherwise observed in the SCI group, indicating that red light treatment restored sensory pathway conduction, rather than impeding it. Furthermore, the rescued magnitude and latency deficits in the SCI+670 group indicates that their reduced sensitivity scores (Fig. 2) were unlikely to have resulted from a generalized reduction of somatic neural conduction.Fig. 3 Dorsal column somatosensory functional deficits from T10 hemicontusion spinal cord injury is reversed by red light treatment. a Schematic of experimental paradigm for evaluating somatosensory (dorsal column pathway) functional integrity illustrating left and right dorsal column pathways (grey), T10 hemicontusion injury on right side, stimulation of sural nerves and location of recording electrode on midline of gracile nucleus. The same electrode position on the midline acquires somatosensory responses independently evoked from both left and right sural nerves, enabling direct comparable quantification of sensory pathways on both sides. Examples of responses (between 5 and 15 ms post-stimulus; 500–3350 Hz bandpass) evoked from left and right sides shown for respective groups (colour-coded as per legend in c and Fig. 2). Arrowheads indicate latency of response onset. b Quantification (integral of rectified signals) of gracile nucleus response magnitudes (right expressed as a percent of left). c Difference in latencies of evoked responses between left and right sides. Note magnitudes and latencies from intact animals are equal on both sides (control group). *p < 0.05; **p < 0.01, Student’s t test, Tukey’s post hoc in c We performed a variety of control experiments to validate our interpretations. There was no observable difference of conduction magnitudes or latencies in any of the sham-injured animals (shamSCI, n = 4; shamSCI+670, n = 4). There was no significant difference between gracile nuclei potentials evoked from the left sural nerve in any of the groups (SCI, 15.9 ± 1.8 μV · ms; SCI+670, 11.9 ± 2.4 μV · ms; control, 16.2 ± 3.6 μV · ms; shamSCI, 10.8 ± 2.6 μV · ms; shamSCI+670, 15.0 ± 2.8 μV · ms; p = 0.70, one-way ANOVA). Similarly, there was no significant difference of response latencies when evoked on the left side for all groups (SCI, 33.7 ± 0.3 μV · ms; SCI+670, 34.0 ± 0.4 μV · ms; control, 34.0 ± 0.4 μV · ms; shamSCI, 34.2 ± 0.5 μV · ms; shamSCI+670, 34.7 ± 0.3 μV · ms; p = 0.51, one-way ANOVA). These control experiments indicated that dorsal column pathway response magnitudes and latencies were similar between the different groups and largely unaffected contralateral to the injury. Red light improves locomotor recovery Could red light treatment cause motor deficits and thereby result in reduced sensitivity scores? To rule out the possibility that the red light impeded the animals’ ability to perform escaping locomotor behaviours, daily locomotor recovery was examined blind to the experimental group (Fig. 4). We found that rather than impeding locomotion, the SCI+670 group (n = 11) demonstrated improved locomotor recovery as early as 2 days post-injury on the ipsilateral side and 3 days post-injury on the contralateral side compared to the sham-treated group (n = 10). Although a group effect of red light improvement was evident on the ipsilateral side (p = 0.026, linear mixed effects model with repeated measures), this failed to reach significance on the contralateral side (p = 0.055). There was a highly significant effect of time for both sides (p < 2e-16). Locomotor improvements observed in the SCI+670 group indicate that reduced sensitivity scores in light-treated animals (Fig. 2) could not have resulted from locomotor deficits.Fig. 4 Locomotor recovery is improved by red light treatment following T10 hemicontusion spinal cord injury. Daily locomotor scores (BBB, see the “Methods” section) following a right-sided hemicontusion spinal cord injury are shown for the contralateral (a) and ipsilateral (b) sides. Red light treatment results in significant locomotor improvements on both sides over the period indicated by the black bar (large asterisk, linear mixed model with repeated measures). Point-wise comparisons between groups for individual time points are also shown (small asterisks, Student’s t test). Individual data points are presented as open square or circular dots; lines indicate the group means. *p < 0.05; **p < 0.01 Red light reduces cell death at the injury zone To examine the effect of red light on cell death following injury, the number of TUNEL+ cells was quantified at 1, 3 and 7 days post-injury in dorsal regions of the T10 spinal cord (Fig. 5, n = 5 for each time point). The SCI group resulted in an increased density of TUNEL+ cells in the dorsal spinal cord ipsilateral to the injury as early as day 1 (contralateral 1.5 ± 1.5 cells/mm2; ipsilateral 96.8 ± 41.1 cells/mm2), reaching maximum levels by day 3 (contralateral 13.1 ± 5.6 cells/mm2; ipsilateral 126.8 ± 41.5 cells/mm2). The contralateral side had much fewer cells where maximum levels were reached by day 7 (Fig. 5; contralateral 32.5 ± 32.5 cells/mm2; ipsilateral 74.2 ± 43.7 cells/mm2). Red light treatment resulted in a significant group reduction of TUNEL+ cells in the ipsilateral side, notably significant at the day 3 time point when TUNEL+ cells were maximal in the sham-treated group (1 dpi: 49.6 ± 25.2 cells/mm2; 3 dpi 18.2 ± 3.9 cells/mm2; 7 dpi 22.0 ± 6.1 cells/mm2). There was no significant difference in TUNEL labelling on the contralateral side between groups (1 dpi: 2 ± 2 cells/mm2; 3 dpi 6.2 ± 2.1 cells/mm2; 7 dpi 5.0 ± 3.9 cells/mm2).Fig. 5 Cell death is reduced by red light following T10 hemicontusion spinal cord injury. Quantification of cells undergoing cell death (TUNEL+) contralateral (a) and ipsilateral (b) to the injury. Example images are from SCI (c) and SCI+670 (d) dorsal horn ipsilateral to the injury at 3 days post-injury. Schematic cross section of spinal cord (bottom) indicates location of injury (dark grey penumbra) and region of quantification (light grey region). Scale bars: 50 μm. *p < 0.05 (Student’s t test); **p < 0.01 (linear mixed model) Red light reduces total activated microglia/macrophages but promotes the expression of the anti-inflammatory/wound-healing (M2) subtype Inflammation has long being implicated in the development of neuropathic pain [27]. We therefore quantified activated microglia/macrophages (ED1+ cells) at 1, 3 and 7 days post-injury in dorsal regions of T10 spinal cord (Fig. 6a–d, n = 5 for each time point). T10 spinal contusion resulted in an increase in ED1+ cell density as early as day 1 post-injury, reaching maximum levels by day 3 in the ipsilateral side. Maximum levels were also reached at day 3 on the contralateral side, but there were negligible ED1+ cells at days 1 and 7. Light treatment significantly reduced ED1 expression ipsilateral to the injury to approximately half that of the SCI group. Despite the low levels of ED1+ cells in the contralateral side, red light treatment also resulted in a significant reduction of ED1+ cells at the 3-day time point.Fig. 6 Anti-inflammatory microglia/macrophages are promoted early by red light treatment following T10 hemicontusion spinal cord injury. a–d Total activated microglia/macrophages (ED1+) per mm2 contralateral (a) and ipsilateral (b) to the injury and example images from SCI (c) and SCI+670 (d) groups. e–h M1 (pro-inflammatory) microglia/macrophages (CD80+ED1+ double labelled) expressed as a proportion of total ED1+ cells contralateral (e) and ipsilateral (f) to the injury and example images from SCI (g) and SCI+670 (h) groups. i–l: M2 (anti-inflammatory) microglia/macrophages (Arginase1+ED1+ double labelled) expressed as a proportion of total ED1+ cells contralateral (i) and ipsilateral (j) to the injury and example images from SCI (k) and SCI+670 (l) groups. All example images are taken from the injury zone of the dorsal horn at 7 days post-injury. Schematic cross section of spinal cord (bottom) indicates location of injury (dark grey penumbra) and region of quantification (light grey region). Scale bars: 50 μm. *p < 0.05 (linear mixed model); **p < 0.01, ***p < 0.001 (Student’s t test); † p < 0.05, †† p < 0.01, ††† p < 0.001 (Wilcoxon rank-sum) Microglia/macrophages can adopt pro- or anti-inflammatory states [30]. To determine the effect of red light treatment on the expression of pro-inflammatory (M1) cells, cells co-expressing CD80 and ED1 were quantified as a proportion of total ED1+ cells (Fig. 6e–h, n = 5 for each time point). The proportion of CD80+ED1+ cells ipsilateral to the injury was maximal at day 1 and remained greater than 40 % of the ED1 population at days 3 and 7 in more than half of animals. CD80+ED1+ cells were only found at day 3 on the contralateral side which coincided with the maximum number of ED1+ cells at that time point. Red light treatment did not have a significant impact on the proportion of M1 cells on either the ipsi- or contralateral sides. Note that no CD80+ED1+ cells were encountered at days 1 and 7 contralateral to the injury as ED1+ cells were also in small quantities at these time points (Fig. 6a). To determine the effect of red light treatment on the expression of anti-inflammatory/wound-healing (M2) microglia/macrophages, ED1+ cells co-expressing Arginase-1 were quantified as a proportion of total ED1+ cells (Fig. 6i–l, n = 5 for each time point). In the SCI group, Arginase-1 expression increased with time ipsilateral to the injury (p = 0.0048, one-way ANOVA) but peaked at day 3 contralateral to the injury at the time when most ED1+ cells were present in that region. Ipsilateral to the injury, the SCI+670 group displayed significantly increased proportions of Arginase1+ED1+ cells from day 1, reaching approximately sevenfold that of the SCI group. This greater Arginase1+ED1+ proportion in light-treated animals was maintained at over one third of ED1+ cells for the entire duration investigated for the majority of animals. No Arginase1+ED1+ cells were detected contralateral to the injury in the SCI+670 group; however, there were very few ED1+ cells in this region (Fig. 6a). The group effect failed to reach significance contralateral to the injury (p = 0.0628) despite a significantly greater level of Arginase1+ED1+ cells in the SCI group. Discussion We demonstrate that following spinal cord injury, 35 mW/cm2 of red (670 nm) light transcutaneously applied for 30 min/day for 7 days to the dorsal surface of rats is sufficient to reach the entire spinal cord and reduce the expression of pain behaviours. These reduced signs of allodynia are not due to sensorimotor deficits, as red light treatment improves both sensory and motor function. Alleviated hypersensitivity, improved tactile/proprioception (dorsal column) pathway functional integrity and locomotor functional outcomes are preceded by reduced numbers of dying cells and reduced numbers of activated microglia/macrophages around the injury zone. Furthermore, the proportion of anti-inflammatory/wound-healing (M2) microglia/macrophages is greatly enhanced by 24 h following light treatment. We are confident that the power output of the red light was sufficient to penetrate the entire rat spinal cord as red light could be seen with the naked eye illuminating through to the ventral surface of the cord in the cadaver models. While penetration through to the rat spinal cord was achievable with an intensity of 35 mW/cm2, future studies would be required to determine the exposure parameters to achieve an equivalent level of irradiation in humans. Our finding of 91 % absorption (9 % excess penetration) is a conservative measure for two main reasons: (i) penetration measurements were obtained through the hair of unshaven animal cadavers (the injury site of all injured animals was shaven) and (ii) deoxygenated haemoglobin absorbs 670 nm significantly more than oxygenated haemoglobin [41, 42]. Measurements from freshly scarified animals are therefore likely to have increased levels of deoxygenated blood, and thus reduced penetration, compared to live animals. Another factor to consider is the small attenuation of light as a function of distance from its source. Our estimation indicates that over 93 % of the light would have reached the spinal cord ventral surface if no intervening tissues were present to absorb the light; thus, the effect of distance appears to be negligible. Nine percent excess penetration (i.e. 91 % absorption) from the surface of the skin (with hair intact) through all intervening tissue layers to the ventral surface of the spinal cord with a device delivering approximately 35 mW/cm2 is consistent with a recent study that demonstrated an excess penetration of 6.6 % through the surface of the skin and the muscle overlying the spinal cord with a device producing approximately 16 mW/cm2 in Fisher rat cadavers [9]. Temperature of sham-treated animals was not significantly different before and after treatment, while light-treated animals’ experienced a significant 1.2 °C increase. This increase does not exceed the normal range for rat tail skin temperature variations which have been reported to oscillate by ± 2 °C within a 2-h time frame [43]. However, as there was a small but significant temperature increase 2 min after light treatment, it is likely that there was a larger temperature increase during the 30 min treatment period. We therefore cannot rule out the possibility that temperature increases did not impact on our findings. Nevertheless, red light treatment does result in significant functional and cellular improvements, regardless if temperature is a contributing factor. If temperature increases were to contribute toward improved outcomes, it would be in contrast to studies of hypothermic treatment which propose superior outcomes following spinal cord injury [44–46]. As the mechanisms of action for light-treatment improvements remain to be elucidated, future investigations isolating the effect of temperature and light are warranted. To our knowledge, our study is the first to report a red light-induced locomotor improvement following a spinal cord injury, which contradicts the only other study by Giacci et al. [9] that examined 670 nm on locomotor recovery with a daily dose of 28.4 J/cm2, an intensity of 15.8 mW/cm2 for 30 min, i.e. less than half the intensity of the present study. The compounded effect of reduced intensity and a more severe contusion injury in their study may explain this difference, and furthermore, suggests that matching the appropriate light dosage to the injury severity is of paramount importance. Our T10 hemicontusion injury model resulted in allodynia within 7 days in a subset of animals. We are confident that our injury model results in neuropathic pain because hypersensitivity developed above and below the level, as well as contralateral to the injury, i.e. at dermatomes that receive their innervation from outside the injury epicentre. This observation is also consistent with findings from an investigation using a C5 hemicontusion injury model and which also found a subset of animals developing allodynia from 7 days post-injury that lasted for least 42 days [47]. Our observation of allodynia on the animals’ dorsum is also consistent with a T13 hemisection injury model that also results in clear development of hypersensitivity in most animals within 7 days and that remains persists for several weeks [48]. We found that red light treatment reduced the severity, but not the incidence of hypersensitivity at 7 days post-injury. As allodynia reached sensitivity levels of almost four times that of the hypersensitivity threshold, we would expect that a milder injury causing sensitivity scores closer to the hypersensitivity threshold boarder would result in a reduction of both the severity and incidence of hypersensitivity. The finding that the shamSCI+670 group demonstrated sensitivity significantly lower than that of the intact control and shamSCI groups was curious. Sham injury may have activated anti-nociceptive descending pathways such as periaqueductal grey/raphe magnus-mediated inhibition of dorsal horn nociceptive inputs [49]. Thus, endogenous central anti-nociceptive mechanisms, compounded by a red light-induced anti-inflammatory microenvironment, could be responsible for the sham-injured-light-treated animals expressing less sensitivity than observed in uninjured animals. This speculation warrants further investigation as red light-augmented relief from pain would have significant clinical relevance for post-surgical pain treatment. Quantification of pain behaviours relies on sensory and motor functional integrity. We are confident that the reduced expression of allodynia in red light-treated animals was not due to diminished general somatic sensation or impeded motor function because red light improved, rather than impeded these parameters. Locomotor recovery and sensitivity testing was scored blind to the experimental group, and therefore, any subjective bias was eliminated. Sural nerve evoked somatosensory potentials in the gracile nuclei provided an objective and precise measure of somatic sensory functional integrity of both left and right dorsal column pathways. Sural nerves were stimulated to recruit all nerve fibres, and therefore, input to the spinal cord was identical on both sides, while the recording conditions on the midline of the gracile nuclei were also identical during the acquisition of evoked potentials elicited from pathways of both sides. Therefore, the only difference in the responses between the left and right sides was due to alterations within their respective dorsal column pathways. We further confirmed this by demonstrating equal magnitudes and latencies of somatosensory potentials in the gracile nuclei when evoked from left and right sural nerves of intact and sham-injured animals. Thus, our data indicates that reduced expression of behavioural signs of pain following red light treatment is unlikely to have resulted from locomotor or sensory deficiencies, but rather, represents a true reduction of pain experienced by the light-treated rats. While our study is the first to demonstrate red light-induced pain relief from spinal cord injury, it is consistent with peripheral nerve injury studies that report pain relief accompanied by light-induced alterations to the inflammatory response [13–15, 50, 51]. The functional improvements found in red light-treated animals were observed after a significant reduction in cell death was apparent at day 3 post-injury, a time coincident with maximal levels of activated microglia/macrophages in the injury zone of sham-treated animals. Our observations of reduced ED1+ cells in 670-nm-treated animals is consistent with that found in retinal damage [7], as well as another study that demonstrated similar proportions of ED1 cell suppression lasting up to 14 days post-corticospinal tract lesion in rats that received daily 810 nm diode laser treatments [16]. In the latter study by Byrnes et al., they also demonstrated functional improvements of some motor tasks, also consistent with improved locomotor function observed in our study. While we cannot speculate on the mechanisms for 810 vs. 670 nm wavelengths to suppress microglia/macrophage activation and improve motor function, it is noteworthy that both wavelengths evoke peak levels of cytochrome C oxidase activity and ATP production [52]. However, wavelength (i.e. 660 vs. 780 nm) has been shown to alter the expression of inflammatory mediators expressed by activated pro-inflammatory microglia/macrophages [37], and light dosage has been shown to alter the balance of M1/M2 cell expression [53]. These in vitro studies suggest that other mechanisms, unrelated to cytochrome C oxidase, may influence the inflammatory microenvironment following light treatment. Furthermore, they highlight the necessity for thorough investigations to establish the therapeutic limits of any wavelength under investigation. While others have demonstrated the impact of wavelength and dose on inflammatory cells in vitro [37, 53], to our knowledge, we are the first to demonstrate the effect of 670 nm light on the polarization of activated microglia/macrophages following spinal cord injury in vivo. The pattern and sequence of pro-inflammatory M1 (CD80+ED1+) cell activation, cell death, followed by anti-inflammatory/wound-healing M2 (Arginase1+ED1+) recruitment observed in sham-treated animals in our study is consistent with what is expected under conditions of spinal cord injury and repair [30, 31]. However, our data indicates that red light-induced reduction of cell death is preceded by the upregulation of M2 cells as early as 1 day post-injury. This is intriguing because the M2 cell expression preceded that of the M1 cells, indicating that red light drastically altered the normal sequence of inflammatory events. We therefore speculate that the early presence of protective M2 cells may have caused the reduced subsequent population of dying (TUNEL+) cells. Insufficient expression of the M2 subtype in spinal cord injury, in contrast to peripheral nerve injury, has been suggested to be a contributing factor to the poorer regenerative capacity and functional outcomes in spinal cord injury compared to peripheral nerve injury [30, 31]. The present study found that red light had a strong impact on promoting the M2 cell types as early as 24 h after treatment which was followed by reduced levels of cell death and subsequent improvement of sensory and motor functional outcomes thereafter. This is consistent with previous suggestions that enhancing the M2 population during recovery from spinal cord injury may indeed significantly contribute to improving functional outcomes following spinal cord injury [31]. Conclusions Modulating the severity of neuropathic pain by simply applying red light is an exciting prospect with great significant clinical relevance, despite not yet fully understanding the mechanism behind photobiomodulation. Our data demonstrates that red light treatment, a non-invasive and cost effective treatment, is able to significantly reduce the severity of pain in rats acutely after spinal cord injury, and these behavioural changes are accompanied by alterations to the alternatively activated macrophage population. Early pain intervention is considered important to avoid the prospects of developing chronic pain [54]. As 670 nm light therapy is FDA approved, it could be quickly adopted as an adjunct to early treatment of spinal cord injury. Not only could this minimise the severity of pain to sufferers, it may also provide collateral benefits which include functional improvements to other sensory/motor systems. However, translation to human patients requires further studies to determine exposure parameters such as the light intensity necessary to penetrate the human spinal cord. Abbreviations ANOVAAnalysis of variance CSSCumulative sensitivity score BBBBasso, Beattie and Bresnahan CNSCentral nervous system FDAFood and Drug Administration LEDLight-emitting diode LSSLevel sensitivity score M1Th1-activated microglia/macrophages M2Th2-activated microglia/macrophages PBSPhosphate-buffered saline RSSRegional sensitivity score shamSCISham-treated sham-injured group shamSCI+670670-nm-treated sham-injured group SCISham-treated spinal cord injured group SCI+670670-nm-treated spinal cord injured group SSCSaline sodium citrate TdTTerminal deoxynucleotidyl transferase Th1T helper cell type 1 Th2T helper cell type 2 TUNELTerminal deoxynucleotidyl transferase (TdT) dUTP nick-end labelling Acknowledgements The authors wish to graciously thank the Gretel and Gordon Bootes Medical Research Foundation for their generous donations which funded this project. The authors also wish to thank Edward Scharrer and Michael Percival for the design and construction of the custom 670 nm light probe. Funding This study was funded by the Gretel and Gordon Bootes Medical Research Foundation. This funding body had no role in the design of the study, collection, analysis and interpretation of data or in preparation of the manuscript. Availability of data and materials The datasets supporting the conclusions of this article are available in the figshare repository, DOI: https://dx.doi.org/10.6084/m9.figshare.3172756. Authors’ contributions JRP and DH conceptualised and designed the experiments; DH and SZ conducted the experiments; DH, JRP and SZ conducted the analysis; DH and JRP wrote the paper. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval All animal experiments were approved by The Australian National University Animal Experimentation Ethics Committee (protocol number: A2014/52). ==== Refs References 1. Gaskin DJ Richard P The economic costs of pain in the United States J Pain 2012 13 8 715 24 10.1016/j.jpain.2012.03.009 22607834 2. Anderson RR Parrish JA The optics of human skin J Invest Dermatol 1981 77 1 13 9 10.1111/1523-1747.ep12479191 7252245 3. Fitzgerald M Hodgetts S Van Den Heuvel C Natoli R Hart NS Valter K Red/near-infrared irradiation therapy for treatment of central nervous system injuries and disorders Rev Neurosci 2013 24 2 205 26 10.1515/revneuro-2012-0086 23492552 4. Grillo SL Duggett NA Ennaceur A Chazot PL Non-invasive infra-red therapy (1072 nm) reduces beta-amyloid protein levels in the brain of an Alzheimer’s disease mouse model, TASTPM J Photochem Photobiol B 2013 123 13 22 10.1016/j.jphotobiol.2013.02.015 23603448 5. El Massri N, Johnstone DM, Peoples CL, Moro C, Reinhart F, Torres N, et al. The effect of different doses of near infrared light on dopaminergic cell survival and gliosis in MPTP-treated mice. Int J Neurosci. 2015. Epub 2014/12/04. doi: 10.3109/00207454.2014.994063. 6. Purushothuman S Johnstone DM Nandasena C Eersel J Ittner LM Mitrofanis J Near infrared light mitigates cerebellar pathology in transgenic mouse models of dementia Neurosci Lett 2015 591 155 9 10.1016/j.neulet.2015.02.037 25703226 7. Albarracin R Eells J Valter K Photobiomodulation protects the retina from light-induced photoreceptor degeneration Invest Ophthalmol Vis Sci 2011 52 6 3582 92 10.1167/iovs.10-6664 21421867 8. Eells JT Henry MM Summerfelt P Wong-Riley MT Buchmann EV Kane M Therapeutic photobiomodulation for methanol-induced retinal toxicity Proc Natl Acad Sci U S A 2003 100 6 3439 44 10.1073/pnas.0534746100 12626762 9. Giacci MK Wheeler L Lovett S Dishington E Majda B Bartlett CA Differential effects of 670 and 830 nm red near infrared irradiation therapy: a comparative study of optic nerve injury, retinal degeneration, traumatic brain and spinal cord injury PLoS One 2014 9 8 e104565 10.1371/journal.pone.0104565 25105800 10. Fitzgerald M Bartlett CA Payne SC Hart NS Rodger J Harvey AR Near infrared light reduces oxidative stress and preserves function in CNS tissue vulnerable to secondary degeneration following partial transection of the optic nerve J Neurotrauma 2010 27 11 2107 19 10.1089/neu.2010.1426 20822460 11. Barbosa RI Marcolino AM de Jesus Guirro RR Mazzer N Barbieri CH de Cassia Registro Fonseca M Comparative effects of wavelengths of low-power laser in regeneration of sciatic nerve in rats following crushing lesion Lasers Med Sci 2010 25 3 423 30 10.1007/s10103-009-0750-8 20135336 12. Medalha CC Di Gangi GC Barbosa CB Fernandes M Aguiar O Faloppa F Low-level laser therapy improves repair following complete resection of the sciatic nerve in rats Lasers Med Sci 2012 27 3 629 35 10.1007/s10103-011-1008-9 22009383 13. Bertolini GR Artifon EL Silva TS Cunha DM Vigo PR Low-level laser therapy, at 830 nm, for pain reduction in experimental model of rats with sciatica Arq Neuropsiquiatr 2011 69 2B 356 9 10.1590/S0004-282X2011000300017 21625765 14. Cidral-Filho FJ Martins DF More AO Mazzardo-Martins L Silva MD Cargnin-Ferreira E Light-emitting diode therapy induces analgesia and decreases spinal cord and sciatic nerve tumour necrosis factor-alpha levels after sciatic nerve crush in mice Eur J Pain 2013 17 8 1193 204 10.1002/j.1532-2149.2012.00280.x 23339021 15. Hsieh YL Chou LW Chang PL Yang CC Kao MJ Hong CZ Low-level laser therapy alleviates neuropathic pain and promotes function recovery in rats with chronic constriction injury: possible involvements in hypoxia-inducible factor 1alpha (HIF-1alpha) J Comp Neurol 2012 520 13 2903 16 10.1002/cne.23072 22351621 16. Byrnes KR Waynant RW Ilev IK Wu X Barna L Smith K Light promotes regeneration and functional recovery and alters the immune response after spinal cord injury Lasers Surg Med 2005 36 3 171 85 10.1002/lsm.20143 15704098 17. Hodgson BD Margolis DM Salzman DE Eastwood D Tarima S Williams LD Amelioration of oral mucositis pain by NASA near-infrared light-emitting diodes in bone marrow transplant patients Support Care Cancer 2012 20 7 1405 15 10.1007/s00520-011-1223-8 21725826 18. Elwakil TF Elazzazi A Shokeir H Treatment of carpal tunnel syndrome by low-level laser versus open carpal tunnel release Lasers Med Sci 2007 22 4 265 70 10.1007/s10103-007-0448-8 17334675 19. Naeser MA Photobiomodulation of pain in carpal tunnel syndrome: review of seven laser therapy studies Photomed Laser Surg 2006 24 2 101 10 10.1089/pho.2006.24.101 16706688 20. Shooshtari SM Badiee V Taghizadeh SH Nematollahi AH Amanollahi AH Grami MT The effects of low level laser in clinical outcome and neurophysiological results of carpal tunnel syndrome Electromyogr Clin Neurophysiol 2008 48 5 229 31 18754533 21. Turhani D Scheriau M Kapral D Benesch T Jonke E Bantleon HP Pain relief by single low-level laser irradiation in orthodontic patients undergoing fixed appliance therapy Am J Orthod Dentofacial Orthop 2006 130 3 371 7 10.1016/j.ajodo.2005.04.036 16979496 22. Maia ML Bonjardim LR Quintans Jde S Ribeiro MA Maia LG Conti PC Effect of low-level laser therapy on pain levels in patients with temporomandibular disorders: a systematic review J Appl Oral Sci 2012 20 6 594 602 10.1590/S1678-77572012000600002 23329239 23. Chow RT Johnson MI Lopes-Martins RA Bjordal JM Efficacy of low-level laser therapy in the management of neck pain: a systematic review and meta-analysis of randomised placebo or active-treatment controlled trials Lancet 2009 374 9705 1897 908 10.1016/S0140-6736(09)61522-1 19913903 24. Ribas ES, Paiva WS, Pinto NC, Yeng LT, Okada M, Fonoff ET, et al. Use of low intensity laser treatment in neuropathic pain refractory to clinical treatment in amputation stumps. International journal of general medicine. 2012;5:739-42. Epub 2012/10/11. doi: 10.2147/ijgm.s18511. PubMed PMID: 23049266; PubMed Central PMCID: PMCPMC3459668. 25. Lees JG Duffy SS Moalem-Taylor G Immunotherapy targeting cytokines in neuropathic pain Front Pharmacol 2013 4 142 10.3389/fphar.2013.00142 24319429 26. Lees JG Duffy SS Perera CJ Moalem-Taylor G Depletion of Foxp3+ regulatory T cells increases severity of mechanical allodynia and significantly alters systemic cytokine levels following peripheral nerve injury Cytokine 2015 71 2 207 14 10.1016/j.cyto.2014.10.028 25461400 27. Schomberg D Ahmed M Miranpuri G Olson J Resnick DK Neuropathic pain: role of inflammation, immune response, and ion channel activity in central injury mechanisms Ann Neurosci 2012 19 3 125 32 25205985 28. Yezierski RP Pain following spinal cord injury: pathophysiology and central mechanisms Prog Brain Res 2000 129 429 49 10.1016/S0079-6123(00)29033-X 11098709 29. Clark AK Gruber-Schoffnegger D Drdla-Schutting R Gerhold KJ Malcangio M Sandkuhler J Selective activation of microglia facilitates synaptic strength J Neurosci 2015 35 11 4552 70 10.1523/JNEUROSCI.2061-14.2015 25788673 30. David S Kroner A Repertoire of microglial and macrophage responses after spinal cord injury Nat Rev Neurosci 2011 12 7 388 99 10.1038/nrn3053 21673720 31. Kigerl KA Gensel JC Ankeny DP Alexander JK Donnelly DJ Popovich PG Identification of two distinct macrophage subsets with divergent effects causing either neurotoxicity or regeneration in the injured mouse spinal cord J Neurosci 2009 29 43 13435 44 10.1523/JNEUROSCI.3257-09.2009 19864556 32. Murray PJ Wynn TA Protective and pathogenic functions of macrophage subsets Nat Rev Immunol 2011 11 11 723 37 10.1038/nri3073 21997792 33. Mantovani A Biswas S Galdiero M Sica A Locati M Macrophage plasticity and polarization in tissue repair and remodelling J Pathol 2013 229 2 176 85 10.1002/path.4133 23096265 34. Mantovani A Locati M Orchestration of macrophage polarization Blood 2009 114 3135 6 10.1182/blood-2009-07-231795 19815678 35. Mantovani A Sica A Sozzani S Allavena P Vecchi A Locati M The chemokine system in diverse forms of macrophage activation and polarization Trends Immunol 2004 25 12 677 86 10.1016/j.it.2004.09.015 15530839 36. Safavi SM Kazemi B Esmaeili M Fallah A Modarresi A Mir M Effects of low-level He-Ne laser irradiation on the gene expression of IL-1beta, TNF-alpha, IFN-gamma, TGF-beta, bFGF, and PDGF in rat’s gingiva Lasers Med Sci 2008 23 3 331 5 10.1007/s10103-007-0491-5 17786499 37. Fernandes KP Souza NH Mesquita-Ferrari RA Silva Dde F Rocha LA Alves AN Photobiomodulation with 660-nm and 780-nm laser on activated J774 macrophage-like cells: effect on M1 inflammatory markers J Photochem Photobiol B 2015 153 344 51 10.1016/j.jphotobiol.2015.10.015 26519828 38. Vijayaprakash KM Sridharan N An experimental spinal cord injury rat model using customized impact device: A cost-effective approach J Pharmacol Pharmacother 2013 4 3 211 3 10.4103/0976-500X.114607 23960429 39. Coderre TJ Fundytus ME McKenna JE Dalal S Melzack R The formalin test: a validation of the weighted-scores method of behavioural pain rating Pain 1993 54 1 43 50 10.1016/0304-3959(93)90098-A 8378102 40. Basso DM Beattie MS Bresnahan JC A sensitive and reliable locomotor rating scale for open field testing in rats J Neurotrauma 1995 12 1 1 21 10.1089/neu.1995.12.1 7783230 41. Tin W Lal M Principles of pulse oximetry and its clinical application in neonatal medicine Semin Fetal Neonatal Med 2015 20 3 192 7 10.1016/j.siny.2015.01.006 25704605 42. Meglinski IV Matcher SJ Quantitative assessment of skin layers absorption and skin reflectance spectra simulation in the visible and near-infrared spectral regions Physiol Meas 2002 23 4 741 53 10.1088/0967-3334/23/4/312 12450273 43. Gordon CJ Puckett E Padnos B Rat tail skin temperature monitored noninvasively by radiotelemetry: characterization by examination of vasomotor responses to thermomodulatory agents J Pharmacol Toxicol Methods 2002 47 107 114 10.1016/S1056-8719(02)00219-8 12459150 44. Maybhate A, Hu C, Bazley FA, Yu Q, Thakor NV, Kerr CL, et al. Potential long-term benefits of acute hypothermia after spinal cord injury: assessments with somatosensory-evoked potentials. Critical care medicine. 2012;40(2):573-9. Epub 2011/10/18. doi: 10.1097/CCM.0b013e318232d97e. PubMed PMID: 22001581; PubMed Central PMCID: PMCPMC3261348. 45. Ahmad FU Wang MY Levi AD Hypothermia for acute spinal cord injury—a review World Neurosurg 2014 82 1-2 207 14 10.1016/j.wneu.2013.01.008 23298671 46. Alkabie S Boileau AJ The role of therapeutic hypothermia after traumatic spinal cord injury—a systematic review World Neurosurg 2016 86 432 49 10.1016/j.wneu.2015.09.079 26433095 47. Detloff MR, Wade RE, Jr., Houle JD. Chronic at- and below-level pain after moderate unilateral cervical spinal cord contusion in rats. Journal of neurotrauma. 2013;30(10):884-90. doi: 10.1089/neu.2012.2632. PubMed PMID: 23216008; PubMed Central PMCID: PMCPMC3660073. 48. Christensen MD Hulsebosch CE Chronic central pain after spinal cord injury J Neurotrauma 1997 14 8 517 37 10.1089/neu.1997.14.517 9300563 49. Ossipov MH Dussor GO Porreca F Central modulation of pain J Clin Invest 2010 120 11 3779 87 10.1172/JCI43766 21041960 50. Giuliani A Fernandez M Farinelli M Baratto L Capra R Rovetta G Very low level laser therapy attenuates edema and pain in experimental models Int J Tissue React 2004 26 1-2 29 37 15573690 51. Lorenzini L Giuliani A Giardino L Calza L Laser acupuncture for acute inflammatory, visceral and neuropathic pain relief: an experimental study in the laboratory rat Res Vet Sci 2010 88 1 159 65 10.1016/j.rvsc.2009.06.001 19560793 52. Desmet KD Paz DA Corry JJ Eells JT Wong-Riley MT Henry MM Clinical and experimental applications of NIR-LED photobiomodulation Photomed Laser Surg 2006 24 2 121 8 10.1089/pho.2006.24.121 16706690 53. von Leden RE Cooney SJ Ferrara TM Zhao Y Dalgard CL Anders JJ 808 nm wavelength light induces a dose-dependent alteration in microglial polarization and resultant microglial induced neurite growth Lasers Surg Med 2013 45 4 253 63 10.1002/lsm.22133 23619903 54. Siddall PJ McClelland JM Rutkowski SB Cousins MJ A longitudinal study of the prevalence and characteristics of pain in the first 5 years following spinal cord injury Pain 2003 103 3 249 57 10.1016/S0304-3959(02)00452-9 12791431

PMC005xxxxxx/PMC5000420.txt

==== Front BMC Complement Altern MedBMC Complement Altern MedBMC Complementary and Alternative Medicine1472-6882BioMed Central London 130410.1186/s12906-016-1304-9Research ArticleEffects of panax notoginseng saponins on the osteogenic differentiation of rabbit bone mesenchymal stem cells through TGF-β1 signaling pathway Wang Yan waynewang007@126.com Huang Xuanping 1017207360@qq.com Tang Yiyao 523664646@qq.com Lin Haiyun waynewang008@163.com Zhou Nuo nuozhou@hotmail.com College & Hospital of Stomatology, Guangxi Medical University, 22, Shuangyong Road, Nanning, Guangxi China 26 8 2016 26 8 2016 2016 16 1 31919 4 2016 18 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Panax Notoginseng is a well-known Chinese medicinal herb which has been used in China for treatment of bone fracture for hundreds of years. However, the specific biological mechanisms of osteogenic effect of PNS are not well understood. Methods In this study, newborn rabbit BMSCs were isolated, and then identified by the positive expression rates of cell surface markers, including CD29, CD45 and HLA-DR, which were detected by flow cytometry(FCM). After the lentivirus-induced cell model of TGF-β1 gene silencing was established, the interference efficiency was tested by q-PCR and Western blot, and the growth curve of silencing cells was drawn by MTT so as to grasp the growth rhythm of silencing cells. In the alizarin red-staining experiment, the effect of 100 mg/L PNS on the activity of intracellular ALP of TGF-β1 gene silencing BMSCs was detected, so as to observe the effect of 100 mg/L PNS on the formation of calcium nodes of gene silencing BMSCs. Results By separating rabbit BMSCs, the lentivirus-induced cell model of TGF-β1 gene silencing was established. Both TGF-β1 mRNA and protein expression were restrained significantly, and the target gene kept silence stably via the verification of q-PCR and Western blot; there was no significant differences of the growth curve between RNAi cells and normal cells; the activity of intracellular APL in 100 mg RNAi group was obviously lower than that in 100 mg group (p < 0.05), but higher than that in the normal group; in the alizarin red-staining experiment, it focused on the effects of PNS on the formation of calcium nodes of gene silencing BMSCs, which showed that calcium nodes could be formed in 100 mg RNAi group but its quantity was lower than that of 100 mg group (p < 0.05). Conclusions It was shown that silencing TGF-β1 gene could interrupt the osteogenic effects of PNS. PNS may have a promoting effect on osteogenic differentiation of rabbits’ BMSCs in vitro by up-regulating the gene expression of TGF-β1. Keywords TGF-β1Bone mesenchymal stem cellsPanax notoginseng saponinsOsteogenic differentiationhttp://dx.doi.org/10.13039/501100001809National Natural Science Foundation of China81360166Zhou Nuo issue-copyright-statement© The Author(s) 2016 ==== Body Background Panax Notoginseng known as San Qi or Tian Qi is a well-known traditional Chinese medicinal herb which has been used in China since ancient time [1]. The key bioactive component of Panax Notoginseng is Panax Notoginseng Saponins (PNS) [2]. Many studies have been reported that PNS might be a potential treatment choice for various diseases. Recent studies have found that the hypoglycemic and anti-obesity properties of PNS may play an important role in the treatment of diabetes [3]. PNS injection has been reported may provide another choice for patients with angina pectoris (AP), although application of PNS alone showed no significant better or worse effect on AP patients, evidence showed PNS combined with traditional western medicine was a better treatment option for AP patient in improving patients’ clinical symptoms [4]. Nan et al. [5] found that PNS could inhibit phenotype switching of vascular smooth muscle cells (VSMCs) induced by Notch3 silencing in a dose-dependent manner, which might provide new evidence for searching effective drug for amending stability of atherosclerotic disease. PNS was also reported could attenuate colitis in azoxymethane/dextran sulfate sodium (DSS)-induced colitis mouse model [6]. In China, PNS has been applied in treatment of promoting bone fracture healing for hundreds of years [7]. Previous studies has been reported that PNS could stimulate alkaline phosphatase activities and increase the number of osteoblasts in vitro, promote the proliferation of bone marrow mesenchymal stromal cells [8], and is effective in promoting blood circulation to remove stasis [9]. Recent studies suggest that PNS may be a potential therapeutic methods for treating bone nonunion, osteoporosis and osteonecrosis [10]. However, the specific biological mechanisms of osteogenic effect of PNS are not well understood. This study focused on the relationship between osteogenesis of PNS and Transforming Growth Factor-Beta 1 (TGF-β1) signaling pathway, so as to reveal the gene targets of pharmacological action of osteogenesis of panax notoginseng at the molecular level. Thus, it may provide a research foundation for the exploration on the application of new Chinese medicines for various bone-repairs. Methods Experimental animals This study was approved by the animal ethics committee of Guangxi Medical University. Healthy 1 to 2-day newborn New Zealand rabbits (without gender limitations) were supplied by the Experimental Animal Center of Guangxi Medical University in Nanning, China (License No.: SCXK GUI 2009–0002). Reagents and instruments Reagents and instruments were showed as following, including PNS (freeze-drying Xue Shuan Tong injection, Wuzhou Pharmaceutical Group Co. Ltd, Wuzhou, China, No.Z20025652, 150 mg/bottle); L-DMEM culture medium and fetal bovine serum (FBS) (GE Healthcare HyClone™ Cell Culture Co., Logan, USA); 0.25 % trypsin (Shanghai Bi Yun Tian Biotechnological Co. Ltd, Shanghai, China); osteogenesis-induced liquid (Nanning Weierkai Biotechnological Co. Ltd, Nanning, China); Methylthiazoletetrazolium (MTT) assay and percoll separating medium (Sigma-Aldrich Co., St. Louis, USA); rabbit anti-human CD29 polyclonal antibody (FITC) (Beijing Biosynthesis Biotechnological Co. Ltd, Beijing, China); rat anti-human and rabbit CD45 monoclonal antibody (FITC) (AbD Serotec Co., Oxford, UK); rat anti-human and rabbit HLA-DR monoclonal antibody(FITC) (Bio-Rad Co., Hercules, USA); rat anti-human IgG1(FITC) and IgG2b(FITC) (Lianke Biotechnological Co., Hangzhou, China); total RNA-extracted kit (Corning Incorporated Co., New York, USA); reverse transcription kit (TaKaRa Co., Tokyo, Japan); real-time PCR primer (Shanghai Bioengineering Co. Ltd, Shanghai, China); DAB kit (Beijing Zhongshan Biotechnological Co. Ltd, Beijing, China); rat anti-rabbit TGF beta 1 antibody (Novus Biologicals Co., Littleton, USA); goat anti-rabbit IgG (Bio-Techne Co., Minneapolis, USA); Alkaline Phosphatase(ALP) detection kit (Nanjing Jiancheng Biotechnological Co. Ltd, Nanjing, China); alizarin red kit (Shanghai SSS Reagent Co. Ltd, Shanghai, China); CO2 calorstat (Themo Forma Co., Waltham, USA); inverted aberration microscope (Zeiss Co., Oberkochen, Germany); PCR detection system (Bio-Rad Co., Hercules, USA). Isolation, culture and identification of rabbit bone marrow mesenchymal stem cells (BMSCs) Following a protocol approved by the animal ethics committee of Guangxi Medical University, rabbit bone marrow were obtained from tibia and femur of new born New Zealand rabbit by flushing the marrow cavity using a 5 ml syringe filled with L-DMEM-medium (including 10 % FBS and 1 % penicillin-streptomycin) attached with a 23G needle. Then the medium with bone marrow was added into the centrifuge tube (pouring slowly along the tube wall and avoiding shaking) with isopyknic Percoll cell separation solution (1.073 g/ml). After 20-minute 2000 rpm centrifugalization, the nebulous liquid in the middle layer was carefully moved into a new centrifuge tube. And medium was added into the tube for 10-minute 1000 rpm centrifugalization. Then resuspended the cells and regulated the cell concentration to 5 × 105/ml. Afterwards, the cells were collected and cultured in a 50 ml culture flask with L-DMEM-medium supplemented with 10 % FBS and 1 % penicillin-streptomycin at 37°Cin a 5%CO2 humidified incubator. The nonadherent cells were removed after 48 h and the medium was changed every 3 days. The cell passaging was achieved when the cell confluence reached to 80–90 %. After that, the growing status of cells was observed under the inverted aberration microscope. The monoclonal antibody, such as CD29, CD45 and HLA-DR, was added in 3rd passage of cells which grew in a good condition, and the FCM was applied to detect the positive rate of each surface marker in cultured cells. Construction of lentivirus-mediated siRNA interference cell model On the basis of NCBI-GenBank (National Center for Biotechnology Information GenBank) information for TGF-β1 related gene sequence of New Zealand rabbit, we chose the relatively complete mRNA sequence (Oryctolagus cuniculus transforming growth factor, beta 1(TGF-β1), mRNA, NCBI Reference Sequence: XM_00272231312, 1164 bp), as the RNA interference target sequence. Nanning Weierkai Biotechnological Co. Ltd. was responsible for the synthesis of siRNA sequence, including the screening of optimal sequence (LipofectamineTM RNAi MAX transient transfection), the construction of TGF-β1 shRNA lentivirus vector (including ZsGreen green-fluorescent protein), the lentiviral packaging and titering and detection of the virus multiplicity of infection (MOI). The verification of silencing effect via q-PCR The total RNA was extracted from 48-hour transfected BMSCs by Trizol method. Extracted RNA(2 μg) was converted into cDNA via reverse transcriptase and was put into the real-time q-PCR: primer TGF-β1-F: 5’ AAGTCGGCACAGCGTCTATA, TGF-β1-R: 5’ TTGCTGCATTTCTGGTACAG, amplified fragment 150 bp; β-actin-75 bp, β-actin-F: 5’ ACTGGAACGGTGAAGGTGACA, β-actin-R: 5’ TCGGCCACATTGCAGAACT; the reaction condition of PCR was showed as following: 50 °C 2 min; 95 °C 2 min; 95 °C 15 s, 60 °C 32 s, reading plates, 40 circulations; 60 °C-95 °C melting curve analysis. The RQ value should be calculated and repeated three times for each sample. The verification of silencing effect via Western blot The total protein was extracted from 48-hour infected BMSCs by cluster-cracking reaction, and those protein samples were quantified via BCA method. 30 μg protein was transferred to PVDF membrane by SDS-PAGE electrophoresis. After Western blotting and development(compared with GAPDH), the sample was taken a picture via infrared imaging system. The gray value of plaques was analyzed and compared among groups. Measurement of transfected BMSCs proliferation via methylthiazoletetrazolium (MTT) assay After two passages, the cultured nontransfected BMSCs, growing in a good condition, were seeded in 50 ml culture flasks(Corning Co.). When the confluence of cells reached 40 % in the flask bottom, the pre-diluted blank virus solution or virus solution was added into the culture flask. After culturing the cells in incubator at 37 °C in 5 % CO2 for 24 h, we added normal culture medium for cell amplification. Then, the transfected cells were seeded in 96-well culture plates. Comparing with normal cells cultured in the same density, the cell growth curves of the 1st, 3rd, 5th, 7th, 9th and 11th day were drew via MTT method, so as to observe the effect of gene silencing on BMSCs proliferation. Effects of PNS on intracellular ALP activities With the density of 103–104/ml, BMSCs, growing in a good condition, were cultured in a 24-well plate. When the confluence of cells reached to 30 %, we divided the cells into five groups, including blank group, 100 mg group, blank virus group, RNAi group and 100 mg RNAi group. The blank virus group was transfected by blank virus diluents, and RNAi group and 100 mg RNAi group were transfected by virus diluents(silencing method) for 24 h. The medium was changed on the next day of all groups, and cells in 2nd and 5th group were also treated with PNS at 100 mg/L. After the 3rd, 5th and 7th day, the medium was eliminated, and each well was washed by 1 ml PBS once. Then, 500 μl TritonX-100 cell lysis solution was added for 40-minute ice-bath lysis. Under the instruction of ALP detection kit, the OD value of 30 μl lysis solution from each well was detected by ELISA in 520 nm wave length. Finally, the King Unit of ALP in each sample should be calculated in accordance with the formula. Effects of PNS on the formation of calcium nodes of gene silencing BMSCs via alizarin red-staining method With the density of 103–104/ml, BMSCs, growing in a good condition, were cultured in the 6-well plate. When cells were occupied 30 % of the bottom, those wells were divided into six groups, including blank group, positive group, 100 mg group, blank virus group, RNAi group and 100 mg RNAi group. The blank virus group was infected by blank virus diluents and the last two groups by virus diluents(silencing method) for 24 h. Next day, the osteogenesis-induced solution was put into the positive group, and PNS 100 mg/L pure culture solution into both 100 mg group and 100 mg RNAi group, while the pure culture solution into the rest three groups. Each solution above was changed every 2 days to maintain the induction. After the next 21-day culture, the alizarin red was added into each well for staining under the instruction of kit. The quantity of calcium nodes should be calculated via 8 visual fields which were chose from each sample randomly under the microscope, so as to use both mean and standard deviation to make the comparison among those groups. Statistical analysis Quantitative data were presented as mean + SD. Statistical significance was determined by two-tailed student’s t-test or one-way ANOVA followed by the LSD t-test for multiple comparisons. A P-value of 0.05 was considered statistically significant. Results Isolation, culture and identification of rabbit BMSCs After the first medium change for primary cultures at 48 h, a large number of adherent cells with irregular shape were observed. Polygon-like tentacles were found on some of the cells. As the cell number increased rapidly, gradually formed cell colonies were observed. Most cell colonies were composed of cells with a characteristic spindle-like or long spindle-like shape, and arranged in a whorl-like array. Primary cultured cells reached 80 % confluence after 7–8 days of seeding. After 2–3 subcultures, cells reached 80–90 % confluence after 6–7 days of culturing (Fig. 1). The results of flow cytometry showed that the positive rate of P3 BMSCs’ surface marker CD29(integrin family member) was more than 95 %, while the positive rate of CD45(development phase marks of hemocytes) and HLA-DR(surface marks of fibroblasts) was less than 1 %(Fig. 2).Fig. 1 Subcultured BMSCs (P3) Cells reached 80–90 % confluence after 7 days of culturing 100× (a:3d b:7d) Fig. 2 The expression of surface markers of BMSCs by FCM. Positive rate of surface marker: CD29: 97.4 %, CD45: less than 1 %, HLA-DR: less than 1 % RNA interfering BMSCs cell model The interference efficiency of siRNA was detected and the optimal sequence(5’-3’) was screened as following: GCUUCAGAUCCACAGAGAAdTdT and antisense strand(AS) UUCUCUGUGGAGCUGAAGCdTdT. After the fluorescence observation, MOI was detected as 100 and the infection rate of cells was 80 % (Figs. 3 and 4).Fig. 3 Observation of BMSCs after 48-hour virus infection 200× (a. fluorescence photo, b. normal photo, c. overlapping photo) Fig. 4 Observation of BMSCs after 72-hour virus infection 200× (a. fluorescence photo, b. normal photo, c. overlapping photo) Results of TGF-β1 gene expression verification by real-time RT-PCR after RNAi As shown by figures above (Figs. 5, 6, 7, 8, 9), the amplification curve and melting curve of each sample had repeatability, as well as both were smooth unimodal curves without irregular peaks. It was showed that both target genes and reference genes were amplified successfully. From the result of qRT-PCR (Fig. 10), the expression of TGF-β1mRNA of the RNAi group decreased significantly by comparing with the normal group. However, the expression of blank virus group was the same as the normal group. It suggested that siRNA-carried virus had an inhibitory effect on the gene expression of TGF-β1, while the blank virus had no effects on the expression of TGF-β1mRNA.Fig. 5 Electrophoretogram of extracted total RNA (1. normal cells, 2.blank virus-infected cells, 3. RNAi cells) Fig. 6 Amplification curve of TGF-β1 gene in each group Fig. 7 Melting curve of TGF-β1 gene in each group Fig. 8 Amplification curve of β-actin in each group Fig. 9 Melting curve of β-actin in each group Fig. 10 Comparison of TGF-β1 mRNA expression of normal group, blank virus group and siRNA group after virus-transfection Results of TGF-β1 gene expression verification by Western blot after RNAi According to the figure and table above, the virus transfection in RNAi group reduced the protein expression of TGF-β1 significantly by the comparison of the normal group; gene silencing was significant while there was no effect in the blank virus group (Fig. 11, Table. 1).Fig. 11 Western blot of TGF-β1 protein of normal group, blank virus group and RNAi group Table 1 Gray level results Groups Normal group Blank Virus group RNAi group TGFβ1 16983.87 16853.84 4280.67 GAPDH 20169.29 22486.75 22750.20 The growth curve of BMSCs after transfection According to the figure above, after blank virus transfection and virus transfection, the growth curve of transfected cells was similar to that of normal cells, which showed that the transfected virus and gene silencing had little effects on cell proliferation (Fig. 12).Fig. 12 Growth curve of normal BMSCs and siRNA-transfected BMSCs The effect of PNS on ALP activity of gene silencing BMSCs According to the table above(Table. 2), the activity of ALP in 100 mg group significantly increased by comparing with that of other groups in each introduction time(**P < 0.01); the ALP activity level in 100 mg RNAi group was higher than that of control group(*P < 0.05), especially on day 3(**P < 0.01), which might indicate that PNS had a promoting effect on osteogenesis of TGF-β1 gene silencing cells; the ALP activity level of 100 mg RNAi group was lower than that of 100 mg group, and there was a significant difference in the day 5 and day 7(△P < 0.05), while the ALP level of 100 mg RNAi group was higher than that of RNAi group, and a significant difference was found in all three induction time (□P < 0.01); comparing with the normal group, there was no statistical significance of the effect on intracellular ALP level among normal group, blank virus group and RNAi group.Table 2 Intracellular ALP activity after RNAi in all groups x¯±s,n=4 Group 3 days 5 days 7 days Normal group 0.490 ± 0.095 1.013 ± 0.334 1.625 ± 0.241 100 mg group 1.031 ± 0.197** 1.675 ± 0.188** 2.578 ± 0.281** Blank virus group 0.388 ± 0.073 0.868 ± 0.260 1.543 ± 0.329 RNAi group 0.358 ± 0.093 0.833 ± 0.079 1.413 ± 0.156 100 mg RNAi group 0.916 ± 0.214**□ 1.295 ± 0.103*△□ 2.144 ± 0.255*△□ *: p<0.05, other 4 groups compare to Normal group **: p<0.01, other 4 groups compare to Normal group △: p<0.05, 100mg RNAi group compare to 100mg group □: p<0.01, 100mg RNAi group compare to RNAi group The effect of PNS on formation of calcium nodes of TGF-β1 gene silencing BMSCs by Alizarin red staining After 21-day induced cultivation, it was observed that there were some white bulges(nodes) with different densities on the bottom of culture plate when the culture solution was absorbed by those induced cells. After the alizarin red staining, it was showed that calcium nodes turned to be orange or dark red, and the counting results of calcium nodes in each group presented as following (Fig. 13).Fig. 13 Result of Alizarin red staining of five groups (a. normal group, b. positive group, c. 100 mg group, d. blank virus group, e. RNAi group, f. 100 mg RNAi group) According to the table above (Table 3), there were several calcium nodes in 100 mg group without silencing after induction, which had a significant difference with the comparison of the normal group(**P < 0.01); there was no significant difference of the calcium nodes count among the blank virus group, the RNAi group (without induction factors) and the normal group(P > 0.05); however, There were also some calcium nodes in 100 mg RNAi group but it was less than that in 100 mg group, which had a significant difference(▲P < 0.01). It was showed that gene silencing had an inhibitory effect on induced-osteogenic effects of PNS in a certain degree.Table 3 Comparison of calcium nodule count in each group x¯±s Groups n Calcium nodes count Normal group 8 2.63 ± 0.52 Positive group 8 120.88 ± 13.80** 100 mg group 8 97.63 ± 12.14** Blank virus group 8 2.38 ± 0.52 RNAi group 8 3.63 ± 1.41 100 mg RNAi group 8 65.38 ± 11.17**▲ **: p<0.01, other 5 groups compare to Normal group ▲: p<0.01, 100mg RNAi group compare to 100mg group Discussion Bone mesenchymal stem cells(BMSCs) is a bone marrow-derived mutipotent stem cells, which can differentiate into several somatic cells under a certain circumstances, such as: osteoblasts, chondroblasts, fibroblasts, adipocytes and endothelial cells [11]. Since BMSCs are easily obtained, possess multiple differentiate ability and strong proliferative ability, and has a positive response to various bone growth factors, they has been considered as an important seed cells in bone tissue engineering [12]. The Percoll density gradient centrifugation was also applied in the experiment [13, 14]. The high purity of BMSCs was achieved, and its rapid growth was in accordance with the typical characteristic of BMSCs [15]. With reference to the surface markers of BMSCs reported in previous studies [16, 17], in our study, the expression of surface markers of rabbit BMSCs, such as CD29, CD44, CD105, CD61 and CD105, was positive, while those of hematopoietic progenitor cells, such as CD34, CD45, CD11a and HLA-DR, was negative. Our study focused on the detection of three surface markers, including CD29, CD45 and HLA-DR. It was found that CD45 and HLA-DR were negative, while CD29 was positive, which was in accordance with literature reports. As a cytokine for regulating bone reconstruction, TGF-β1 plays an important role in fracture healing, new bone regeneration as well as the balance between resorption and formation during bone reconstruction. It was reported [18, 19] that there were five positive effects on regulating osteogenesis of TGF-β1, including the transformation of mesenchymal cells, the differentiation of osteoblasts and chondroblasts, the formation and excretion of extracellular matrix, the repairing and reconstruction of bone tissues via the regulation of osteoclasts, as well as the regulation of other hormones and growth factors. TGF-β1 was likely to be the target of osteogenesis-promoting effect for most drugs. It was reported that PNS with a certain concentration had a promoting effect on osteogenic differentiation of BMSCs of rats and rabbits in vitro, which was related to the down-regulation of expressions of RANKL, reduction of formations and activity of osteoclasts [20], as well as restriction of adipogenic differentiation of BMSCs. In this study, it was found that a certain dosage of PNS had an up-regulating effect on the gene expression of TGF-β1, which suggested that the osteogenesis of PNS was likely related to the up-regulation of gene expression of TGF-β1. Our study focused on not only whether PNS still had osteogenic effects on RNAi cells, but the relationship between PNS and the expression of TGF-β1. RNA interference (RNAi) can transfect target gene-homologous small interferencing RNA(siRNA) into target cells in a certain way. After that, the specific mRNA which had the homologous sequence in target cells was induced to degrade, so as to restraint the protein translation process of target genes. It is similar to the gene knockout but the gene is not to be removed forever, so it is also called as “gene silencing”. The greatest advantage of RNAi is the blockage of both transmission pathway of cell signalings and target spots by aiming at a specific factor. In this experiment, siRNA-carried recombinant pLVX-TGF-β1 shRNA-ZsGreen plasmids was established and packed as lentivirus. Then, the lentivirus was applied to infect target cell BMSCs, which made the siRNA to express in BMSCs and degrade the TGF-β1mRNA. As a result, the gene silencing of TGF-β1 was achieved in BMSCs. The stability of silencing effect was verified via FQ-PCR and Western blot. Osteoblasts can secrete alkaline phosphatase(ALP), and the activity of ALP elevates when the activity of osteoblasts enhances. High expression of ALP is the early marker of differentiation and maturation of osteoblasts [21], and the deposition of calcium salt and formation of calcium nodes can manifest the further differentiation and maturation of osteoblasts [22]. Therefore, the observation of the drug effect on activity of ALP and mineralization of osteoblasts is an essential approach for the research on osteogenetic effect of drugs. In this study, the blank virus group and the pure RNAi alone group were designed to eliminate the interruption of results caused by virus and gene RNAi. According to the results of intracellular ALP activity detection, the activity of ALP in 100 mg group was higher than that in the normal group, which indicated that PNS might promote the osteogenesis of osteoblasts; the activity of ALP in 100 mg RNAi group was higher than that in normal group but lower than that in 100 mg group, which indicated that PNS might still promote the osteogenesis of osteoblasts after TGFβ1 silencing but the osteogenic potential decreased significantly; and no significant differences of intracellular ALP level was found among normal group, blank virus group and RNAi group. From the results of Alizarin red staining, calcium nodes was rare in the normal group, which suggested that only few cells differentiated into osteblasts. Moreover, the results also showed that the cultured BMSCs had a poor ability in osteogenic differentiation, which was in accordance with the feature of multi-directional differentiation potential of BMSCs. The number of calcium nodes was few in both blank virus group and RNAi group, and it was no significant difference by comparing with the control group. This was also showed that both blank virus and gene silencing could not promote the osteogenic differentiation of BMSCs. The number of calcium nodes in the 100 mg RNAi group was more than that in normal group but fewer than that in 100 mg group. This results might suggest that TGF-β1 RNAi has an inhibitory effect on osteogenic differentiation potential of PNS, and further indicate that the osteogenic potential of PNS may partly mediated by TGFβ1, and was related to the mechanism of TGF-β signaling pathway. However, the inhibitory effect of gene silencing on osteogenic differentiation could not reduce the number of calcium nodes to the level of normal cells and silencing cells. Thus, this suggests that PNS had a promoting effect on osteogenesis in a certain degree, and PNS might exert its osteogenic effect via other cytokines and other signaling pathways, such as bone morphogenetic proteins(BMPs) act upstream and Smads signaling act downstream. The mechanism might be related to various monomer compositions contained in PNS. However, its efficacy was extensive and uncertain because the specific function of each composition was not revealed entirely. In addition, the mechanism of osteogenesis might involve in several cell signal pathways, including TGF-β pathway, Notch pathway [23, 24], Wnt pathway [25, 26], Hedgehog pathway [27] and MAPK pathway [28]. Further research will be launched to identify which pathway and target plays a dominant role in promoting osteogenesis. Conclusion In summary, in this study we found that silencing TGF-β1 gene could interrupt the osteogenic effect of PNS. And we suggest that PNS may have a promoting effect on osteogenic differentiation of rabbits’ BMSCs in vitro by up-regulating the gene expression of TGF-β1. However, further studies, including in vitro and in vivo studies, should be conducted to better verify the effect on osteogenic differentiation of PNS. Abbreviations ALPAlkaline phosphatase BMPsBone morphogenetic proteins BMSCsBone marrow mesenchymal stem cells FBSFetal bovine serum MTTMethylthiazoletetrazolium PNSPanax notoginseng saponins RNAiRNA interference TGF-β1Transforming growth factor beta 1 Acknowledgements No acknowledgements. Funding This study was funded by National Natural Science Foundation of China, Award Number: 81360166. Availability of data and materials The data will not be shared, for no related and specific data is included in our manuscript. Authors’ contributions YW: Study design, collection and assembly of data, data analysis and interpretation, literature research, writing the article. XH: Study design, collection and assembly of data, data analysis and interpretation. YT: collection of data, data analysis. HL: collection of data, statistical analysis. NZ: Study concept and study design, guarantor of integrity of entire study, critical revision of the article, final approval of article. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethic approval and consent to participate This study was approved by the animal ethics committee of Guangxi Medical University. (License No.: SCXK GUI 2009–0002). ==== Refs References 1. Gao XM Chinese pharmacy 2007 Beijing China Press of Traditional Chinese Medicine 296 7 2. Dong TT Cui XM Song ZH Zhao KJ Ji ZN Lo CK Tsim KW Chemical assessment of roots of Panax notoginseng in China: regional and seasonal variations in its active constituents J Agric Food Chem 2003 51 4617 23 10.1021/jf034229k 14705886 3. Uzayisenga R Ayeka PA Wang Y Anti-diabetic potential of Panax notoginseng saponins (PNS): a review Phytother Res 2014 28 510 6 10.1002/ptr.5026 23846979 4. Yang X Xiong X Wang J Sanqi panax notoginseng injection for angina pectoris Evid Based Complement Alternat Med 2014 2014 963208 24693328 5. Liu N Shan D Li Y Chen H Gao Y Huang Y Panax notoginseng saponins attenuate phenotype switching of vascular smooth muscle cells induced by Notch3 silencing Evid Based Complement Alternat Med 2015 2015 162145 26539217 6. Wen XD Wang CZ Yu C Zhao L Zhang Z Matin A Wang Y Li P Xiao SY Du W He TC Yuan CS Panax notoginseng attenuates experimental colitis in the azoxymethane/dextran sulfate sodium mouse model Phytother Res 2014 28 892 8 10.1002/ptr.5066 24142591 7. Yang J Ren T Li XD Qiao ZG The experiment study of total saponins of Panax notoginseng on promoting fracture healing J. Med. Res. 2010 39 69 71 8. Li XD Wang JS Chang B Chen B Guo C Hou GQ Huang DY Du SX Panax notoginseng saponins promotes proliferation and osteogenic differentiation of rat bone marrow stromal cells J Ethnopharmacol 2011 134 268 74 10.1016/j.jep.2010.11.075 21167926 9. Liu Y Hao F Zhang H Cao D Lu X Li X Panax notoginseng saponins promote endothelial progenitor cell mobilization and attenuate atherosclerotic lesions in apolipoprotein E knockout mice Cell Physiol Biochem 2013 32 814 26 10.1159/000354484 24080960 10. Du WX Duan SF Yu XL Yin LM Panax notoginseng saponins suppress radiation-induced osteoporosis by regulating bone formation and resorption Phytomedicine 2015 22 813 19 10.1016/j.phymed.2015.05.056 26220628 11. Bianco P Rininucci M Gronthos S Robey PG Bone marrow stromal stem cells: nature, biology, and potential applications Stem Cells 2001 19 180 92 10.1634/stemcells.19-3-180 11359943 12. Matsumura N Ochi K Ichimura M Mizushima T Harada H Harada M Study on free radicals and panceaticibrosis, pancereatic fibrosis induced by repeated injections of superoxide dismutase inhibitor Oanceras 2001 22 53 7 13. Ma L Liu DJ Li DT Liu PT Effects of different separate methods and culture conditions on growth proliferation and biological characteristics of rabbit bone marrow mesenchymal stem cells J Clin Rehabil Tissue Eng Res 2008 38 7401 06 14. Li X Zhang Y Qi G Evaluation of isolation methods and culture conditions for rat bone marrow mesenchymal stem cells Cytotechnology 2013 65 323 34 10.1007/s10616-012-9497-3 23011741 15. Friedenstein AJ Chailakhyan RK Gerasimov UV Bone marrow osteogenic stem cells: in vitro cultivation and transplantation in diffusion chambers Cell T issue Kinet 1987 20 263 67 16. Wang XF Tong PJ Jin HT Approaches to isolate, culture and identify the rabbit marrow mesenchymal stem cells J Tradit Chinese Orthop Traum 2011 23 3 6 17. Liu Q Shi Y Wang H Experimental study on isolation, culture and identify of rabbit bone marrow-derived mesenchymal stem cells in vitro J Tissue Eng Reconstr Surg 2013 17 1366 69 18. Erlebacher A Filvaroff EH Ye JQ Derynck R Osteoblastic responses to TGF-β during bone remodeling Mol Biol Cell 1998 9 1903 18 10.1091/mbc.9.7.1903 9658179 19. Schuldiner M Yanuka O Itskovitz-Eldor J Melton DA Benvenisty N Effects of eight growth factors on the differentiation of cells derived from human embryonic stem cells Proc Natl Acad Sci U S A 2000 97 11307 12 10.1073/pnas.97.21.11307 11027332 20. Wang DW Pan H Li HB Wang HX Chen YP Influence of total saponins of Panax Notoginseng on the alcohol-induced differentiation of bone marrow stromal stem cells J Clin Rehabil Tissue Eng Res 2008 12 1410 13 21. Yao XM Cheng Y Fang F Experimental study on YIGUTANG containing serum on osteoblast proliferation and ALP expression J Zhejiang University Tradit Chinese Med 2007 31 1 2 22. Maniatopoulos C Sodek J Melcher AH Bone formation in vitro by stromal cells obtained from bone marrow of young adult rats Cell Tissue Res 1998 254 317 25 3197089 23. Hilton MJ Tu X Wu X Bai S Zhao H Kobayashi T Kronenberg HM Teitelbaum SL Ross FP Kopan R Long F Notch signaling maintains bone marrow mesenchymal progenitors by suppressing osteoblast differentiation Nat Med 2008 14 306 14 10.1038/nm1716 18297083 24. Dishowitz MI Mutyaba PL Takacs JD Barr AM Engiles JB Ahn J Hankenson KD Systemic inhibition of canonical Notch signaling results in sustained callus inflammation and alters multiple phases of fracture healing PLoS One 2013 8 e68726 10.1371/journal.pone.0068726 23844237 25. Lin GL Hankenson KD Integration of BMP, Wnt, and Notch signaling pathways in osteoblast differentiation J Cell Biochem 2011 112 3491 501 10.1002/jcb.23287 21793042 26. Chen B Li XD Liu DX Wang H Xie P Liu ZY Hou GQ Chang B Du SX Canonical Wnt signaling is required for Panax Notoginseng Saponin-mediated attenuation of the RANKL/OPG ratio in bone marrow stromal cells during osteogenic differentiation Phytomedicine 2012 19 1029 34 10.1016/j.phymed.2012.06.002 22818895 27. Plaisant M Fontaine C Cousin W Rochet N Dani C Peraldi P Activation of Hedgehog signaling inhibits osteoblast differentiation of human mesenchymal stem cells Stem Cells 2009 27 703 13 10.1634/stemcells.2008-0888 19096040 28. Hu Y Chan E Wang SX Li B Activation of p38 mitogen-activated protein kinase is required for osteoblast differentiation Endocrinology 2003 114 2068 74 10.1210/en.2002-220863 12697715

PMC005xxxxxx/PMC5000421.txt

==== Front BMC Evol BiolBMC Evol. BiolBMC Evolutionary Biology1471-2148BioMed Central London 73110.1186/s12862-016-0731-zResearch ArticleStructuring evolution: biochemical networks and metabolic diversification in birds http://orcid.org/0000-0002-4487-6915Morrison Erin S. esmorr@email.arizona.edu Badyaev Alexander V. abadyaev@email.arizona.edu Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, AZ USA 25 8 2016 25 8 2016 2016 16 1 16815 2 2016 1 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Recurrence and predictability of evolution are thought to reflect the correspondence between genomic and phenotypic dimensions of organisms, and the connectivity in deterministic networks within these dimensions. Direct examination of the correspondence between opportunities for diversification imbedded in such networks and realized diversity is illuminating, but is empirically challenging because both the deterministic networks and phenotypic diversity are modified in the course of evolution. Here we overcome this problem by directly comparing the structure of a “global” carotenoid network – comprising of all known enzymatic reactions among naturally occurring carotenoids – with the patterns of evolutionary diversification in carotenoid-producing metabolic networks utilized by birds. Results We found that phenotypic diversification in carotenoid networks across 250 species was closely associated with enzymatic connectivity of the underlying biochemical network – compounds with greater connectivity occurred the most frequently across species and were the hotspots of metabolic pathway diversification. In contrast, we found no evidence for diversification along the metabolic pathways, corroborating findings that the utilization of the global carotenoid network was not strongly influenced by history in avian evolution. Conclusions The finding that the diversification in species-specific carotenoid networks is qualitatively predictable from the connectivity of the underlying enzymatic network points to significant structural determinism in phenotypic evolution. Electronic supplementary material The online version of this article (doi:10.1186/s12862-016-0731-z) contains supplementary material, which is available to authorized users. Keywords Network structureMetabolic pathwaysPhenotypic diversityDavid and Lucille Packard Foundationn/aAmherst College Graduate Fellowshipn/aissue-copyright-statement© The Author(s) 2016 ==== Body Background Only a small proportion of theoretically possible changes seemed to be realized in phenotypic evolution and diversification, with some outcomes appearing recurrently whereas others are seemingly forbidden [1–5]. Such determinism and predictability of phenotypic outcomes is surprising considering the dimensionality of the genome, the proteome, and the developmental dynamics linking them and point to the existence of constraints in phenotypic variation. Theoretical and empirical studies have suggested that such constraints may be a reflection of the connectivity of the network of interactions among elements such as genes, proteins, enzymes and metabolites (defined here as a deterministic network) caused by genomic or developmental epistasis [1, 6–11], internal integration during development [12–15], and physical stability or historical contingency of gene and protein associations [16–22]. Direct examination of the correspondence between opportunities for diversification imbedded in such networks and realized phenotypic diversity is needed to illuminate the structural properties of networks that delineate phenotypic diversity. Phenotypic diversification on a deterministic network is the result of the gain or loss of elements and interactions that convey different fitness [1, 3, 22]. Mechanistically, the evolutionary representation and variability of network elements tends to be associated with their topological positions [23–28]. In particular, two structural properties of networks – the number of reactions per element, which represents the connectivity of the network, and the number of reactions that separate elements in a network, which defines the length of pathways between elements in the network – provide distinct ways by which elements and interactions in the network are gained or lost and result in different patterns of phenotypic diversification (Fig. 1) [29–33].Fig. 1 The structure of a deterministic network and potential evolutionary trajectories. The possible interactions (arrows) between elements (small circles) represent potential opportunities for diversification on a deterministic network (shown in grey). The black, purple and orange shaded portions of the network show examples of different expressed networks, with each color denoting a different functional module made up of different elements and interactions. (a) Under the pathway diversification scenario, elements with the most interactions (higher connectivity) should be most conserved across networks, and the number and identity of the interactions associated with these connected elements should differ across networks. (b) Under the pathway elongation scenario, elements at the beginning of a sequential pathway of reactions should be the most conserved across networks, and the pathway length (the number of reactions that separate one element from another) and elements located further away from the start of the pathway should differ between networks. (c) Under the module diversification scenario, differences between networks are the result of the gain or loss of entire modules (unique groups of functionally coupled elements and interactions) and the gain or loss of elements would not be related to their connectivity or to their distance from a starting element in a pathway Greater connectivity of an element – the number of direct interactions it has with other elements in a network – enables an evolving lineage to include different elements that both directly interact with the same element [34–36]. In this mode of network diversification (hereafter pathway diversification), the gain of different interactions associated with the same element represents the start of divergent pathways comprised of unique elements and interactions (Fig. 1a). For example, in metabolic networks, the use of different enzymatic reactions from the same substrate metabolite produces different products resulting in distinct metabolic pathways. Theory and empirical data suggest that metabolic and protein networks commonly evolve by the preferential attachment of new enzymatic reactions or protein interactions to the most connected elements in these networks [24, 34, 37, 38]. Correspondingly, the genes underlying proteins and enzymes with greater connectivity tend to be represented in a greater number of taxa, have longer evolutionary persistence and lower rates of evolutionary change than elements with fewer direct interactions in a network [23, 39, 40]. Thus, the divergence among species’ networks should be driven by the gain or loss of interactions among highly connected elements, whereas the connected elements themselves should be conserved across species. Differences in the number of interactions that start from these conserved elements should be reflected in differences in the overall network connectivity (number of interactions per element) across species’ networks, because a greater number of opportunities exist for species to express different interactions at densely connected compounds. If pathway diversification causes divergence among species’ networks, then we expect differences in the elements and interactions present across species networks to increase with the differences in the connectivity of their networks, such that interactions and elements associated with the most connected compounds in the network should vary the most across species. The length of pathways – the number of interactions (e.g., enzymatic reactions) that connect elements in a network – enables an evolving lineage to express different elements and reactions along the same pathway. This mode of network diversification (hereafter pathway elongation), results from differences in the number of sequential interactions from the same starting element (Fig. 1b). Most genes, proteins, and metabolites are regulated by multistep interactions [35, 41] and thus in most cases, the activation or expression of an element is dependent on several prior interactions. Changes in interactions at the beginning of a pathway may prevent the expression of interactions located further downstream in the pathway and result in shorter pathways and the loss of elements. Alternatively, the addition of a new interaction to the end of a pathway can increase the length of the pathway and produce a novel product. Models of network growth and empirical results suggest that most of the change in networks occurs at their periphery, such that terminal elements are most likely to be gained or lost, whereas the central or upstream elements are the most conserved [39, 42–44]. Longer pathways between elements in a network therefore provide more opportunities for the use of different numbers of sequential reactions from the same starting element, such that some species networks only express the intermediate elements that lie along a pathway of interactions from one element to another and the final product is never expressed. If network diversification is driven by differences in the elongation of a sequence of interactions among species, then we expect species’ networks to have different pathway lengths from the same starting element. The difference in the length of the pathways among species’ networks should be reflected in the diversification among the elements and interactions present in each species. In this case, the elements located at the beginning of pathways should be conserved across networks, and species’ networks should diverge more from each other at elements located closer to the ends of potential pathways. Networks are often organized in discrete functional modules in which a group of metabolites, enzymes, genes, or proteins interact more often with each other than with other elements in the network [45, 46]. Functional modules play an important role in the evolvability of organisms [47–51]. Empirical studies have shown that genes in the same regulatory modules tend to be co-expressed [52–55], resulting in similar evolutionary rates of proteins in the same modules [56, 57]. Additionally, genes that underlie within-module enzymatic reactions have similar rates of evolutionary gain and loss (e.g., [58, 59]), such that multiple enzymatic reactions that comprise a pathway are gained or lost together. Therefore, another mode of network divergence among species could be the result of the gain or loss of complete functional modules (hereafter module diversification) (Fig. 1c). If this is the case, then species should differ in modules they express, and neither the connectivity of elements nor the length of a pathway between elements in a network should be related to the differences in species’ networks. Here we examined the extent to which the structure of enzymatic reactions in the global carotenoid network – that comprises all of the documented enzymatic reactions among naturally occurring carotenoids (Additional file 1a) – is associated with patterns of avian diversification in carotenoid-producing metabolic networks. The connectivity and topology of enzymatic reactions of the global carotenoid network have evolved largely in the context of bacterial evolution (e.g., [60, 61]) and subsets of this global network are utilized in the carotenoid metabolism of all lineages studied to date, such as fungi, plants, insects and animals (e.g., [62, 63]). Here we studied the patterns of utilization of this network associated with the production of carotenoid pigmentation in the plumage and integument of 250 bird species. Specifically, we were interested in the effect of the structure of the global metabolic network on the frequency of occurrence of individual carotenoid compounds and reactions across species. In birds, metabolism of carotenoids expressed in feathers and integument necessarily starts with the consumption of dietary carotenoids (e.g., [64, 65]). This property of avian carotenoid biosynthesis allows for the identification of the starting points of metabolic pathways in species’ networks and provides an opportunity to distinguish the effects of pathway diversification from the effects of pathway elongation and module diversification on network divergence across species. In birds, pathway diversification from the same highly connected compounds, pathway elongation starting at the same dietary compounds, or the consumption of different dietary compounds representing different functional modules in the network could produce evolutionary transitions across species’ networks. In the global carotenoid network, opportunities for pathway diversification and elongation vary across metabolic pathways that start at different dietary carotenoids (Figs. 2 and 3). Additionally, the consumption of different dietary compounds results in access to different enzymatic reactions and metabolites that could comprise different functional modules (Fig. 2). Here, we first mapped species’ carotenoid networks onto the global avian carotenoid metabolic network [66] and examined whether differences in enzyme connectivity or relative pathway position of individual carotenoid compounds were associated with their evolutionary representation among species. We then repeated these analyses for biochemical modules of interconnected elements and examined their evolutionary representation in relation to their structural properties. We examined the relative contribution of enzymatic connectivity, metabolic pathway lengths, and module representation on network divergence and identified the structural properties of both individual compounds and modules associated with diversification hotspots on the global carotenoid network. We discuss the extent to which the structure of the carotenoid metabolic network can be used to understand and predict patterns of realized phenotypic diversity.Fig. 2 Schematics of the connected global enzymatic network of carotenoid compounds (66 compounds, 97 enzymatic reactions) found in species under this study (Additional files 1 and 2). Green nodes show dietary carotenoids. The distinct shaded areas represent the module assignments for the 53 compounds expressed at least once across species’ networks using simulated annealing [71, 72]. The numbers in the squares for each module denote the module number that corresponds to the module assignments for each compound in Additional file 1c Fig. 3 Structural diversity of carotenoid compounds in the avian space of the global carotenoid metabolic network (Fig. 2). Compounds differ in connectivity (reactions per compound), shown in the histogram on the left, and their distance (number of reactions) from the four main dietary (starting) compounds (lutein, zeaxanthin, β-carotene, β-cryptoxanthin), shown in the graph on the right Methods Data collection and metabolic network construction The global carotenoid biosynthesis network includes all of the enzymatic reactions that occur among naturally-occurring carotenoids in bacteria, plants, fungi and animals (Additional file 1a, [66]). This network delineates biochemical pathways of carotenoid biosynthesis based on the chemical properties of the compounds. We collected an exhaustive list of all the carotenoid compounds and reactions documented in birds (n = 339 species), using carotenoids that are found in plumage, integument (bill, tarsi, skin), plasma, liver, fat, feces, retina, and seminal fluid, or are known to be consumed in the diet (Additional file 1b; data current as of July 2015). The chromatography and mass spectrometry methods that are listed in Additional file 1b document the presence or absence of specific compounds against known standards. All of the distinct compounds identified in the species of birds were then used to construct the “avian subset” of the global carotenoid metabolic network, consisting of 66 carotenoids and 97 enzymatic reactions (Fig. 2). The global metabolic network was then used as a template to construct 250 species-specific carotenoid metabolic networks between known dietary carotenoid compounds (the upstream elements of carotenoid metabolic networks in birds), metabolized compounds (e.g., circulating in plasma or found in other organism tissues), and the expressed compounds identified from species’ plumage and integument (Additional file 2). Briefly, after mapping compounds found in the diet, plasma, and plumage or integument of species under this study on the “avian space” of the global carotenoid biosynthesis network (Fig. 2), we recorded biochemical pathways that link dietary, intermediate and plumage-expressed compounds for each species (Additional files 1b and 2; details and justification in Badyaev et al. [66], which also see for phylogenetic analyses of avian carotenoid networks). For species that had no known dietary or intermediate compounds (but not both), missing compounds and reactions were assigned based on the mapping of the species’ known compounds and reactions on the global network and recording all biochemically possible connections (e.g., between a known dietary and a known expressed compound or between a known intermediate and a known expressed compound and a possible dietary compound). Networks were not built for species if the carotenoids expressed in their plumage or integument were unknown even when all other components of the network were documented. Thus, not all of the compounds and reactions in the avian subset of the global carotenoid metabolic network (Additional file 1a, Fig. 2) were present in the species-specific networks. In the 250 species-specific complete networks that were constructed, 53 compounds and 81 enzymatic reactions occurred at least once. Species under this study represent eleven avian orders (Anseriformes, Charadriiformes, Ciconiiformes, Columbiformes, Galliformes, Passeriformes, Pelecaniformes, Phaethontifromes, Phoenicopteriformes, Piciformes, Trogoniformes) and span over 110 MYA of avian carotenoid diversification (Fig. 4a, 4b, 4c, 4d and 4e, Additional file 3) [66].Fig. 4 (a) Consensus tree of the non-passerine species in this study showing, for each species’ metabolic network, the number of compounds (number of bars; green bars –distinct dietary carotenoids; yellow, orange and red bars – metabolically derived compounds), average degree (y-axis of the legend), number of modules (number of bar groups), pathway length (x –axis of the legend, number of enzymatic reactions from the closest dietary compound). The tree is a part of a majority rule consensus tree of 249 species based on 1,000 randomly sampled trees from the Hackett All Species pseudo posterior distribution from Jetz et al. [116] (Additional file 3). The other subsets of the tree, show in the inset in the lower left corner, are displayed in Figures 4b, 4c, 4d, and 4e Fig. 4 (b) Consensus tree of the suboscine species under this study. Legend in Figure 4a Fig. 4 (c) Consensus tree of a subset the oscine species under this study. Legend in Figure 4a Fig. 4 (d) Consensus tree of a subset of the oscine species under this study. Legend in Figure 4a Fig. 4 (e) Consensus tree of a subset the oscine species under this study. Legend in Figure 4a Metabolic distance and modularity in networks We used a modified metabolic distance based on the Jaccard distance [67] and Rodrigues and Wagner [68] to calculate the fraction of reactions and compounds differing between any two metabolic networks. Species’ networks were coded based on the presence of compounds and reactions in the avian subset of the global carotenoid metabolic network. The uncorrected P-distance is the fraction of the number of compounds and reactions that differ between each pair of networks (d) out of the total number of compounds and reactions in the global network (NG): P=dNG The pairwise P-distances were computed in Mesquite (version 3.03) [69] using the PDAP:PDTREE (version 1.16) package [70]. The metabolic distance (D) between networks represents the fraction of compounds and reactions in which two networks differ out of the total number of compounds and reactions that occur in each of the networks: D=dN1+N2 where N1 and N2 are the total number of compounds and reactions in networks S1 and S2, respectively. The 53 compounds expressed in the global carotenoid network at least once among the species’ networks were partitioned into ten structurally defined modules based on the density of the compounds’ enzymatic interconnectivity using the simulated annealing program netcarto (https://amaral.northwestern.edu/resources/software/netcarto) [71, 72]. This approach to module partitioning has previously been used to reliably assign metabolites to the correct functional pathway based only on the structural properties of the metabolites [71]. In the avian carotenoid metabolic network, the modules are partitioned by different dietary compounds; seven of the ten modules include at least one starting, upstream dietary compound. For module assignments of the individual compounds in the global carotenoid metabolic network refer to Fig. 2 and Additional file 1c. Network structural measurements For each compound in the avian carotenoid network (Fig. 2) we calculated the number of directly linked enzymatic reactions [73] and the distance from a dietary compound (minimum number of reactions between a compound and any of the dietary compounds in the network) to represent the connectivity and the pathway position of each compound, respectively. The connectivity (C) of each of the modules in the global network and each of the species’ networks was the average number of reactions per compound: C=rn where r is the total number of reactions in the module or network and n is the total number of compounds in the module or network. The diameter of each of the species’ networks is the shortest distance (number of reactions) between the two most distant dietary and expressed compounds in the network. The diameter of each of the modules in the global network is the fewest number of reactions between the two most distant compounds in the module. Both the connectivity of the species' networks and the modules and the diameter of the modules were computed using Cytoscape 2.8.2 [74] with NetworkAnalyzer 2.7 [75, 76] and RandomNetworks 1.0 [77]. Species representation and realized phenotypic diversification The species representation of a compound or reaction is the number of species that have this compound or reaction (e.g., [39]). Whereas species representation characterizes the evolutionary representation of a compound, it does not include information on species’ phylogenetic relationships, and instead enables the examination of metabolic network evolution from a structural, rather than historical perspective (e.g., [39]). In a companion study we found that the phylogenetic relationships among the species in this study were not reflected in the similarity of their biochemical networks; the small biochemical space on which birds diversify and the structure of the biochemical network instead leads to recurrent convergence of distantly related and ecologically distinct taxa in metabolic networks [66]. Having examined the historical sequence of exploration of the global carotenoid network by extant avian species in that study, here we explore whether the structure of the global carotenoid network is reflected in the pattern of network exploration across avian lineages. Several other studies have taken similar approaches to compare structural features of metabolic networks across species of bacteria, eukaryotes, and archaea independently of their phylogenetic relationships (e.g., [24, 35, 78]). The realized diversification (R) of an enzymatic reaction was measured as the fraction of species that do not have a reaction among all of the species that have the substrate compound for the reaction (nc), where nr is the number of species that have the reaction: R=nc−nrnc An enzymatic reaction with a realized diversification score of zero represents a location in the network with little or no divergence between species’ networks along that part of a pathway; meaning that the enzyme is conserved across species that also have the enzyme’s substrate compound. The realized diversification of an enzymatic reaction with a score close to 1 represents a point of major divergence between species (i.e., the enzyme is only present in a small fraction of the total species that have the enzyme’s substrate compound). Results Global carotenoid network structural properties and diversity of species’ networks Connectivity and the distance from dietary carotenoids of compounds varied widely in the avian subset of the global carotenoid network (Figs. 2 and 3). All but one compound were associated with at least one reaction to a maximum of 10 reactions. Non-dietary compounds were one to eight reactions away from starting dietary carotenoids (Fig. 3). The species’ networks (Fig. 4a, 4b, 4c, 4d and 4e; Additional file 1b) differed widely in the number of total compounds (1-21), number of reactions (0-46), connectivity (0-4.53 average reactions per compound), diameter length (0-8 reactions), number of modules (1-6), and number of dietary carotenoids (1-6). Structural determinants of compound occurrence among species The connectivity of a compound contributed the most to its species representation; carotenoids with higher connectivity had greater species representation (Fig. 5a; bST = 0.73, t = 7.63, P < 0.001, n = 55). Species representation of a compound did not vary with its distance from a dietary carotenoid (Fig. 5b; bST = -0.07, t = -0.72, P = 0.48, n = 55).Fig. 5 A compound’s connectivity contributed more to the compound’s occurrence than did the compound’s relative distance from a dietary compound. Shown are partial regressions of a compound’s species representation on (a), the number of reactions per compound and (b), its distance from a dietary compound The role of modules in compound occurrence among species The representation of functional modules of the avian carotenoid network varied across species' networks (Fig. 6a and b). Modules of higher connectivity occurred in more species (Fig. 6a; Spearman’s ρ = 0.80, P = 0.006, n = 10), but the diameter of a module was not related to the occurrence of the module across species (Fig. 6b; ρ = 0.49, P = 0.15, n = 10). Differences in the numbers of species with each of the compounds in a module were correlated with the connectivity of the module (Fig. 6c; ρ = 0.74, P = 0.01, n = 10), but not with the diameter of the module (Fig. 6d; ρ = 0.59, P = 0.07, n = 10).Fig. 6 Species representations of interconnected compounds within modules were related to the connectivity, but not the length of pathways of these modules. Compounds in modules characterized by (a), greater overall connectivity were overrepresented across species’ networks, whereas the occurrence of compounds in modules was not related to (b), the diameter of the module. Vertical bars represent the standard error. Differences in the species representation of compounds in the same module increased with (c), greater module enzymatic connectivity, but was not related to (d), the diameter of the module Structural determinants of metabolic distance among species networks In pairs of species networks that shared dietary carotenoids, differences in network connectivity accounted for more of the metabolic distance between species’ networks (Fig. 7a; bST = 0.67, t = 75.24, P < 0.001, n = 4,839) than did differences in the diameters of the networks (Fig. 7b; bST = 0.28, t = 31.50, P < 0.001, n = 4,839). Pairs of networks with large differences in the average number of reactions per compound were more metabolically distinct than networks with large differences in their diameters.Fig. 7 Differences in enzymatic connectivity contributed more to network divergence than differences in diameter. Shown are partial regression plots of the metabolic distance between pairs of species’ networks that share the same dietary (starting) compounds and the difference in (a), network connectivity and (b), diameter length between each pair of networks Structural properties of realized diversification of enzymatic reactions The connectivity of a substrate compound contributed to the realized diversification across species of the reactions associated with the substrate compound (Fig. 8a; bST = 0.38, t = 3.10, P = 0.003, n = 81). The realized diversification of reactions in the network was not predicted by the distance of their substrate compounds from dietary compounds (Fig. 8b; bST = -0.05, t = -0.39, P = 0.70, n = 81).Fig. 8 Realized diversification of the reactions associated with a compound (the fraction of species that do not have a reaction among all of the species that have the substrate compound for the reaction) was predicted by the connectivity of the substrate compound (reactions per compound), but not by the substrate compound’s distance from a dietary compound. Shown are partial regressions of the realized diversification of a reaction on (a), the enzymatic connectivity and (b), the distance from a dietary compound of the reaction’s substrate compound Discussion To what extent is the exploration of a deterministic network and its associated phenotypic diversification the result of the network’s structural properties? The divergence between species’ networks could be driven by either the exploration of pathways from conserved compounds, the elongation of conserved pathways, or the addition of different modules. Our findings suggest that pathway diversification is the main mechanism of divergence among species’ metabolic networks; differences in the enzymatic connectivity among species’ networks contributed more to their metabolic divergence than did differences in the length of their diameters (Fig. 7). In the avian subset of the global carotenoid metabolic network, the connectivity of a compound strongly contributed to further network diversification: compound connectivity contributed the most to both the frequency of compound occurrence across species (Fig. 5a) and the realized diversification of the reactions associated with the compound among species’ networks (Fig. 8a). In contrast, pathway elongation did not play a major role in the diversification of avian carotenoid networks: the relative distance from a dietary compound was not related to a compound’s representation across species (Fig. 5b) or to the realized diversification of reactions associated with the compound among species’ networks (Fig. 8b). The presence of distinct structural modules and differences in the species representation of compounds within these modules contributed to the metabolic divergence across species: the most densely connected modules were the most prevalent across species’ networks. Metabolic divergence across species, however, was not due to the concurrent gain or loss of all of the compounds in a module (Fig. 6c and d). Thus, pathway diversification strongly contributes to metabolic divergence among species: modules characterized by greater connectivity provided more opportunities for the use of distinct pathways. A central assumption of these tests and their interpretation, is that species are co-opting elements (genes or enzymes) that comprise the global avian carotenoid metabolic network and are selectively expressing a particular subset of these elements, rather than evolving them de novo. Several lines of evidence support this assumption. First, there was no correspondence between the historical relationships across study species and their utilization of carotenoid network space (i.e., use or disuse of particular reactions and compounds; [66, 79]). Instead the structure of networks, in particular the link between pathway elongation and pathway diversification, accounted for recurrent convergence of phylogenetically distant and ecologically distinct species in the utilization of network space and expression of carotenoid compounds (ibid.). Although such a pattern could be produced by the independent evolution of enzymes with identical functions, it is highly unlikely (e.g., [80]). In other taxa, horizontal gene transfer [58, 81–84] and symbiotic events [85] accounted for enzymatic convergence in carotenoid metabolism between unrelated species, but neither of these processes play a significant role in avian carotenoid biosynthesis. Gene duplications could similarly account for the evolution of convergent enzymes [24, 83, 86, 87], but the rate of gene duplications in birds [88] seems orders of magnitude lower that would be required to explain the documented rates of carotenoid enzyme convergence across bird species [66]. Instead, species-specific expression of compounds and reactions by the selective expression of different enzyme-encoding genes from the global carotenoid network, appears to be the dominant mode of avian carotenoid network evolution [88, 89], with de novo evolution of new carotenoid pathways (e.g., [90–92]) playing a secondary role (Additional file 1b). A potential mechanism that could drive pathway diversification of enzymatic reactions at these connected compounds is differences in the control of metabolic flux among species across different pathways [93]. Alternatively, different threshold concentrations of a substrate compound associated with several enzymatic reactions may be required to activate different enzymatic reactions [94, 95], such that the diversification of these pathways among species should be dependent on changes in the concentrations of these connected compounds. We showed that the evolutionary representation of compounds and enzymatic reactions reflected their structural properties in the global carotenoid network (Fig. 5a). Why do compounds with the greatest connectivity tend to be overrepresented across species? The longer evolutionary persistence of the most connected elements is a common property of protein and gene deterministic networks across many taxa [e.g., 23, 24, 39, 40] and could reflect their role in maintaining the overall structural cohesiveness and function of the network. The removal or modification of highly connected elements could have greater pleiotropic effects that are more harmful to the function of the network than the removal of less connected compounds [96–98]. This property can result in stronger selection against the loss of these elements (e.g., [99]) or, alternatively, in lesser effectiveness of purifying selection for the deletion of centrally located elements in the network [100, 101]. Further, metabolic flux theory suggests that enzymes with the highest flux control coefficients should be located at the branching points of pathways in metabolic networks [102–105]. Such enzymes experience stronger stabilizing selection than those that contribute less to the flow of metabolites through metabolic pathways (e.g., [106]), accounting for the link between enzymatic connectivity and evolutionary persistence found in this study (Fig. 5a). These conclusions are corroborated by the models of network evolution and empirical studies of network growth that find that new elements in a network preferentially attach to evolutionarily stable elements that have greater connectivity rather than to sparsely connected, but more evolutionary labile downstream elements [24, 28, 34, 38]. It is possible that dietary compounds – the upstream-most elements of avian carotenoid networks – are not evolutionarily stable enough to contribute to incremental pathway elongation over evolutionary time. The evolutionary rates of the gain and loss of dietary carotenoids were orders of magnitude higher than the evolutionary lability of other compounds across avian metabolic networks [66], and our results show that dietary compounds were no more likely to be present in a network than metabolized downstream compounds (Fig. 5b). Theory predicts that rate-limiting enzymes should occur at upstream positions in pathways (e.g., [44]), however the evolutionary instability of dietary compounds can decrease the effectiveness of selection on these compounds. Instead, due to the high enzymatic connectivity of some compounds in carotenoid networks, pathways from different dietary starting points can ultimately produce the same end products (Fig. 2). Thus, network robustness to evolutionary labile dietary compounds – a central feature of avian carotenoid networks [66, 107] – may also contribute to the evolutionary stability of the connected compounds and explain why the diversification of species’ networks was centered on connected compounds instead of the continued lengthening of pathways from specific dietary compounds. Variance in the species representations of compounds and enzymatic reactions within the same modules (Fig. 6c and d) implies that the modules partitioned by their structural properties do not correspond to actual biological processes (e.g. [108]), despite the fact that the structural modules used in this study were associated with different dietary compounds. Differences in the number of species with each compound in a module, however, could be the result of the connectivity of each of the compounds to other modules, which has been shown to explain the evolutionary rate of genes in protein interaction networks [109]. Furthermore, it is possible that species utilize all of the enzymatic reactions and produce all of the compounds in a module but selectively express only some of the compounds in their plumage [107, 110–112], and so the variation of the species representations of compounds in modules captures this selective compound deposition of the products of a module. By identifying the topological structural properties in a deterministic network that underlie phenotypic differences we can begin to establish specific mechanisms for the microevolutionary sequences behind observed macroevolutionary patterns. For example, if highly connected network elements determine phenotypic differences, then phenotypic diversification in a lineage might not occur in sequential order (structural or temporal) because different pathways can be explored from the same initial conserved element, and so we would expect weak phylogenetic signal among phenotypes. If pathway elongation is the source of phenotypic differences, then the dependence between downstream and upstream elements imposes a clear sequential order to phenotypic diversification along the pathway, resulting in stronger historical associations across species’ networks. The incorporation or loss of entire modules of elements in a deterministic network may be ordered or unordered, depending on their relative positions, but either would result in recurrent bursts of diversification across lineages’ phenotypes [113–115]. Because we found no evidence of avian carotenoid network diversification due to pathway elongation, we would not expect a sequential order in patterns of realized diversification in carotenoid pathways during avian evolutionary history. Instead, our finding that differences among species’ networks were due to pathway diversification from highly connected compounds, suggests that related species should have similar carotenoid networks only when they utilize the same pathways from the same shared compound. The results of this study thus explain why phenotypic diversification in expressed carotenoids between related species was overwhelmingly due to unordered periodic bursts of biochemical diversification of several compounds at once in the same pathway module across species, with ecological divergence in the use of dietary carotenoids – the process closely associated with ecological speciation, pathway elongation, and species relatedness – playing a significantly weaker role [66, 107]. Conclusions The goal of this study was to explicitly consider how the structural interactions among elements of a trait affect its diversification. Our results show that the structure of the enzymatic reactions in the avian space of the global carotenoid network delineates opportunities for diversification of expressed carotenoids in birds. Within-species studies can establish the proximate mechanisms underlying the observed association of network topology, enzymatic connectivity and evolutionary diversification in carotenoid compounds. Explicit consideration of spatial and temporal organization of interactions between genes, proteins, enzymes and other elements of deterministic networks brings us closer to an understanding of the relationship between potential and realized phenotypic diversity. Abbreviation MYA, million years ago Additional files Additional file 1: (a) Appendix S1: Confirmed enzymatic reactions in the “avian space” of global carotenoid biosynthesis network in bacteria, plants, and animals. This appendix contains references supporting the presence of specific compounds and the enzymatic reactions that comprise the avian carotenoid biosynthesis global network. (b) Appendix S2: Characteristics of carotenoid metabolic networks for species used in the study. This appendix contains the structural measurements and references for compound identification and the method of identification for each of the species’ metabolic networks. (c) Appendix S3: Module assignments in the avian subset of the global carotenoid metabolic network. This appendix contains the module assignments for each of the compounds in the global avian carotenoid metabolic network. The number of the module corresponds to the partitioned regions in Fig. 2. (PDF 1501 kb) Additional file 2: Species’ binary metabolic networks. This appendix contains binary metabolic networks for each of the species included in the study. (XLSX 123 kb) Additional file 3: Majority rule consensus phylogeny of species included in the study. This appendix contains the Newick tree format of the majority rule consensus phylogeny visually presented in Fig. 4, 5, 6, 7 and 8. The tree is based on 1,000 randomly sampled trees from the Hackett All Species pseudo-posterior distribution downloaded from birdtree.org that is based on Jetz et al. 2012. (TXT 20 kb) Acknowledgements We thank V. Belloni, V. Farrar and J. Andrews for help with the data collection, and R. Duckworth, M. Sanderson, D. Higginson, A. Potticary, C. Gurguis, G. Semenov and three anonymous reviewers for thorough comments on previous versions and helpful suggestions. Funding This work was supported by the David and Lucille Packard Foundation, Amherst College graduate fellowships, and the University of Arizona Open Access Publishing Fund. Availability of data and material The datasets supporting the results of this article are available as additional files (Additional files 1, 2 and 3). Authors’ contributions ESM designed the study. ESM and AVB analyzed the data. ESM wrote the manuscript with help from AVB. Both authors have read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate Not applicable. ==== Refs References 1. Gavrilets S Fitness landscapes and the origin of species 2004 Princeton Princeton University Press 2. Gerhart J Kirschner M The theory of facilitated variation Proc Natl Acad Sci U S A 2007 104 8582 9 10.1073/pnas.0701035104 17494755 3. Maynard SJ Natural selection and the concept of a protein space Nature 1970 225 563 4 10.1038/225563a0 5411867 4. Wagner GP Homology, genes, and evolutionary innovation 2014 Princeton Princeton University Press 5. Newman SA The developmental genetic toolkit and the molecular homology—analogy paradox Biol Theory 2006 1 12 6 10.1162/biot.2006.1.1.12 6. Badyaev AV Walsh JB Charmantier A Garant D Kruuk LEB Epigenetic processes and genetic architecture in character origination and evolution Quantitative genetics in the wild 2014 Oxford Oxford University Press 177 89 7. Bershtein S Segal M Bekerman R Tokuriki N Tawfik DS Robustness-epistasis link shapes the fitness landscape of a randomly drifting protein Nature 2006 444 929 32 10.1038/nature05385 17122770 8. Breen MS Kemena C Vlasov PK Notredame C Kondrashov FA Epistasis as the primary factor in molecular evolution Nature 2012 490 535 8 10.1038/nature11510 23064225 9. Gravner J Pitman D Gavrilets S Percolation on fitness landscapes: effects of correlation, phenotype, and incompatibilities J Theor Biol 2007 248 627 45 10.1016/j.jtbi.2007.07.009 17692873 10. Poelwijk FJ Kiviet DJ Weinreich DM Tans SJ Empirical fitness landscapes reveal accessible evolutionary paths Nature 2007 445 383 6 10.1038/nature05451 17251971 11. Rice SH Wolf JB Brodie ED III Wade MJ The evolution of developmental interactions: Epistasis, canalization, and integration Epistasis and the evolutionary process 2001 New York Oxford University Press 82 98 12. Alberch P From genes to phenotype: Dynamical systems and evolvability Genetica 1991 84 5 11 10.1007/BF00123979 1874440 13. Arthur W Developmental drive: An important determinant of the direction of phenotypic evolution Evol Dev 2001 3 271 8 10.1046/j.1525-142x.2001.003004271.x 11478524 14. Forgacs G Newman SA Biological physics of the developing embryo 2005 Cambridge Cambridge University Press 15. Whyte LL Internal factors in evolution 1965 New York George Braziller 16. Bloom JD Labthavikul ST Otey CR Arnold FH Protein stability promotes evolvability Proc Natl Acad Sci U S A 2006 103 5869 74 10.1073/pnas.0510098103 16581913 17. Bridgham JT Ortlund EA Thornton JW An epistatic ratchet constrains the direction of glucocorticoid receptor evolution Nature 2009 461 515 9 10.1038/nature08249 19779450 18. Harms MJ Thornton JW Historical contingency and its biophysical basis in glucocorticoid receptor evolution Nature 2014 512 203 7 10.1038/nature13410 24930765 19. Newman SA Physico-genetic determinants in the evolution of development Science 2012 338 217 9 10.1126/science.1222003 23066074 20. Pagel M Pomiankowski A Evolutionary genomics and proteinomics 2008 Sunderland Sinauer Associates 21. Povolotskaya IS Kondrashov FA Sequence space and the ongoing expansion of the protein universe Nature 2010 465 922 7 10.1038/nature09105 20485343 22. Wagner A The molecular origins of evolutionary innovations Trends Genet 2011 27 397 410 10.1016/j.tig.2011.06.002 21872964 23. Fraser HB Hirsh AE Steinmetz LM Scharfe C Feldman MW Evolutionary rate in the protein interaction network Science 2002 296 750 2 10.1126/science.1068696 11976460 24. Light S Kraulis P Elofsson A Preferential attachment in the evolution of metabolic networks BMC Genomics 2005 6 159 10.1186/1471-2164-6-159 16281983 25. Liu WC Lin WH Davis AJ Jordán F Yang HT Hwang MJ A network perspective on the topological importance of enzymes and their phylogenetic conservation BMC Bioinformatics 2007 8 121 10.1186/1471-2105-8-121 17425808 26. Yamada T Bork P Evolution of biomolecular networks-lessons from metabolic and protein interactions Nat Rev Mol Cell Biol 2009 10 791 803 10.1038/nrm2787 19851337 27. Zhao J Ding G-H Tao L Yu H Yu Z-H Luo J-H Modular co-evolution of metabolic networks BMC Bioinformatics 2007 8 311 10.1186/1471-2105-8-311 17723146 28. Maslov S Krishna S Pang TY Sneppen K Toolbox model of evolution of prokaryotic metabolic networks and their regulation Proc Natl Acad Sci U S A 2009 106 9743 8 10.1073/pnas.0903206106 19482938 29. Banerjee A Structural distance and evolutionary relationship of networks BioSyst 2011 107 186 96 10.1016/j.biosystems.2011.11.004 30. Borenstein E Kupiec M Feldman MW Ruppin E Large-scale reconstruction and phylogenetic analysis of metabolic environments Proc Natl Acad Sci U S A 2008 105 14482 7 10.1073/pnas.0806162105 18787117 31. Ebenhöh O Handorf T Kahn D Evolutionary changes of metabolic networks and their biosynthetic capacities IEE P Syst Biol 2006 153 354 8 10.1049/ip-syb:20060014 32. Mithani A Hein J Preston GM Comparative analysis of metabolic networks provides insight into the evolution of plant pathogenic and non pathogenic lifestyles in Pseudomonas Mol Biol Evol 2011 28 483 99 10.1093/molbev/msq213 20709733 33. Navlakha S Kingsford C Network archaeology: uncovering ancient networks from present-day interactions PLoS Comp Biol 2011 7 10.1371/journal.pcbi.1001119 34. Barabási A-L Albert R Emergence of scaling in random networks Science 1999 286 509 12 10.1126/science.286.5439.509 10521342 35. Jeong H Tombor B Albert R Oltvai ZN Barabási A-L The large-scale organization of metabolic networks Nature 2000 407 651 4 10.1038/35036627 11034217 36. Thieffry D Huerta AM Pérez-Rueda E Collado-Vides J From specific gene regulation to genomic networks: a global analysis of transcriptional regulation in Escherichia coli Bioessays 1998 20 433 40 10.1002/(SICI)1521-1878(199805)20:5<433::AID-BIES10>3.0.CO;2-2 9670816 37. Barabási A-L Luck or reason Nature 2012 489 507 8 10.1038/nature11486 22972190 38. Eisenberg E Levanon EY Preferential attachment in the protein network evolution Phys Rev Lett 2003 91 138701 10.1103/PhysRevLett.91.138701 14525344 39. Bernhardsson S Gerlee P Lizana L Structural correlations in bacterial metabolic networks BMC Evol Biol 2011 11 20 10.1186/1471-2148-11-20 21251250 40. Hahn MW Kern AD Comparative genomics of centrality and essentiality in three Eukaryotic protein-interaction networks Mol Biol Evol 2005 22 803 6 10.1093/molbev/msi072 15616139 41. Xu K Bezakova I Bunimovich L Yi SV Path lengths in protein–protein interaction networks and biological complexity Proteomics 2011 11 1857 67 10.1002/pmic.201000684 21480527 42. Ramsay H Rieseberg LH Ritland K The correlation of evolutionary rate with pathway position in plant terpenoid biosynthesis Mol Biol Evol 2009 26 1045 53 10.1093/molbev/msp021 19188263 43. Rausher MD Miller RE Tiffin P Patterns of evolutionary rate variation among genes of the anthocyanin biosynthetic pathway Mol Biol Evol 1999 16 266 74 10.1093/oxfordjournals.molbev.a026108 10028292 44. Wright KM Rausher MD The evolution of control and distribution of adaptive mutations in a metabolic pathway Genetics 2010 184 483 502 10.1534/genetics.109.110411 19966064 45. Hartwell LH Hopfield JJ Leibler S Murray AW From molecular to modular cell biology Nature 1999 402 C47 52 10.1038/35011540 10591225 46. Ravasz E Somera AL Mongru DA Oltvai ZN Barabasi A-L Hierarchical organization of modularity in metabolic networks Science 2002 297 1551 5 10.1126/science.1073374 12202830 47. Badyaev A Evolvability and robustness in color displays: Bridging the gap between theory and data Evol Biol 2007 34 61 71 10.1007/s11692-007-9004-5 48. Nagy L Changing patterns of gene regulation in the evolution of arthropod morphology Am Zool 1998 38 818 28 10.1093/icb/38.6.818 49. Raff EC Raff RA Dissociability, modularity, evolvability Evol Dev 2000 2 235 7 10.1046/j.1525-142x.2000.00069.x 11252551 50. von Dassow G Munro E Modularity in animal development and evolution: elements of a conceptual framework for EvoDevo J Exp Zool 1999 285 307 25 10.1002/(SICI)1097-010X(19991215)285:4<307::AID-JEZ2>3.3.CO;2-M 10578108 51. Wagner GP Altenberg L Perspective: Complex adaptations and the evolution of evolvability Evolution 1996 50 967 76 10.2307/2410639 52. Eisen MB Spellman PT Brown PO Botstein D Cluster analysis and display of genome-wide expression patterns Proc Natl Acad Sci U S A 1998 95 14863 8 10.1073/pnas.95.25.14863 9843981 53. Halfon MS Grad Y Church GM Michelson AM Computation-based discovery of related transcriptional regulatory modules and motifs using an experimentally validated combinatorial model Genome Res 2002 12 1019 28 12097338 54. Ihmels J Friedlander G Bergmann S Sarig O Ziv Y Barkai N Revealing modular organization in the yeast transcriptional network Nat Genet 2002 31 370 77 12134151 55. Niehrs C Pollet N Synexpression groups in eukaryotes Nature 1999 402 483 7 10.1038/990025 10591207 56. Campillos M von Mering C Jensen LJ Bork P Identification and analysis of evolutionarily cohesive functional modules in protein networks Genome Res 2006 16 374 82 10.1101/gr.4336406 16449501 57. Chen Y Dokholyan NV The coordinated evolution of yeast proteins is constrained by functional modularity Trends Genet 2006 22 416 9 10.1016/j.tig.2006.06.008 16797778 58. Pál C Papp B Lercher MJ Adaptive evolution of bacterial metabolic networks by horizontal gene transfer Nat Genet 2005 37 1372 5 10.1038/ng1686 16311593 59. Wagner A Evolutionary constraints permeate large metabolic networks BMC Evol Biol 2009 9 231 10.1186/1471-2148-9-231 19747381 60. Klassen JL Phylogenetic and evolutionary patterns in microbial carotenoid biosynthesis are revealed by comparative genomics PLoS One 2010 5 10.1371/journal.pone.0011257 20582313 61. Umeno D Tobias AV Arnold FH Diversifying carotenoid biosynthetic pathways by directed evolution Microbiol Mol Biol Rev 2005 69 51 78 10.1128/MMBR.69.1.51-78.2005 15755953 62. Britton G Liaaen-Jensen S Pfander H Carotenoids 2004 Boston Birkhäuser Verlag 63. Schmidt K Connor A Britton G Goodfellow M O'Donnell AG Analysis of pigments: carotenoids and related polyenes Chemical methods in prokaryotic systematics 1994 Chichester John Wiley & Sons 403 61 64. Brush AH Metabolism of carotenoid-pigments in birds FASEB J 1990 4 2969 77 2394316 65. McGraw KJ Hill GE McGraw KJ The mechanics of carotenoid coloration in birds Bird coloration volume 1: Mechanisms and measurements 2006 Cambridge Harvard University Press 177 242 66. Badyaev A Morrison E Belloni V Sanderson M Tradeoff between robustness and elaboration in carotenoid networks produces cycles of avian color diversification Biol Direct 2015 10 45 10.1186/s13062-015-0073-6 26289047 67. Jaccard P The distribution of the flora in the alpine zone New Phytol 1912 11 37 50 10.1111/j.1469-8137.1912.tb05611.x 68. Rodrigues JFM Wagner A Genotype networks, innovation, and robustness in sulfur metabolism BMC Syst Biol 2011 5 39 10.1186/1752-0509-5-39 21385333 69. Maddison WP, Maddison DR. Mesquite: a modular system for evolutionary analysis. Version 3.03. 2015. [http://mesquiteproject.wikispaces.com/] 70. Midford PE, Garland T, Jr., Maddison WP. PDAP package of Mesquite. Version 1.16. 2011. [http://mesquiteproject.org/pdap_mesquite/]. 71. Guimerà R Amaral LAN Functional cartography of complex metabolic networks Nature 2005 433 895 900 10.1038/nature03288 15729348 72. Guimerà R, Amaral LAN. Cartography of complex networks: Modules and universal roles. J Stat Mech Theor Exp. 2005;2005:P02001-1-13. 73. Harary F Graph theory 1969 Reading Addison-Wesley 74. Smoot ME Ono K Ruscheinski J Wang P-L Ideker T Cytoscape 2.8: New features for data integration and network visualization Bioinformatics 2011 27 431 2 10.1093/bioinformatics/btq675 21149340 75. Assenov Y Ramírez F Schelhorn S-E Lengauer T Albrecht M Computing topological parameters of biological networks Bioinformatics 2008 24 282 4 10.1093/bioinformatics/btm554 18006545 76. Doncheva NT Assenov Y Domingues FS Albrecht M Topological analysis and interactive visualization of biological networks and protein structures Nat Protoc 2012 7 670 85 10.1038/nprot.2012.004 22422314 77. McSweeney PJ. Randomnetworks. Version 1.0. 2008. [http://apps.cytoscape.org/apps/randomnetworks] 78. Ebenhöh O Handorf T Heinrich R A cross species comparison of metabolic network functions Genome Inform 2005 16 203 13 16362923 79. Thomas DB McGraw KJ Butler MW Carrano MT Madden O James HF Ancient origins and multiple appearances of carotenoid-pigmented feathers in birds Proc R Soc B 2014 281 20140806 10.1098/rspb.2014.0806 24966316 80. Furnham N Sillitoe I Holliday GL Cuff AL Laskowski RA Orengo CA Exploring the evolution of novel enzyme functions within structurally defined protein superfamilies PLoS Comp Biol 2012 8 10.1371/journal.pcbi.1002403 81. Altincicek B Kovacs JL Gerardo NM Horizontally transferred fungal carotenoid genes in the two-spotted spider mite Tetranychus urticae Biol Lett 2012 8 253 7 10.1098/rsbl.2011.0704 21920958 82. Kreimer A Borenstein E Gophna U Ruppin E The evolution of modularity in bacterial metabolic networks Proc Natl Acad Sci U S A 2008 105 6976 81 10.1073/pnas.0712149105 18460604 83. Moran NA Jarvik T Lateral transfer of genes from fungi underlies carotenoid production in aphids Science 2010 328 624 7 10.1126/science.1187113 20431015 84. Nováková E Moran NA Diversification of genes for carotenoid biosynthesis in aphids following an ancient transfer from a fungus Mol Biol Evol 2012 29 313 23 10.1093/molbev/msr206 21878683 85. Sloan DB Moran NA Endosymbiotic bacteria as a source of carotenoids in whiteflies Biol Lett 2012 8 986 9 10.1098/rsbl.2012.0664 22977066 86. Kelley BP Sharan R Karp RM Sittler T Root DE Stockwell BR Conserved pathways within bacteria and yeast as revealed by global protein network alignment Proc Natl Acad Sci U S A 2003 100 11394 399 10.1073/pnas.1534710100 14504397 87. Kondrashov FA Gene duplication as a mechanism of genomic adaptation to a changing environment Proc R Soc B 2012 279 5048 57 10.1098/rspb.2012.1108 22977152 88. Zhang G Li C Li Q Li B Larkin DM Lee C Comparative genomics reveals insights into avian genome evolution and adaptation Science 2014 346 1311 20 10.1126/science.1251385 25504712 89. Walsh N Dale J McGraw KJ Pointer MA Mundy NI Candidate genes for carotenoid coloration in vertebrates and their expression profiles in the carotenoid-containing plumage and bill of a wild bird Proc R Soc B 2012 279 58 66 10.1098/rspb.2011.0765 21593031 90. Hudon J Anciães M Bertacche V Stradi R Plumage carotenoids of the pin-tailed manakin (Ilicura militaris ): Evidence for the endogenous production of rhodoxanthin from a colour variant Comp Biochem Physiol B Biochem Mol Biol 2007 147 402 11 10.1016/j.cbpb.2007.02.004 17400013 91. Prum RO LaFountain AM Berro J Stoddard MC Frank HA Molecular diversity, metabolic transformation, and evolution of carotenoid feather pigments in cotingas (Aves: Cotingidae) J Comp Physiol B 2012 182 1095 1116 10.1007/s00360-012-0677-4 22669477 92. Prum R LaFountain A Berg C Tauber M Frank H Mechanism of carotenoid coloration in the brightly colored plumages of broadbills (Eurylaimidae) J Comp Physiol B 2014 184 651 72 10.1007/s00360-014-0816-1 24647990 93. Morrison ES, Badyaev AV. The landscape of evolution: Reconciling structural and dynamic properties of metabolic networks in adaptive diversifications. Integr Comp Biol. 2016;56:235-46. 94. Bongaerts GP Vliegenthart JS Effect of aminoglycoside concentration on reaction rates of aminoglycoside-modifying enzymes Antimicrob Agents Chemother 1988 32 740 6 10.1128/AAC.32.5.740 2840015 95. Matsuno R Nakanishi K Ohnishi M Hiromi K Kamikubo T Threshold in a single enzyme reaction system: reaction of maltose catalyzed by saccharifying α-Amylase from B. Subtilis J Biochem 1978 83 859 62 417077 96. Albert R Jeong H Barabási A-L Error and attack tolerance of complex networks Nature 2000 406 378 82 10.1038/35019019 10935628 97. Jeong H Mason SP Barabási A-L Oltvai ZN Lethality and centrality in protein networks Nature 2001 411 41 42 10.1038/35075138 11333967 98. Schmidt S Sunyaev S Bork P Dandekar T Metabolites: a helping hand for pathway evolution? Trends Biochem Sci 2003 28 336 41 10.1016/S0968-0004(03)00114-2 12826406 99. Aris-Brosou S Determinants of adaptive evolution at the molecular level: The extended complexity hypothesis Mol Biol Evol 2005 22 200 9 10.1093/molbev/msi006 15483330 100. Badyaev AV “Homeostatic hitchhiking”: A mechanism for the evolutionary retention of complex adaptations Integr Comp Biol 2013 53 913 22 10.1093/icb/ict084 23868466 101. Kauffman S Levin S Towards a general theory of adaptive walks on rugged landscapes J Theor Biol 1987 128 11 45 10.1016/S0022-5193(87)80029-2 3431131 102. Heijnen JJ van Gulik WM Shimizu H Stephanopoulos G Metabolic flux control analysis of branch points: an improved approach to obtain flux control coefficients from large perturbation data Metab Eng 2004 6 391 400 10.1016/j.ymben.2004.07.002 15491867 103. LaPorte DC Walsh K Koshland DE The branch point effect. Ultrasensitivity and subsensitivity to metabolic control J Biol Chem 1984 259 14068 75 6389540 104. Pritchard L Kell DB Schemes of flux control in a model of Saccharomyces cerevisiae glycolysis Eur J Biochem 2002 269 3894 3904 10.1046/j.1432-1033.2002.03055.x 12180966 105. Rausher MD The evolution of genes in branched metaoblic pathways Evolution 2013 67 34 48 10.1111/j.1558-5646.2012.01771.x 23289560 106. Flowers J Sezgin E Kumagai S Duvernell D Matzkin L Schmidt P Adaptive evolution of metabolic pathways in Drosophila Mol Biol Evol 2007 24 1347 54 10.1093/molbev/msm057 17379620 107. Higginson DM Belloni V Davis SN Morrison ES Andrews JE Badyaev AV Evolution of long-term coloration trends with biochemically unstable ingredients Proc R Soc B 2016 283 20160403 10.1098/rspb.2016.0403 27194697 108. Wang Z Zhang J In search of the biological significance of modular structures in protein networks PLoS Comp Biol 2007 3 10.1371/journal.pcbi.0030107 109. Fraser HB Modularity and evolutionary constraint on proteins Nat Genet 2005 37 351 2 10.1038/ng1530 15750592 110. Fox DL Metabolic fractionation, storage and display of carotenoid pigments by flamingoes Comp Biochem Physiol 1962 6 1 24 10.1016/0010-406X(62)90040-3 13959019 111. Fox DL Smith VE Wolfson AA Carotenoid selectivity in blood and feathers of lesser (African), Chilean and greater (European) flamingos Comp Biochem Physiol 1967 23 225 32 10.1016/0010-406X(67)90490-2 6075163 112. McGraw KJ Beebee MD Hill GE Parker RS Lutein-based plumage coloration in songbirds is a consequence of selective pigment incorporation into feathers Comp Biochem Physiol B Biochem Mol Biol 2003 135 689 96 10.1016/S1096-4959(03)00164-7 12892761 113. Gerhart J Kirschner M Evolution and evolvability Cells, embryos, and evolution 1997 Malden Blackwell Science 580 614 114. Reid RGB Biological emergence: Evolution by natural experiment 2007 Cambridge MIT Press 115. Yang AS Modularity, evolvability, and adaptive radiations: a comparison of the hemi- and holometabolous insects Evol Dev 2001 3 59 72 10.1046/j.1525-142x.2001.003002059.x 11341675 116. Jetz W Thomas GH Joy JB Hartmann K Mooers AO The global diversity of birds in space and time Nature 2012 491 444 8 10.1038/nature11631 23123857

PMC005xxxxxx/PMC5000422.txt

==== Front BMC CancerBMC CancerBMC Cancer1471-2407BioMed Central London 272510.1186/s12885-016-2725-zResearch ArticleHigh sensitivity isoelectric focusing to establish a signaling biomarker for the diagnosis of human colorectal cancer Padhan Narendra narendra.padhan@igp.uu.se 1Nordling Torbjörn E. M. tn@kth.se 124Sundström Magnus magnus.sundstrom@igp.uu.se 1Åkerud Peter peterakerud@gmail.com 3Birgisson Helgi helgi.birgisson@surgsci.uu.se 3Nygren Peter peter.nygren@igp.uu.se 1Nelander Sven sven.nelander@igp.uu.se 1http://orcid.org/0000-0003-4275-2000Claesson-Welsh Lena lena.welsh@igp.uu.se 11 Department of Immunology, Genetics and Pathology, Rudbeck Laboratory, Uppsala University, Dag Hammarskjöldsv 20, Uppsala, 751 85 Sweden 2 Stockholm Bioinformatics Centre, Science for Life Laboratory, Box 1031, 171 21 Solna, Sweden 3 Department Surgical Sciences, Uppsala University, 751 85 Uppsala, Sweden 4 Current address: Department of Mechanical Engineering, National Cheng Kung University, No. 1 University Road, Tainan, 70101 Taiwan 25 8 2016 25 8 2016 2016 16 1 68322 9 2015 15 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background The progression of colorectal cancer (CRC) involves recurrent amplifications/mutations in the epidermal growth factor receptor (EGFR) and downstream signal transducers of the Ras pathway, KRAS and BRAF. Whether genetic events predicted to result in increased and constitutive signaling indeed lead to enhanced biological activity is often unclear and, due to technical challenges, unexplored. Here, we investigated proliferative signaling in CRC using a highly sensitive method for protein detection. The aim of the study was to determine whether multiple changes in proliferative signaling in CRC could be combined and exploited as a “complex biomarker” for diagnostic purposes. Methods We used robotized capillary isoelectric focusing as well as conventional immunoblotting for the comprehensive analysis of epidermal growth factor receptor signaling pathways converging on extracellular regulated kinase 1/2 (ERK1/2), AKT, phospholipase Cγ1 (PLCγ1) and c-SRC in normal mucosa compared with CRC stage II and IV. Computational analyses were used to test different activity patterns for the analyzed signal transducers. Results Signaling pathways implicated in cell proliferation were differently dysregulated in CRC and, unexpectedly, several were downregulated in disease. Thus, levels of activated ERK1 (pERK1), but not pERK2, decreased in stage II and IV while total ERK1/2 expression remained unaffected. In addition, c-SRC expression was lower in CRC compared with normal tissues and phosphorylation on the activating residue Y418 was not detected. In contrast, PLCγ1 and AKT expression levels were elevated in disease. Immunoblotting of the different signal transducers, run in parallel to capillary isoelectric focusing, showed higher variability and lower sensitivity and resolution. Computational analyses showed that, while individual signaling changes lacked predictive power, using the combination of changes in three signaling components to create a “complex biomarker” allowed with very high accuracy, the correct diagnosis of tissues as either normal or cancerous. Conclusions We present techniques that allow rapid and sensitive determination of cancer signaling that can be used to differentiate colorectal cancer from normal tissue. Electronic supplementary material The online version of this article (doi:10.1186/s12885-016-2725-z) contains supplementary material, which is available to authorized users. Keywords Colorectal cancerIsoelectric focusingSignal transductionProliferationERKc-SRChttp://dx.doi.org/10.13039/501100002794CancerfondenCAN2013/661Claesson-Welsh Lena http://dx.doi.org/10.13039/501100004063Knut och Alice Wallenbergs StiftelseKAW 2015.0275Claesson-Welsh Lena http://dx.doi.org/10.13039/501100007051Uppsala UniversitetGAj 2012Padhan Narendra http://dx.doi.org/10.13039/501100007287Worldwide Cancer Research13-1295Claesson-Welsh Lena Swedish Strategic research program eSSENCEhttp://dx.doi.org/10.13039/501100004663Ministry of Science and Technology, Taiwan105-2218-E-006-016-MY2Nordling Torbjörn E. M. issue-copyright-statement© The Author(s) 2016 ==== Body Background Although the prognosis of patients with colorectal cancer (CRC) is steadily improving, the disease remains the second most common cause of cancer-related deaths in Europe [1]. The treatment of CRC is dependent on the disease stage and the location of the tumor. Conventional treatment includes surgery, radiation and chemotherapy (5-fluorouracil, irinotecan and/or oxaliplatin) [2], often combined with bevacizumab (a neutralizing antibody against vascular endothelial growth factor; VEGF) or cetuximab/panitumumab (neutralizing antibodies against epidermal growth factor receptor; EGFR), depending on disease stage and patient-related factors [3]. During the course of CRC, mutations accumulate in genes controlling cell survival and proliferation. Several of the genes afflicted in CRC belong to the RAS pathway [4]. The RAS pathway involves at least 4 key protein families (RAS, RAF, mitogen-activated protein kinase kinase (MEK) and extracellular regulated kinase (ERK)) that are activated in a consecutive manner, creating a signaling cascade that eventually results in gene regulation. Approximately 50 % of metastatic CRCs have activating mutations in the KRAS or NRAS genes [5–7]. Patients with RAS mutations do not respond favorably to treatment with neutralizing anti-EGFR antibodies [8]. BRAF is the best characterized of three closely related RAF proteins [9]. The BRAF gene harbors an activating mutation (V600E) in 5–12 % of all CRC [10]. Tumors may have mutations either in KRAS or BRAF though, as a rule, not in both [11]. Activation of certain protein kinase C (PKC) isoforms, such as PKCɛ, by phospholipase Cγ1 (PLCγ1), promotes RAF activation [12]. BRAF in turn activates the dual tyrosine and serine/threonine kinase MEK, which is mutated only very rarely in CRC [13]. The serine/threonine kinases ERK1/2, downstream of MEK, are also not mutated in CRC [13]. Cell proliferation is regulated also by the cytoplasmic tyrosine kinase c-SRC, which is activated when phosphorylated on tyrosine residue (Y) 418 in the kinase domain and which is inhibited when phosphorylated on the C-terminal Y527 [14]. c-SRC expression is reported to be 5–8 fold higher in premalignant colorectal polyps than in normal mucosa and a correlation between elevated c-SRC levels and CRC progression/metastatic potential has been suggested [15–17]. c-SRC kinase inhibitors are being developed for therapeutic purposes [18, 19]. Resistance to BRAF inhibition in melanoma can be overcome by inhibiting c-SRC activity [20], indicating a convergence of the pathways. Cell survival is regulated by the phosphoinositide 3-kinase (PI3K)/AKT pathway which, via mammalian target of rapamycin complex 1 (mTORC1), eventually results in activation of p70S6 kinase and gene induction [21]. The serine/threonine kinase AKT is activated by phosphorylation of threonine (T) 308 located in the kinase domain and serine (S) 473 in the C-terminal end, by phosphatidylinositol-dependent kinase 1 (PDK1) and mTORC2, respectively. The PI3K/AKT pathway is negatively regulated by the lipid phosphatase, phosphatase and tensin homolog (PTEN) [22], which has been identified as a tumor suppressor [23]. About 15 % of all CRCs have activating or suppressing mutations in the PI3KCA gene, encoding the p110α catalytic subunit of PI3K, as well as the PTEN gene [24]. Moreover, in wild type (non-mutated) KRAS gene tumors, the presence of PI3K and PTEN mutations indicates a poor prognosis [25]. To identify mutations in cancer is part of an effort to individualize each patient’s treatment. However, mutations may not result in changes in protein expression levels and/or activity, and the mutation status of a particular cancer may fail to convey information about additional events occurring during progression of the disease, which may override a particular mutation, e.g. compensatory upregulation of other proteins and pathways [26]. There is no doubt that the EGFR/RAS pathway and downstream ERK1/ERK2 activities are essential in CRC etiology and disease progression [27]. However, predicting RAS pathway activity is particularly complex as there are several different upstream and parallel activators on different levels and many alternative feedback loops [26]. Apart from the regulation of RAS activity through GTPase regulatory proteins (GAPs and GEFs), downstream signaling in the RAS pathway can be induced or modulated through activities in several other pathways, including the PLCγ/PKC, PI3K/AKT and c-SRC pathways. Another complicating aspect of RAS signaling in CRC is chromosomal fragility. 85 % of sporadic CRC cases display chromosomal instability, chromosome amplification and translocation leading to aneuploidy (see [28] and refs therein), whereas the remaining 15 % of patients have high-frequency microsatellite instability phenotypes i.e. frameshift mutations and base pair substitutions [29]. The chromosomal instability of CRC clearly influences the biological consequence of the mutations. Thus, taken together, the presence of a mutation in a signaling protein does not necessarily predict activity in the corresponding signaling pathway. Due to the existing challenges in CRC therapy, the development of rapid and sensitive screens to measure the biological activity of key signal transducers, which could serve as drug targets or as predictive or prognostic biomarkers, is warranted. Previously, the CRC proteome has been investigated using mass spectrometry to identify up- and downregulation of proteins, using mostly cell lines but also, to some extent, patient samples [30]. However, this is the first study to comprehensively address the proliferative signaling proteome in CRC tissues. For this purpose, we have developed protocols for highly sensitive, robotized isoelectric focusing, to show that signaling in the RAS pathway is dysregulated in human CRC primary tumors compared with normal mucosa. Moreover, by computational and geometric assessment of the signal transduction patterns in the different tissues examined (normal, stage II and stage IV CRC), we show that combinations of patterns from several pathways could serve as biomarkers and be exploited for the classification of tissues as normal or cancerous. We suggest that further refinement of complex signatures can be exploited for prognostic purposes. Methods Tumor biopsy collection The colorectal tumor sampling and characterization of the anonymous samples was approved by the Uppsala Regional Ethical Review Board (no 2007/005 and 2000/001). Prior to the operation the patient was asked by the responsible surgeon to donate tumor tissue and blood samples for future molecular studies. Patients agreeing to participate were given written study information and signed an informed consent form. When the surgical specimen (colon) was removed from the patient, it was immediately transported on ice to the histopathological department and a clinical pathologist cut a 5x5x5 mm biopsy from the periphery of the primary tumor and a 10x10 mm normal mucosa more than 5 cm from the primary tumor. The biopsies were immediately placed, without addition of medium, in test tubes, which were stored at -80 °C until analyses were made. Thirty-three colon cancer samples were selected from a set of frozen tumor biopsies collected from patients operated upon for colorectal cancer at the hospitals in Karlstad or Västerås, Sweden. Seventeen of the 33 patients had stage II colon cancer and 16 had stage IV colon cancer. Samples of normal mucosa from 18 patients were available for analyses. Cell culture and VEGF treatment Human umbilical vein endothelial cells (HUVECs; ATCC; Manassas, VA) were cultured on gelatin-coated 10 cm tissue culture petri dishes in endothelial cell basal medium MV2 (EBM-2, C-22221; PromoCell, Heidelberg, Germany) with supplemental pack C-39221, containing 5 % FCS, epidermal growth factor (5 ng/ml), VEGF (0.5 ng/ml), basic FGF (10 ng/ml), Insulin-like Growth Factor (Long R3 IGF, 20 ng/ml), hydrocortisone (0.2 μg/ml), and ascorbic acid (1 μg/ml). HUVECs at passages 3–6 were used. For experimental purposes, ECs were serum-starved overnight and plated in EBM-2 medium, 1 % FCS without growth factor supplement and treated with/without VEGF (50 ng/ml, Preprotech, Rocky Hill, NJ) for 7.5 min or 15 min. The cells were lysed in a commercial RIPA buffer containing protease inhibitor mix (# 040-482, ProteinSimple, Santa Clara, CA) and phosphatase inhibitors (# 040-510, ProteinSimple). The lysates were clarified by centrifugation and protein concentrations were determined by using BCA Protein Assay Kit (Pierce ThermoFisher Scientific, Rockford, IL, USA). Isoelectric focusing CRC tissue samples were lysed in RIPA buffer containing phosphatase and protease inhibitors (ProteinSimple). The tissue lysates were clarified by centrifugation and protein concentration was measured by using BCA Protein Assay Kit (Pierce/ThermoFisher Scientific). Samples were run in triplicates. Lysates were mixed with ampholyte premix (# 040-972, G2 pH 5-8 or # 040-968, G2 pH 3-10) and fluorescent isoelectoric point (pI) standards (# 040-646, pI Standard Ladder 3) before being loaded into the NanoPro 1000 system (ProteinSimple) for analysis. Isoelectric focusing was performed in capillaries filled with a mixture of cell lysate (0.05–0.2 mg/ml protein), fluorescently labeled pI standards, and ampholytes. The separated proteins were cross-linked onto the capillary wall using UV light, and immobilization was followed by immunoprobing with anti-ERK1/2 (1:50, # 9102), anti-pERK1/2 (# 4377, 1:50) and anti-PLCγ1 (# 2822, 1:50) antibodies from Cell Signaling Technology (Danvers, MA); anti-AKT (# sc-8312, 1:20), p70S6 kinase (# sc-8418, 1:50), and MEK 1/2 (# sc-436, 1:50) antibodies from Santa Cruz Biotechnology Inc. (Dallas, Texas); anti c-SRC (# ab47405, 1:50) antibodies from Abcam; and anti-EGFR (# 05-484, 1:50) antibodies from Merck Millipore (Darmstadt, Germany). Analysis of HSP 70 (# NB600-571, 1:500), Novus Biologicals (Littleton, CO) was run in parallel for normalization. HRP-conjugated secondary antibodies were used, either from ProteinSimple (Goat anti rabbit-HRP IgG, # 041-081 and Goat anti mouse-HRP IgG, # 040-655 both at 1:100) or from Jackson ImmunoResearch (West Grove, PA) (Donkey anti-Rabbit IgG, # 711-035-152 and Donkey anti-Mouse-HRP IgG # 711-035-150, both at 1:300), to detect the signal. In some cases, signal amplification steps were employed by using an amplified rabbit (# 041-126, 1:100) or amplified mouse (# 041-127, 1:100) secondary antibody detection kit (ProteinSimple). The signal was visualized by enhanced chemiluminescence (ECL) and captured by a charge-coupled device (CCD) camera. The digital image was analyzed and peak area quantified with Compass software (ProteinSimple). The peak area of the protein of interest was normalized to the area of heat shock protein 70 (HSP70) in the sample, analyzed in parallel. Lambda phosphatase digestion Some samples were enzymatically dephosphorylated by incubating 8–15 μg of cell lysate with 50 units of lambda phosphatase (# 14-405; Upstate Biotechnology, Charlottesville, VA), for 5-30 min at 30 °C, where incubation time was titrated independently for each signaling component. Digested samples were subjected to immunoblotting or isoelectric focusing as described above. Mutation analysis KRAS pyro-sequencing mutational analysis was performed according to the manufacturer’s protocol for the PyroMark™ Q24 KRAS Pyro kit (QIAGEN GmbH, Hilden, Germany) and the use of PCR primers previously described for KRAS codon 12/13 [31], codon 61 [32], and for BRAF codon 600 [31]. Ten ng genomic DNA from the patients tumor tissue was used for each PCR reaction. Twenty μl PCR product was then subjected to Pyro-sequencing analysis using Streptavidin Sepharose High Performance beads (GE Healthcare, Chicago IL), PyroMark Gold Q96 reagents, PyroMark Q24 2.0.6 software, and a Q24 instrument (QIAGEN). Sequencing primer for KRAS codon 12/13 was 5′-AACTTGTGGTAGTTGGAGCT-3′, for codon 61 5′-TCTTGGATATTCTCGACACAGCAG-3′, and for BRAF codon 600 5′-TGATTTTGGTCTAGCTACA-3′. Due to sub-optimal DNA quality, two samples were not suitable for mutation analysis (denoted “unclear” in the figures). Immunoblotting Ten μg of CRC tissue- or cell lysate was mixed with lithium dodecylsulfate sample buffer and Sample Reducing Agent and heated at 70 °C for 10 min. The proteins were resolved on NuPAGE Novex 4–12 % Bis-Tris SDS PAGE Gel (Life Technologies, Carldsbad, CA) and transferred onto PVDF membranes (Immobilon-P IPVH00010; Merck Millipore). The membranes were blocked by using 5 % (w/v) nonfat dry milk/BSA in TBS with 0.1 % Tween 20 for 1 h at RT, which was followed by incubation over night at 4 °C with primary antibodies pERK 1/2 (# 4377, 1:1000), ERK1/2 (# 9102, 1:1000), SRC pY416 (# 2101, 1:1000), SRC pY527 (# 2105, 1:1000), pAKT (# 4060, 1:1000), AKT (# 9272, 1:1000), PLCγ1 (# 2822, 1:1000), all from Cell Signaling Technology. SRC (# ab47405, 1:1000) and β2M (# ab75853, 1:2000) were from Abcam. EGFR (# 05-484, 1:2000) and GAPDH (# MAB374, 1:1500) from Merck Millipore, α-Tubulin (# T9026, 1:1000) from Sigma-Aldrich (Saint Louis, MI), p70S6 kinase (# sc-8418, 1:2000) from Santa Cruz Biotechnologies Inc, HSP 70 (# NB600-571, 1:1000) from Novus Biologicals. Proteins of interest were detected with HRP-conjugated donkey anti-rabbit IgG antibody (# NA934, 1: 15000) or sheep anti-mouse IgG antibody (# NA931, 1: 15000), visualized with using ECL Prime (# RPN2232) and exposed to either the Hyperfilm ECL (# 28906837) all from GE Healthcare. Signals were visualized using the ChemiDoc™ MP Imaging System (Bio-Rad Laboratories, Herkules, CA) according to the provided protocol. All antibodies used for the isoelectric focusing were tested for specificity by immunoblotting of HUVEC lysates (for AKT, p70S6 kinase, PLCγ1, c-SRC, SRC pY527, ERK1/2, HSP 70 and MEK 1/2) and lysates from A431 cells (#12-302, Merck Millipore) for EGFR (see Additional file 1: Figure S1). Certain antibodies, such as the anti-c-Src antibodies were also validated elsewhere for example at the MD Anderson Functional Proteomics resource (RPPA core facility, see https://www.mdanderson.org/research/research-resources/core-facilities/functional-proteomics-rppacore/antibody-information-and-protocols.html. Statistical analysis The Mann-Whitney U test was used to calculate two-tailed p-values of the null hypothesis that the populations of the two compared features (proteins) are the same. p < 0.05 was considered statistically significant. *, p < 0.05; **, p < 0.01; ***, p < 0.001 and ****, p < 0.0001. The Mann-Whitney test is a conservative, non-parametric test that was chosen to preclude false detections arising from assumptions of data distribution. Identification of tissue signatures For assessment of data sets and the creation and evaluation of convex hulls for classification of the tissue samples based on signatures, see Additional file 1: Figure S3, Characteristics of the data set and errors. Results Regulation of EGFR expression and activity in CRC Whereas activating mutations in the EGFR gene are rare in CRC, protein levels may be increased as a result of gene amplification or through other mechanisms e.g. involving increased translation or decreased internalization and degradation. We used isoelectric focusing for sensitive and high-resolution detection of EGFR expression in tissues, comparing normal mucosa (18 samples) with CRC samples (17 samples from stage II and 16 samples from stage IV). The mutation status of the CRC samples was determined for KRAS and BRAF. Tissues were lysed and, in a robotized procedure, proteins were immobilized to the wall of thin capillaries using UV exposure, followed by incubation with primary and secondary antibodies and ECL-detection, as outlined schematically in Fig. 1a. Tissue lysates and antibodies were loaded at desired concentrations in 384-well plates placed under the capillary holder in the instrument. As shown in Fig. 1b and c, there was no significant difference in the expression levels of EGFR when comparing normal tissue with stage II and IV CRC using isoelectric focusing, although the median was numerically lower in stage IV samples. The peaks corresponding to antibody detection of EGFR were normalized to those of HSP70 run in parallel. There was no correlation between EGFR levels and the KRAS or BRAF mutation status, in this analysis.Fig. 1 Sensitive isoelectric focusing of EGFR in normal mucosa and CRC. a. Schematic outline of the isoelectric focusing procedure. 400 nl of protein lysates from cultured cells or tissues are passed through the capillaries, followed by probing with antibodies and detection using ECL, resulting in an electropherogram. b. Representative electropherogram showing EGFR protein peaks. c. Plot of individual HSP70-normalized peak areas from EGFR electropherograms on normal mucosa or CRC samples. Symbols in plots indicate the mutation status of CRC biopsies: Red; KRAS mutated, green; BRAF mutated, blue; wild type (WT) with regard to KRAS and BRAF, black; unidentifiable (unclear) for KRAS and BRAF Regulation of AKT and p70S6K pathways in CRC Signaling in the PI3K/AKT pathway results in downstream activation of mTOR and p70S6 kinase and ultimately, cell survival and proliferation [33]. The level of AKT expression and activity was first analyzed by immunoblotting on normal mucosa and CRC samples (Fig. 2a). The level of AKT pS473 was elevated in stage II CRC, but the variability was considerable in this conventional analysis. Isoelectric focusing followed by detection of AKT resulted in a reproducible pattern with several peaks, when probed with an antibody against total AKT proteins, AKT1, AKT2 and AKT3 (Fig. 2b). The pattern of AKT-peaks was reminiscent but not identical to that described in previous reports where isoelectric focusing was used to investigate the in vitro regulation of the AKT pathway in cell lines from breast cancer and acute myeloid leukemia [34, 35].Fig. 2 Detection of total AKT protein and phospho-protein by isoelectric focusing. Plots (d-f, h) show values after normalization to HSP70 levels analyzed in parallel in each sample. Symbols in plots: Red; KRAS mutated, green; BRAF mutated, blue; wild type (WT) with regard to KRAS and BRAF, black; unclear for KRAS and BRAF. a. Immunoblotting of selected tissue samples with antibodies against pAKT (AKT pS473) and total AKT protein. Blotting for β2 microglobulin (β2M) was used as a loading control. b. Representative electropherogram showing phosphorylated and non-phosphorylated AKT peaks. Blue and green lines indicate electropherograms of samples digested (green) or not (blue) with lambda phosphatase. Inset; electropherogram showing HSP70 run in parallel. c. Immunoblotting of HUVEC (±VEGF for 15 min) cell lysate with antibodies against pAKT and total AKT protein. Blotting for tubulin was used as to monitor equal loading. Control; without any incubation; Buffer; control lysate incubated with buffer, Phosphatase; lysate incubated with lambda phosphatase enzyme. d. Plot of total AKT peak areas (P1–P8) in the different samples. e. Plot of pAKT peak areas (present before but not after lambda phosphatase treatment; P1–P4 and P6). f. Plot of the ratio of pAKT/AKT peak areas. g. Representative electropherogram of p70S6 kinase expression. h. Plot of p70S6 kinase expression normalized to HSP70 Phospho-specific AKT antibodies did not permit specific detection of protein species in the isoelectric focusing (data not shown). Through lambda phosphatase digestion, however, several pAKT isoforms were identified (Fig. 2b; P1-4 and P6), possibly representing distinct AKT family members phosphorylated on different residues. The optimal conditions for lambda phosphatase digestion were determined by immunoblotting of phosphatase-treated control (HUVEC) cell lysates, which showed that the phosphatase treatment resulted in phosphate stripping without digestion of protein (Fig. 2c). Of note, the levels of pAKT and total AKT as detected in the capillary isoelectric focusing were significantly higher in the CRC tissues compared with normal mucosa (Fig. 2d and e). The ratio of pAKT/AKT did not change, however, indicating that the relative AKT phosphorylation level was not affected by the disease (Fig. 2f). The protein level of p70S6 kinase, a serine/threonine kinase activated downstream of PI3K/AKT, was similar in the cancer samples as compared to normal mucosa (Fig. 2g-h). Upregulation of PLCγ1 protein in CRC stage II and stage IV PLCγ1 is known to activate the RAS pathway to promote cell proliferation via PKC. Conventional immunoblotting for PLCγ1 allowed detection of a very faint band in the tissue lysates of normal samples while CRC stage II showed a prominent upregulation of PLCγ1 protein. In CRC stage IV samples, the signal was slightly lower (Fig. 3a). Capillary isoelectric focusing resulted in two very closely migrating peaks (Fig. 3b) which were both resistant to lambda phosphatase treatment. Antibodies against phosphorylated PLCγ1 failed to yield a signal in the isoelectric focusing (data not shown). Quantification of the combined areas of the two peaks showed a significant increase in PLCγ1 expression in stage II and IV samples (Fig. 3c), in agreement with the immunoblotting data. Moreover, the variability in expression level was higher in the cancer samples than in the normal tissue biopsies. Combined, these data indicate that while total PLCγ1 was upregulated in CRC, there was low or no accumulation of phosphorylated PLCγ1.Fig. 3 Detection of PLCγ1 total protein by isoelectric focusing. a. Immunoblotting of selected tissue samples with antibodies against PLCγ1. Blotting for GAPDH and β2 microglobulin (β2M) were used as loading control. b. Representative electropherogram showing PLCγ1 total protein peaks. Inset; electropherogram showing HSP70 run in parallel. c. Plot of PLCγ1 peak areas in samples from normal tissue, CRC stage II and IV biopsies. Values were normalized to HSP70 levels. Symbols in plots: Red; KRAS mutated, green; BRAF mutated, blue; wild type (WT) with regard to KRAS and BRAF, black; unclear for KRAS and BRAF Decreased c-SRC phosphorylation in CRC We also investigated the expression and activity of c-SRC, as its activity results in the downstream induction of several signaling pathways regulating cell proliferation. An antibody against total c-SRC detected several species upon immunoblotting of normal and stage II samples. In contrast, stage IV samples showed very faint or no expression of c-SRC (Fig. 4a). Moreover, all samples lacked reactivity with antibodies against c-SRC pY418, indicating low or no c-SRC activity in the colon (Fig. 4a, upper panel). Control immunoblotting of lysates from growth factor stimulated cells verified that the anti c-SRC pY418 antibodies recognized the expected 60 kDa species (Fig. 4a, lower panel). Moreover, immunoblotting with antibodies against the inactivating c-SRC pY527 residue revealed prominent bands in both the control cell lysate and in selected CRC samples (Fig. 4a, lower panel). Thus, conventional immunoblotting for total c-SRC and the phosphorylated variants showed a complex and variable pattern.Fig. 4 Detection of c-SRC total protein and phosphorylated forms by isoelectric focusing. Plots (d–f) show values after normalization to HSP70 levels. Symbols in plots: Red; KRAS mutated, green; BRAF mutated, blue; wild type (WT) with regard to KRAS and BRAF, black; unclear for KRAS and BRAF. a. Immunoblotting of selected tissue samples with antibodies against c-SRC pY418, c-SRC pY527 and total SRC protein. For upper panel, loading control β2M was same as Fig. 2a. GAPDH was used as a loading control for lower panel. Control; HUVEC cell lysate was used as a positive control. b. Representative electropherogram showing c-SRC total protein and phosphoprotein peaks. Phosphorylated peaks (blue line) were identified by virtue of their sensitivity to lambda phosphatase digestion (green line). Inset; electropherogram showing HSP70 run in parallel. c. Representative electropherogram showing c-SRC pY527 peaks. d. Plot of combined c-SRC peak (P1–P6) areas in normal mucosa, stage II and stage IV CRC. e. Plot of phosphorylated c-SRC peak (P1–P5) areas. f. Plot of the ratio pSRC/SRC Isoelectric focusing detected six major c-SRC species (Fig. 4b); five peaks with a more acidic isoelectric point disappeared with lambda phosphatase digestion and were collected in one peak with a more basic pI of 6.5 (Fig. 4b). Probing with the c-SRC pY527 antibodies showed that the majority of the pSRC species in peaks (P)1-3,5 contained phosphorylation at the inactivating Y527 (Fig. 4c). The various pY527 antibody-reactive phosphospecies focusing at different pI may correspond to c-SRC variants with different posttranslational modifications such as serine/threonine phosphorylation [36]. We can not exclude that certain molecular species may correspond to c-SRC related proteins, containing highly similar epitopes. However, the normalized peak areas for all peaks (Fig. 4d, denoted “SRC”) showed that c-SRC expression was significantly lower in CRC stages II and IV, compared with normal tissues. The area of the combined “pSRC” peaks P1-P5 (Fig. 4e) was also lower in the CRC samples. Moreover, the ratio of pSRC/SRC (Fig. 4f) was lower in CRC than in normal mucosa, indicating that the level of inactivating pY527 phosphorylation was reduced in the cancer compared with normal tissues. There was no apparent correlation between the decreased levels of pSRC/SRC and KRAS/BRAF mutation status. Decreased level of pERK1, but not expression level, in CRC Growth factors regulate cell proliferation in the RAS pathway by modifying downstream phosphorylation of the serine/threonine kinases ERK1, on T202/Y204, and ERK2, on T185/Y187. Phosphorylated and nuclearly translocated ERK1/2 catalyze phosphorylation and thereby activation of a range of nuclear transcription factors [37, 38]. Immunoblotting for pERK1/2 showed variable expression in normal mucosa, high expression in stage II and lower expression again in stage IV CRC. The levels of pERK1/2 were variable over the panel of immunoblotted samples (Fig. 5a). Isoelectric focusing on the other hand resolved total ERK1/2 into six major peaks representing both phosphorylated and non-phosphorylated ERK isoforms (Fig. 5b). Using a combination of antibodies reactive with both ERK1 and ERK2, antibodies specifically recognizing only one of the two, and, dephosphorylation by lambda phosphatase, the identity of each peak could be mapped (Fig. 5b). Quantification of the normalized peak areas showed no difference in expression levels of ERK1 between normal mucosa and cancer stage II and IV. However, accumulation of pERK1 decreased in the CRC samples compared to the normal tissue resulting in a significantly decreased pERK1/ERK1 ratio (Fig. 5c). Although ERK2 levels increased in the CRC samples, the pERK2/ERK2 ratios remained unchanged (Fig. 5d). The decrease in pERK1 levels dominated over the increase in pERK2 levels, as a cross-reactive pERK1/2 antibody also showed lower phosphoprotein levels in the cancer samples (Fig. 5e).Fig. 5 Detection of ERK1/2 total protein and phosphorylated forms by isoelectric focusing. Plots (c–e) show values after normalization to HSP70 levels. Symbols in plots: Red; KRAS mutated, green; BRAF mutated, blue; wild type (WT) with regard to KRAS and BRAF, black; unclear for KRAS and BRAF. a. Immunoblotting of selected tissue samples with antibodies against pERK1/2 and total ERK1/2 protein. Loading control β2M was same as shown in Fig. 2a. b. Representative electropherogram showing ERK1/2 total protein peaks. c. Plot of individual peak areas from ERK1 (ppERK1+ pERK1 + ERK1) analyses of normal mucosa and CRC stage II and IV (top) and of pERK1 (ppERK1 + pERK1)/ERK1 peak areas (bottom) after normalization for HSP70 run in parallel. d. Plot of normalized ERK2 (ppERK2 + ERK2) protein peaks (top) and pERK2 (ppERK2)/ERK2 (bottom). e. Plot of normalized, combined ERK1/2 total protein peaks (top) and combined pERK1/2 over total ERK1/2 peaks (bottom) Computational selection of proteins to distinguish CRC from normal tissue Since individual pathways associated with epithelial cell proliferation showed a very complex pattern in the CRC tissues, we conducted a computational search for combinations of proteins from several pathways that would allow for the discrimination of normal tissue samples from CRC. The overlap between the convex hulls of the data points from normal tissue and CRC stage II or stage IV was examined for every possible combination of up to three features. In addition to the measured 23 different variants (represented by individual peaks in the electropherograms shown in Figs. 1, 2, 3, 4 and 5) for EGFR, AKT, p70S6K, PLCγ1, c-SRC, ERK1, ERK2, and MEK1/2 (see Additional file 1: Figure S2 for MEK1/2 analyses), we also included 15 features constructed as the sum of phosphorylated or non-phosphorylated forms of the seven proteins and their ratios. For detailed description on computational analyses and machine learning see Additional file 1: Figure S3; Characteristics of the data set and errors. In mathematics, the convex hull of a set is the minimal convex set that covers all points in the set. Applied in this context, the convex hull represents the region in protein space that encompasses all observations for either one of the cancer stage or the normal tissue. As shown by the minimal overlap of the convex hulls in Fig. 6, the combination of total pERK1, SRC peak 6 and p70S6K peak 3, separated normal tissue from CRC II and CRC IV. In other words, these three patterns yield a “signature” that was distinct for normal and cancer tissue and measurement of these proteins was sufficient for classification of a tissue sample as normal or CRC. Only one CRC stage IV sample fell within the convex hull of the normal tissues. The convex hulls of the two CRC stages overlapped implying that the combination used (pERK1, SRC peak 6 and p70S6K peak 3) was not appropriate for classification of the disease stage. Monte Carlo simulations revealed that the separation of the non-cancer versus cancer sets was highly unlikely to occur by chance (p-value <10-6; multiple hypothesis corrected p-value <10-2). Thus, with this strategy, a unique signature for normal tissue versus cancer tissue was obtained.Fig. 6 Convex hulls separating normal, CRC stage II and IV tissues. Convex hulls of the sets of all data points of each tissue class representing total pERK1 (ppERK1 + pERK1) peaks, SRC P6 and p70S6K P3 allowed separation of normal tissues (green) from CRC stage II (blue) and stage IV (red). Each dot represents a computationally analyzed data point Discussion Substantial research efforts over the last decades have resulted in increased understanding of CRC mutations and molecular consequences; still, due to the complexity of the tumor biology and the heterogeneity of the cancer, CRC remains a fatal disease. Here, we show that signaling pathways regulating cell survival and proliferation were differently regulated in CRC tissues compared to normal mucosa. Expression of ERK1 and SRC appeared significantly suppressed in CRC tissues compared with normal mucosa while expression of AKT and PLCγ1 were upregulated. See Table 1 for a summary of the pattern of proliferative CRC signaling identified in this study.Table 1 Summary of changes in signaling components between normal and CRC tissues Signaling component Normal CRC II CRC IV Comment EGFR + + + Similar levels in benign, CRC II and IV. pAKT/AKT + + + Total levels of pAKT and AKT upregulated in CRC but pAKT/AKT ratios were similar in the different samples. p70S6K + + + Similar levels in benign, CRC II and IV. PLCγ1 - +++ + + Low or no PLCγ1 expression in benign samples and higher levels in CRC. pPLCγ1 was not detected in any samples. SRC pY527/SRC +++ + + Low or no SRC pY418 in all samples. Lower SRC pY527/SRC ratios in CRC II and IV compared to benign. pERK1/ERK1 +++ + + Reduced pERK1/ERK1 ratio in CRC II and IV compared to benign. pERK2/ERK2 ++ ++ ++ Similar ratios benign, CRC II and IV. MEK1/2 + + + Similar levels in benign, CRC II and IV. +, ++ and +++ indicate different relative levels or ratios for a particular signaling component when comparing normal, CRCII and CRCIV samples, and should not be applied to compare levels/ratios between the different signaling components Signaling was analyzed using capillary isoelectric focusing, which we found to be superior to conventional immunoblotting in sensitivity and resolution. After loading of samples and antibodies, the processing was robotized, resulting in highly reproducible and sensitive detection. For example, ERK1/2 protein was detected in 2.5 ng of CRC lysate per capillary (corresponding to 6.25 μg/ml total lysate). Moreover, protein variants, phosphorylated at different residues, could be separated and quantified independently. For ERK1/2 proteins, six of the isoforms (pERK1, ppERK1, ERK1, pERK2, ppERK2, ERK2) could be identified and quantified in relation to the house keeping proteins analyzed in parallel. In comparison, conventional immunoblotting run on the same samples required much more protein for each analysis. It often failed to resolve protein phospho-variants and reproducibility was low, in part due to problems with transfer of proteins to the filter. Ongoing efforts include adapting the isoelectric focusing protocol for the detection of signal transducers in formalin-fixed, paraffin-embedded samples to make the procedure applicable in clinical routines. Using the isoelectric focusing strategy, several important observations were made that can be related to earlier reports on CRC signaling (see also summary in Table 1):I) AKT: In agreement with our findings on increased AKT protein expression in the CRC tissues, colorectal adenomas and carcinomas frequently overexpress AKT [39] at an early stage in the disease. Moreover, other components in the PI3K/AKT pathway are affected in CRC. The most common event is a loss of expression of, or mutation in PTEN, which occurs in close to 50 % of the premalignant lesions [40]. II) PLCγ1: Studies on a limited number of CRC samples showed increased PLCγ1 protein levels whereas other PLC family members, PLCβ1 and PLCδ1, remained unaffected [41]. However, whether the increased protein levels are accompanied by increased phospholipase activity in CRC remains unclear. Phosphorylation of PLCγ1 is known to induce its catalytic activity however, we failed to detect phosphorylated PLCγ1 in the CRC samples studied here. III) c-SRC is a key signal transducer whose activity may initiate most, if not all, other pathways related to cell proliferation [42], and the expression and activity of c-SRC have been associated with CRC progression [15, 16, 42]. However, in several studies, c-SRC activity has been analyzed using an in vitro immune complex kinase assay on cell lines, rather than on clinical samples [43–45]. The lack of pY418 phosphorylated c-SRC and the decrease in expression in disease shown here (Fig. 4), indicate that c-SRC does not drive CRC tumor cell proliferation. Also, pathways potentially induced as a consequence of c-SRC activation in CRC, such as the Scatter factor/c-Met pathway, may not be crucial [46]. c-SRC kinase activity is regulated by tyrosine phosphorylation/dephosphorylation. We detected c-SRC pY527 in all samples, although the amount decreased in disease. As there was no parallel increase in c-SRC pY418, it appears that overall, there is limited c-SRC activity in CRC. The decrease in pY527 levels may depend on phosphatase activity with c-SRC being dephosphorylated e.g. by the tyrosine phosphatase PTPRO [47]. Apart from the well characterized positive regulatory pY418 and negative regulatory pY527, there are other phosphorylation sites in c-SRC including pS17 and pY215 whose functions have remained unclear [36]. The many phospho-SRC peaks identified in the isoelectric focusing indicate that, in CRC, c-SRC can become modified at yet additional sites. However, as the critical pY418 is lacking, it is questionable whether c-SRC is a suitable target for CRC therapy. Another complicating aspect of studying c-SRC’s role in cancer biology is the high degree of structural relatedness with other SRC family tyrosine kinases (SFKs), first and foremost the ubiquitously expressed FYN and YES. Thus, we cannot exclude that c-SRC, YES, and FYN phosphoproteins may all have been detected by the c-SRC reagents used here, due to the highly conserved phosphorylation sites in all three members. Overall, insight on the role of the different SFKs in CRC is lacking. IV) ERK1/2: Aberrant colon crypt foci, which are believed to predict a malignant process, were analyzed using a similar methodology to that applied in this study, revealing elevated levels of both pERK1 and pERK2 irrespective of KRAS and BRAF mutation status [48]. ERK1 and ERK2 are highly related structurally and are largely co-regulated and indeed, in many aspects, redundant. However, ERK2, but not ERK1, has been shown to contribute to RAS-induced oncogenic signaling [49], and yet, ERK1 has been implicated in the negative regulation of ERK2 [50]. Therefore, the reduced pERK1 levels in CRC that we describe here may unleash ERK2 activity, resulting in increased oncogenic signaling in primary tumors. Regulation of ERK1/2 signaling is truly complex, with scaffold proteins, including KSR1/2, IQGAP1, MP1, and β-Arrestin1/2, participating in the regulation of the ERK1/2 MAP kinase cascade [26]. Furthermore, ERK1/2 are dephosphorylated by several different phosphatases [51] that may be differently expressed. Several decades of ambitious basic and clinical research have demonstrated the challenges in identifying reliable biomarkers in cancer. Challenges include the complexity of the primary tumor tissue consisting of, apart from the tumor cells, a range of host-derived endothelial, fibroblast and inflammatory cells; potential differences between the primary tumor and metastasis; and the possibility that biopsies may not be representative. In this study, the proportion of tumor cells ranged from about 30–60 % in most samples, based on the estimation of mutated DNA/total DNA in the samples (data not shown). An important conclusion from the current study is that the combination of several features from the conducted analyses allows a very high confidence in classifying the tissues as normal or cancerous. The particular combination of pERK1, SRC peak 6, and p70S6K peak 3 selected here to distinguish cancer tissue from normal tissue, may or may not indicate convergence of the included pathways in CRC signaling. The main objective of the selection was to allow unbiased diagnosis. Thus, we propose that reliable prognostic and diagnostic biomarkers should be designed using complex patterns rather than a single molecular or genetic marker. For clinical translation, the isoelectric focusing analyses can easily be made routine and scaled up. For example, 96 unique samples could be run in parallel to yield information on three or more selected pathways (by mixing several appropriate antibodies yielding non-overlapping patterns) in a 10 h run in a robotized set-up. Combined with the powerful computational evaluation to identify sets of signaling components showing significant characteristics, this strategy could prove to be clinically feasible for diagnostic purposes beyond the treatment of CRC. Based on the results obtained this far, we predict that the measurement of seven protein forms (i.e. selected peaks from the electropherograms) would be sufficient for the correct classification of both non-cancerous versus cancerous tissue as well as for the CRC grade. Moreover, analysis of a larger cohort of samples, combined with information on chosen therapy and disease outcomes, would allow the use of supervised learning for identification of clinically relevant subtypes. Conclusions Highly sensitive robotized isoelectric focusing was established and a wide range of signal transduction pathway antibodies were validated. The set-up was shown to allow detection of signaling status also in extremely scarce samples, not amenable to conventional analysis performed in parallel. The study revealed dysregulated signal transduction in several proliferative pathways in human colorectal cancer tissue, which did not correlate with the mutation status. Computational analysis was used to identify signal activities consisting of three components that, when combined, could accurately identify normal mucosa from cancer. The study suggests that such combinations of different signalling activities could serve as predictive or prognostic complex biomarkers. Additional file Additional file 1: Figure S1. Validation of antibodies used in the study by conventional immunoblotting. All antibodies showed immunoreactivity with the expected molecular species, in conventional immunoblotting on endothelial lysates. Figure S2. Detection of MEK1/2 protein by isoelectric focusing. There was no significant difference in MEK protein expression between normal, CRCII and CRCIV tissues. Detailed description of computational analyses; “Characterization of the data set and errors”. Figure S3. Distribution function for data subsets by Monte Carlo simulation. (DOCX 1215 kb) Abbreviations CCDCharge-coupled device CRCColorectal cancer EGFREpidermal growth factor receptor ECLEnhanced chemiluminescence ERKExtracellular regulated kinase HSP70Heat shock protein 70 IP3Inositol 1,4,5-trisphosphate MEKMitogen-activated protein kinase kinase mTORMammalian target of rapamycin complex PDK1Phosphatidylinositol-dependent kinase 1 pIIsoelectric point PIP2Phosphatidylinositol 4,5 bisphosphate PLCγ1Phospholipase Cγ1 PTENPhosphatase and tensin homolog WTWild type Acknowledgements The authors gratefully acknowledge Ross Smith, Uppsala University, for linguistic revision and Marcus Thuresson, Statisticon AB, for revision of statistical analyses. Funding These studies were supported by grants to Lena Claesson-Welsh from the Swedish Cancer foundation (CAN2013/661), Worldwide Cancer Research (13-1295) and the Knut and Alice Wallenberg foundation (KAW 2015.0275). Narendra Padhan was supported by a Post Doctoral fellowship (Gustaf Adolf Johansson Foundation) from Uppsala University. Torbjörn Nordling was supported by the Swedish strategic research program eSSENCE and by the Ministry of Science and Technology, Taiwan (grant 105-2218-E-006-016-MY2). Availability of data and materials For complete data sets used for computational analyses, see data deposited at Dryad (datadryad.org), doi:10.5061/dryad.h35p4. Authors’ contributions All authors planned the study. NP established the isoelectric focusing protocols and carried out the experimental work. PÅ, HB, PN and MS guided the clinical aspects of the study including collection of tissue samples and mutational analyses. TEMN and SN guided the computational analyses of the study. LCW and NP wrote the manuscript with input from all authors. All authors approved the final version of the manuscript. Authors’ informations NP is a vascular/cancer biologist with a PhD from the Jawaharlal Nehru University, New Delhi, India, and currently a senior postdoc with Prof. Lena Claesson-Welsh. TEMN is a Systems Biologist with a PhD in Automatic Control from KTH Royal Institute of Technology, Stockholm, Sweden; currently an Assistant Professor at the National Cheng Kung University, Tainan, Taiwan. MS is a cancer and molecular biology researcher with a PhD from Uppsala University, Sweden, and currently a development manager at the molecular pathology unit at Uppsala University Hospital. PÅ is an MD and expert in colorectal surgery, with a PhD in neuroscience from Karolinska Institute, Sweden. HB is an MD, PhD, working as a colorectal surgeon and associate professor at the Department of surgical sciences, Uppsala University Hospital, Sweden. PN is an MD and expert in clinical oncology with a PhD from Uppsala University, Sweden, and a professor in Oncology at this university. SN is a cancer systems biologist with a PhD from Gothenburg University, Sweden; he is currently an associate professor and group leader at Uppsala University, Sweden. LCW is a vascular/cancer biologist with expertise in signal transduction. She is a professor in Medical Biochemistry at Uppsala University. Competing interests The authors declare that they have no competing interests. Consent for publication Consent to publish was granted in the ethical permit (see above). Ethics approval and consent to participate This study was approved by the regional ethics committee of the University of Uppsala, Sweden (Reference number 2007/005 and 2000/001). All patients whose samples were used in this study gave informed consent to the storage of tissue, isolation of protein, the use of the material in research projects. ==== Refs References 1. Ferlay J Steliarova-Foucher E Lortet-Tieulent J Rosso S Coebergh JW Comber H Forman D Bray F Cancer incidence and mortality patterns in Europe: estimates for 40 countries in 2012 Eur J Cancer 2013 49 6 1374 1403 10.1016/j.ejca.2012.12.027 23485231 2. Sargent DJ Patiyil S Yothers G Haller DG Gray R Benedetti J Buyse M Labianca R Seitz JF O’Callaghan CJ End points for colon cancer adjuvant trials: observations and recommendations based on individual patient data from 20,898 patients enrolled onto 18 randomized trials from the ACCENT Group J Clin Oncol 2007 25 29 4569 4574 10.1200/JCO.2006.10.4323 17876008 3. Tol J Koopman M Cats A Rodenburg CJ Creemers GJ Schrama JG Erdkamp FL Vos AH van Groeningen CJ Sinnige HA Chemotherapy, bevacizumab, and cetuximab in metastatic colorectal cancer N Engl J Med 2009 360 6 563 572 10.1056/NEJMoa0808268 19196673 4. Normanno N Tejpar S Morgillo F De Luca A Van Cutsem E Ciardiello F Implications for KRAS status and EGFR-targeted therapies in metastatic CRC Nat Rev Clin Oncol 2009 6 9 519 527 10.1038/nrclinonc.2009.111 19636327 5. Samowitz WS Curtin K Schaffer D Robertson M Leppert M Slattery ML Relationship of Ki-ras mutations in colon cancers to tumor location, stage, and survival: a population-based study Cancer Epidemiol Biomarkers Prev 2000 9 11 1193 1197 11097226 6. Douillard JY Oliner KS Siena S Tabernero J Burkes R Barugel M Humblet Y Bodoky G Cunningham D Jassem J Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer N Engl J Med 2013 369 11 1023 1034 10.1056/NEJMoa1305275 24024839 7. Brink M de Goeij AF Weijenberg MP Roemen GM Lentjes MH Pachen MM Smits KM de Bruine AP Goldbohm RA van den Brandt PA K-ras oncogene mutations in sporadic colorectal cancer in The Netherlands Cohort Study Carcinogenesis 2003 24 4 703 710 10.1093/carcin/bgg009 12727799 8. Sullivan KM Kozuch PS Impact of KRAS Mutations on Management of Colorectal Carcinoma Pathol Res Int 2011 2011 219309 10.4061/2011/219309 9. Matallanas D Birtwistle M Romano D Zebisch A Rauch J von Kriegsheim A Kolch W Raf family kinases: old dogs have learned new tricks Genes Cancer 2011 2 3 232 260 10.1177/1947601911407323 21779496 10. Garnett MJ Marais R Guilty as charged: B-RAF is a human oncogene Cancer Cell 2004 6 4 313 319 10.1016/j.ccr.2004.09.022 15488754 11. Rajagopalan H Bardelli A Lengauer C Kinzler KW Vogelstein B Velculescu VE Tumorigenesis: RAF/RAS oncogenes and mismatch-repair status Nature 2002 418 6901 934 10.1038/418934a 12198537 12. Toton E Ignatowicz E Skrzeczkowska K Rybczynska M Protein kinase Cepsilon as a cancer marker and target for anticancer therapy Pharmacol Rep 2011 63 1 19 29 10.1016/S1734-1140(11)70395-4 21441608 13. Krasinskas AM EGFR Signaling in Colorectal Carcinoma Pathol Res Int 2011 2011 932932 10.4061/2011/932932 14. Sen B Johnson FM Regulation of SRC family kinases in human cancers J Signal Transduct 2011 2011 865819 10.1155/2011/865819 21776389 15. Talamonti MS Roh MS Curley SA Gallick GE Increase in activity and level of pp60c-src in progressive stages of human colorectal cancer J Clin Invest 1993 91 1 53 60 10.1172/JCI116200 7678609 16. Cartwright CA Meisler AI Eckhart W Activation of the pp60c-src protein kinase is an early event in colonic carcinogenesis Proc Natl Acad Sci U S A 1990 87 2 558 562 10.1073/pnas.87.2.558 2105487 17. Aligayer H Boyd DD Heiss MM Abdalla EK Curley SA Gallick GE Activation of Src kinase in primary colorectal carcinoma: an indicator of poor clinical prognosis Cancer 2002 94 2 344 351 10.1002/cncr.10221 11900220 18. Kim LC Song L Haura EB Src kinases as therapeutic targets for cancer Nat Rev Clin Oncol 2009 6 10 587 595 10.1038/nrclinonc.2009.129 19787002 19. Creedon H Brunton VG Src kinase inhibitors: promising cancer therapeutics? Crit Rev Oncog 2012 17 2 145 159 10.1615/CritRevOncog.v17.i2.20 22471705 20. Girotti MR Pedersen M Sanchez-Laorden B Viros A Turajlic S Niculescu-Duvaz D Zambon A Sinclair J Hayes A Gore M Inhibiting EGF receptor or SRC family kinase signaling overcomes BRAF inhibitor resistance in melanoma Cancer Discov 2013 3 2 158 167 10.1158/2159-8290.CD-12-0386 23242808 21. Song MS Salmena L Pandolfi PP The functions and regulation of the PTEN tumour suppressor Nat Rev Mol Cell Biol 2012 13 5 283 296 22473468 22. Yamada KM Araki M Tumor suppressor PTEN: modulator of cell signaling, growth, migration and apoptosis J Cell Sci 2001 114 Pt 13 2375 2382 11559746 23. Manning BD Cantley LC AKT/PKB signaling: navigating downstream Cell 2007 129 7 1261 1274 10.1016/j.cell.2007.06.009 17604717 24. Voutsina A Tzardi M Kalikaki A Zafeiriou Z Papadimitraki E Papadakis M Mavroudis D Georgoulias V Combined analysis of KRAS and PIK3CA mutations, MET and PTEN expression in primary tumors and corresponding metastases in colorectal cancer Mod Pathol 2013 26 2 302 313 10.1038/modpathol.2012.150 22936063 25. Bohanes P LaBonte MJ Winder T Lenz HJ Predictive molecular classifiers in colorectal cancer Semin Oncol 2011 38 4 576 587 10.1053/j.seminoncol.2011.05.012 21810517 26. Kolch W Coordinating ERK/MAPK signalling through scaffolds and inhibitors Nat Rev Mol Cell Biol 2005 6 11 827 837 10.1038/nrm1743 16227978 27. Stephen AG Esposito D Bagni RK McCormick F Dragging ras back in the ring Cancer Cell 2014 25 3 272 281 10.1016/j.ccr.2014.02.017 24651010 28. Cunningham D Atkin W Lenz HJ Lynch HT Minsky B Nordlinger B Starling N Colorectal cancer Lancet 2010 375 9719 1030 1047 10.1016/S0140-6736(10)60353-4 20304247 29. Bardelli A Siena S Molecular mechanisms of resistance to cetuximab and panitumumab in colorectal cancer J Clin Oncol 2010 28 7 1254 1261 10.1200/JCO.2009.24.6116 20100961 30. de Wit M Fijneman RJ Verheul HM Meijer GA Jimenez CR Proteomics in colorectal cancer translational research: biomarker discovery for clinical applications Clin Biochem 2013 46 6 466 479 10.1016/j.clinbiochem.2012.10.039 23159294 31. Richman SD Seymour MT Chambers P Elliott F Daly CL Meade AM Taylor G Barrett JH Quirke P KRAS and BRAF mutations in advanced colorectal cancer are associated with poor prognosis but do not preclude benefit from oxaliplatin or irinotecan: results from the MRC FOCUS trial J Clin Oncol 2009 27 35 5931 5937 10.1200/JCO.2009.22.4295 19884549 32. Loupakis F Ruzzo A Cremolini C Vincenzi B Salvatore L Santini D Masi G Stasi I Canestrari E Rulli E KRAS codon 61, 146 and BRAF mutations predict resistance to cetuximab plus irinotecan in KRAS codon 12 and 13 wild-type metastatic colorectal cancer Br J Cancer 2009 101 4 715 721 10.1038/sj.bjc.6605177 19603018 33. Hernandez-Aya LF Gonzalez-Angulo AM Targeting the phosphatidylinositol 3-kinase signaling pathway in breast cancer Oncologist 2011 16 4 404 414 10.1634/theoncologist.2010-0402 21406469 34. Sabnis H Bradley HL Bunting ST Cooper TM Bunting KD Capillary nano-immunoassay for Akt 1/2/3 and 4EBP1 phosphorylation in acute myeloid leukemia J Transl Med 2014 12 166 10.1186/1479-5876-12-166 24923301 35. Iacovides DC Johnson AB Wang N Boddapati S Korkola J Gray JW Identification and quantification of AKT isoforms and phosphoforms in breast cancer using a novel nanofluidic immunoassay Mol Cell Proteomics 2013 12 11 3210 3220 10.1074/mcp.M112.023119 23929892 36. Roskoski R Jr Src kinase regulation by phosphorylation and dephosphorylation Biochem Biophys Res Commun 2005 331 1 1 14 10.1016/j.bbrc.2005.03.012 15845350 37. Johnson GL Lapadat R Mitogen-activated protein kinase pathways mediated by ERK, JNK, and p38 protein kinases Science 2002 298 5600 1911 1912 10.1126/science.1072682 12471242 38. Gerdes MJ Sevinsky CJ Sood A Adak S Bello MO Bordwell A Can A Corwin A Dinn S Filkins RJ Highly multiplexed single-cell analysis of formalin-fixed, paraffin-embedded cancer tissue Proc Natl Acad Sci U S A 2013 110 29 11982 11987 10.1073/pnas.1300136110 23818604 39. Roy HK Olusola BF Clemens DL Karolski WJ Ratashak A Lynch HT Smyrk TC AKT proto-oncogene overexpression is an early event during sporadic colon carcinogenesis Carcinogenesis 2002 23 1 201 205 10.1093/carcin/23.1.201 11756242 40. Waniczek D Snietura M Mlynarczyk-Liszka J Piglowski W Kopec A Lange D Rudzki M Arendt J PTEN expression profiles in colorectal adenocarcinoma and its precancerous lesions Pol J Pathol 2013 64 1 15 20 10.5114/pjp.2013.34598 23625595 41. Noh DY Lee YH Kim SS Kim YI Ryu SH Suh PG Park JG Elevated content of phospholipase C-gamma 1 in colorectal cancer tissues Cancer 1994 73 1 36 41 10.1002/1097-0142(19940101)73:1<36::AID-CNCR2820730108>3.0.CO;2-5 8275435 42. Chen J Elfiky A Han M Chen C Saif MW The role of Src in colon cancer and its therapeutic implications Clin Colorectal Cancer 2014 13 1 5 13 10.1016/j.clcc.2013.10.003 24361441 43. Weber TK Steele G Summerhayes IC Differential pp60c-src activity in well and poorly differentiated human colon carcinomas and cell lines J Clin Invest 1992 90 3 815 821 10.1172/JCI115956 1381724 44. Cartwright CA Kamps MP Meisler AI Pipas JM Eckhart W pp60c-src activation in human colon carcinoma J Clin Invest 1989 83 6 2025 2033 10.1172/JCI114113 2498394 45. Bolen JB Veillette A Schwartz AM DeSeau V Rosen N Activation of pp60c-src protein kinase activity in human colon carcinoma Proc Natl Acad Sci U S A 1987 84 8 2251 2255 10.1073/pnas.84.8.2251 2436227 46. Herynk MH Zhang J Parikh NU Gallick GE Activation of Src by c-Met overexpression mediates metastatic properties of colorectal carcinoma cells J Exp Ther Oncol 2007 6 3 205 217 17552361 47. Asbagh LA Vazquez I Vecchione L Budinska E De Vriendt V Baietti MF Steklov M Jacobs B Hoe N Singh S The tyrosine phosphatase PTPRO sensitizes colon cancer cells to anti-EGFR therapy through activation of SRC-mediated EGFR signaling Oncotarget 2014 5 20 10070 10083 10.18632/oncotarget.2458 25301722 48. Drew DA Devers T Horelik N Yang S O’Brien M Wu R Rosenberg DW Nanoproteomic analysis of extracellular receptor kinase-1/2 post-translational activation in microdissected human hyperplastic colon lesions Proteomics 2013 13 9 1428 1436 10.1002/pmic.201200430 23467982 49. Shin S Dimitri CA Yoon SO Dowdle W Blenis J ERK2 but not ERK1 induces epithelial-to-mesenchymal transformation via DEF motif-dependent signaling events Mol Cell 2010 38 1 114 127 10.1016/j.molcel.2010.02.020 20385094 50. Lloyd AC Distinct functions for ERKs? J Biol 2006 5 5 13 10.1186/jbiol46 16879721 51. Cagnol S Rivard N Oncogenic KRAS and BRAF activation of the MEK/ERK signaling pathway promotes expression of dual-specificity phosphatase 4 (DUSP4/MKP2) resulting in nuclear ERK1/2 inhibition Oncogene 2013 32 5 564 576 10.1038/onc.2012.88 22430215

PMC005xxxxxx/PMC5000423.txt

==== Front BMC PsychiatryBMC PsychiatryBMC Psychiatry1471-244XBioMed Central London 100610.1186/s12888-016-1006-3Research ArticleLipid profiles in schizophrenia associated with clinical traits: a five year follow-up study http://orcid.org/0000-0003-0328-9092Solberg Dag K. +47-97-04-74-60dagksol@online.no 12Bentsen Håvard havard.bentsen@diakonsyk.no 2Refsum Helge helge.refsum@diakonsyk.no 2Andreassen Ole A. o.a.andreassen@medisin.uio.no 341 Institute for Military Psychiatry, Norwegian Defense Medical Services, Pb 1550 Sentrum, 0015 Oslo, Norway 2 Center for Psychopharmacology, Diakonhjemmet Hospital, Oslo, Norway 3 NORMENT, KG Jebsen Centre for Psychosis Research, Institute of Clinical Medicine, University of Oslo, Oslo, Norway 4 Division of Mental Health and Addiction, Oslo University Hospital, Oslo, Norway 25 8 2016 25 8 2016 2016 16 1 29915 5 2016 18 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Alterations in serum and membrane lipids may be involved in schizophrenia pathophysiology. It is not known whether lipid profiles are associated with disease severity or current symptom level. Methods Clinical and lipid data were gathered from 55 patients with schizophrenia admitted to psychiatric emergency wards in an acute stage of the disease (T1). The patients were re-examined after 5 years at a stable phase (T2). The clinical assessments included Positive and Negative Syndrome Scale (PANSS total, positive, negative) and Global Assessment of Functioning (GAF S, symptom and F, function). Serum lipids (cholesterol and triglyceride) and membrane polyunsaturated fatty acids (PUFA, LCPUFA) were measured. Healthy controls were recruited among hospital workers. Results Serum triglyceride was significantly higher in patients with schizophrenia compared to healthy controls both at T1 and T2 (p < 0.001), while serum cholesterol did not differ significantly. The levels of serum lipids in patients remained stable over time. At T1, serum lipids and symptoms were not significantly correlated. At T2, higher serum lipids were associated with more severe symptoms and poorer functioning. Higher serum lipid levels at T1 were associated with more severe symptoms and poorer functioning at T2; cholesterol with GAF-S (p < 0.05), triglyceride with PANSS total (p < 0.05), GAF-S (p < 0.01) and GAF-F (p < 0.01). Membrane lipids were significantly lower in the patient group compared to healthy controls at T1 (PUFA p < 0.001, LCPUFA p < 0.001), but not at T2. Membrane lipids were not significantly correlated with symptoms at T1, but significantly associated with negative symptoms and functioning at T2 as previously reported. Conclusions The present findings suggest different roles of membrane and serum lipids in schizophrenia pathophysiology. To further elucidate the relation of lipid biology to disease traits, replication in independent studies of longitudinal samples are warranted. http://dx.doi.org/10.13039/501100006095Helse Sør-Øst RHF2005132Research Council of Norway223273Josef and Haldis Andresen's legacyDiakonhjemmet Hospital, Norwayissue-copyright-statement© The Author(s) 2016 ==== Body Background Abnormal lipid biology may play a significant role in the pathophysiology of schizophrenia. Most studies show that patients with schizophrenia have higher levels of serum lipids (cholesterol and triglyceride) than a healthy population [1, 2]. This dyslipidemia has been regarded as a result of antipsychotic medication and lifestyle factors [3], but dyslipidemia has also been demonstrated in unmedicated schizophrenia patients [4–7]. Altered metabolism of membrane lipids (polyunsaturated fatty acids, PUFA) is another aspect of lipid biology suggested to be involved in schizophrenia pathophysiology [8]. Lower levels of PUFA in cell membranes have been found in schizophrenia [9, 10], both in the acute and chronic stage of the disease [11]. The course and outcome of schizophrenia is regarded as heterogeneous. The nature of the relation between lipid profiles, lipid metabolism and clinical characteristics of schizophrenia is mainly unknown. Particularly, it is not known whether lipid profiles are associated with the disease itself, and / or current symptoms. Conflicting findings of associations between lipid levels and symptom severity may represent fluctuations of lipid levels as the disease progresses [12, 13]. It is possible that lipid levels are stable while symptoms, especially positive psychotic symptoms, fluctuate during the course of the disease. The presence of abnormal lipid metabolism from the onset of the disease that remains stable independent of disease symptoms and antipsychotic treatment, may suggest that lipid abnormalities are a disease trait, and thus involved in the pathophysiological development of the disease, as suggested for cholesterol [14]. Lipids that are aberrant during an acute psychotic episode of schizophrenia and normalized after the acute episode may indicate a role for lipids in relation to disease symptoms, which has been suggested for membrane lipids (PUFA) [15]. The role of lipid biology in relation to disease symptoms can best be investigated in a longitudinal study, following patients during different stages of the disease. Several lines of evidence suggest that the pathophysiology of schizophrenia involves immune- and inflammatory pathways, integrated with redox-regulation [16, 17]. It has been suggested that the composition of membrane lipids is abnormal [18, 19], potentially due to disturbed redox-regulation [20, 21]. Oxidative stress can also affect serum lipids and cause dyslipidemia [22–24]. In schizophrenia, levels of both serum- and membrane lipids seem aberrant [9, 25, 26] Thus, an alteration in redox-regulation can be a common factor linking abnormalities of both serum and membrane lipids in schizophrenia. A change in membrane lipid composition in neuronal cells can affect neurotransmission, symptoms and behavior in schizophrenia [27]. Both serum and membrane lipids have been found to predict the outcome of treatment [28, 29]. We hypothesize that both serum lipid and membrane lipid alterations may be involved in the pathophysiology of schizophrenia. We have reported earlier that symptom levels were positively associated with two types of lipids, both serum lipids and membrane lipids in patients with schizophrenia [26]. How these relationships change over time is unknown. Here we investigate with a longitudinal design, using a sample with repeated assessment, if the lipid profiles vary in relation to clinical characteristics; positive and negative symptoms (PANSS) and general symptoms and functioning (GAF). We hypothesize that there is a core abnormality in both membrane and serum lipid systems in schizophrenia, reflected in abnormal membrane and serum lipid levels, and this is independent of disease phase. In order to test this hypothesis, we examined a group of patients and healthy controls at baseline and after 5 years follow-up. The aims of the current study are first to determine if there are differences in lipid profiles (serum lipids and membrane lipids) between people with schizophrenia and healthy controls and if they are stable over a 5 years period. The second aim is to explore the relationship between lipid profiles and clinical characteristics during the course of schizophrenia. We measured the levels of serum and membrane lipids at admission to emergency psychiatric wards and after 5 years follow-up in out-patients clinics or at long-term care facilities, and their association with the disease and clinical symptoms. Repeated measures of lipid levels during the course of schizophrenia and their relationship with clinical symptoms may elucidate whether lipid profiles are associated with stable disease characteristics (traits). Methods Participants In a longitudinal study, socioeconomic, clinical and biological data were gathered from a group of 55 patients with schizophrenia and schizoaffective disorders. The patients were a sub-sample of a group of 99 patients participating in a trial of an omega-3 fatty acid and antioxidants in schizophrenia [30]. The patients were recruited when admitted to psychiatric emergency wards in southern Norway in 2001 to 2003 (T1) [30]. They were examined again between 2006 and 2010 (T2), with a mean follow-up time of 61 months. Of the 44 patients not included in the follow-up study, 21 did not wish to participate, twelve were not located, nine were dead, and two had moved to other regions of Norway. The sub-sample of 55 patients did not statistically differ from the main sample of 99 patients with regard to demographics or main outcome variables at T1. When examined at the inclusion (T1), the patients were admitted to an emergency psychiatric ward and thus considered in an acute stage of the disease. Five years later (T2), the patients were treated at out-patient clinics or at psychiatric long term care facilities, and considered to be in a more stable chronic stage. The patients were screened for somatic illness at inclusion and at follow-up. The data from an overlapping sample at T2 have been presented earlier [26]. In addition, healthy controls were recruited among hospital employees from the same age group as patients. At T1, 20 healthy controls were included. At T2, 51 healthy controls were examined. Of these, 16 were a subsample from T1, and an additional 35 healthy participants were included as healthy controls at T2. To be included as healthy controls, participants and their first-degree relatives could not have any ongoing or past severe psychiatric disorder. This was determined with an interview assessing severe mental illness, and screening for ongoing and past psychiatric disorder using the Mini International Neuropsychiatric Interview (MINI). In addition, their physical health was assessed with self-report and short screening interview addressing current or history of somatic illness in the healthy control participants. Demographics of patients and healthy controls are reported in Table 1. The study was approved by the Regional Committee for Medical Research Ethics. All participants gave written informed consent. Inclusion of patients ended in June of 2010 and results were submitted in 2015. This delay was partly caused by technological difficulties which led to re-analyzes and data analysis could not start until late 2011, and partly due to clinical duties for the first author.Table 1 Demographics Patients T1 Healthy controls T1 Patients T2 Healthy controls T2 n = 55 n = 20 n = 55 n = 51¤ Age 26.5 ± 6.1 31.1 ± 5.3 31.3 ± 5.7 33.0 ± 6.1 Sex (% male) 69.1 55.0 69.1 54.9 Smokers (%) 31 (56.4) 7 (35.0) 29 (52.7) 9 (17.6) Primary school (%) a 14 (25.5) 3 (15.0) 9 (17.0) 1 (2.0) Secondary school (%) 30 (54.5) 8 (40.0) 30 (56.6) 8 (16.0) University / college (%) 11 (20.0) 9 (45.0) 14 (26.4) 41 (82.0) PANSS total 81 (71,93) 82 (57,101) PANSS negative 22 (16,27) 25 (18,31) PANSS positive 17 (13,21) 16 (11,19) GAF-S 35 (30,40) 45 (37,58) GAF-F 37 (31,40) 47 (38,58) PANSS positive and negative symptoms scale PANSS total positive and negative symptoms scale PANSS positive positive component: items P1 + P3 + P5 + P6 + G9 PANSS negative negative component: items N1 + N2 + N3 + N4 + N6 + G7 + G8 + G16 GAF global assessment of functioning, S symptoms, F functioning Age mean ± standard deviation PANSS, GAF: median (25,75 percentiles) a data missing from patient T2 n = 2, control T2 = 1 ¤ 16 healthy controls from T1, 35 additional healthy controls included at T2 Clinical assessment All patients were diagnosed with the Structural Clinical Interview for DSM-IV (SCID). To measure the severity of symptoms the Positive and Negative Syndrome Scale Structured Interview Version (SCI-PANSS), and Global Assessment of Functioning (GAF) were used. In addition to PANSS total, we reported positive and negative PANSS components from a model established by van der Gaag et al. [31]. The split version of GAF was used, which includes Symptom (GAF-S) and Functioning (GAF-F) scales [32]. At T1 the patients were assessed by a group of 16 investigators from the participating hospitals. DKS was among the investigators at T1, who underwent trainings sessions and the interrater reliability was assessed by rating ten (median) videotaped interviews [33]. At T2, all patients were assessed by the same clinical investigator (DKS). Biochemical assays Blood for lipid analyses was sampled after overnight fasting. Serum lipids and membrane lipids were measured. At T1 serum lipids were not obtained from healthy controls. Serum lipids (cholesterol and triglyceride) were analyzed at the Department of Clinical Chemistry at Aker Hospital (T1) and Diakonhjemmet Hospital (T2) with standard enzymatic methods from Roche Diagnostics Norge AS, Oslo, Norway. For analyses of polyunsaturated fatty acids washed blood cells were stored at−70 °C and sent within 3 months in dry ice to Mylnefield Research Services LTD, Dundee, United Kingdom, who did the analysis. The lipids were extracted, converted into fatty acid methyl esters, and analyzed by gas chromatography, yielding fatty acid profiles. In total 28 species of fatty acids from C14:0 to C24:1 is reported as micrograms per gram of RBC (red blood cells). The sum of omega-3 fatty acids is C18:3 + C18:4 + C20:3 + C20:5 + C22:5 + C22:6. The sum of omega-6 fatty acids is C18:2 + C18:3 + C20:2 + C20:3 + C20:4 + C22:4 + C22:5. The sum of omega-3 and omega-6 is named polyunsaturated fatty acids (PUFA). The sum of PUFA with 20 or 22 carbon atoms is named long-chain polyunsaturated fatty acids (LCPUFA). Pharmacological treatment Use of antipsychotic medication may affect the levels of serum lipids. The current and previous use of antipsychotics and other pharmacological agents in the patients were registered from interviews and medical records information. At T2, detectable levels of antipsychotic drug in serum samples were measured to control for adherence (therapeutic drug monitoring). Use of medication for dyslipidemia was registered. Statistics All statistical analyses were performed using SPSS version 20 (SPSS Inc, Chicago, IL, USA / IBM, New York, USA). Demographical and clinical variables are presented as average values or proportions. Parametric or non-parametric tests were chosen depending on the distribution of variables. The Wilcoxon signed rank test was used to compare lipid levels within patients. The Mann–Whitney test was used to compare lipid levels between patients and healthy controls. Two-sided tests were used, and the significance level was set to p < 0.05. Spearman’s correlation coefficients (rs) were used to evaluate the relationship between lipid data and clinical symptoms (PANSS and GAF). Results Clinical characteristics The patients had significantly lower GAF score at T1 than T2; GAF-S (p < 0.001) and GAF-F (p < 0.001), while the PANSS scores were not significantly different between T1 and T2. See Table 1 for details. Use of medication No patients or healthy controls used medication for dyslipidemia. At T1 (admitted to an emergency psychiatric ward) all 55 of the patients used antipsychotic medication, but only 26 (47.3 %) used antipsychotic medication before admission. At T2, 44 patients (80 %) used antipsychotic medication. Repeated lipid measures Serum lipids The levels of triglyceride were significantly higher in patients with schizophrenia than in healthy controls both at T1 (p < 0.001), and at T2 (P < 0.001) (Table 2). The difference in cholesterol between patients (T1 and T2) and healthy controls (T2) was not significant. Fasting serum lipid data were not available in the healthy control group at T1. The levels of serum lipids (triglyceride and cholesterol) remained stable in the patient group over time, with no significant difference between T1 and T2. In the patient group, serum lipid levels remained stable over the 5 year follow-up period, illustrated with significantly correlated levels at T1 and T2 for both cholesterol (rs = 0.63, p < 0.001) and triglyceride (rs = 0.54, p < 0.001).Table 2 Serum and membrane lipids, patients and healthy controls, case–control and longitudinal data Patients T1 Patients T2 Healthy controls T1 Healthy controls T2 n = 55 n = 55 n = 20 n = 51 PUFA 434 (171,507) 471 (440,513)*** 478 (446,495)*** 470.0 (437,508) LCPUFA 283 (89,337) 308 (291,334)*** 307 (293,330)*** 306 (287,336) S-cholesterol 5.00 ± 1.12 5.36 ± 1.20 ¤ 5.05 ± 0.84 S-triglyceride 1.50 (0.80,2.48)### 1.33 (0.95,2.66)### ¤ 0.85 (0.59,1.12) *** P <0.001 vs. patients T1, ### P < 0.001 vs. healthy controls T2 ¤not measured Serum cholesterol: normally distributed, mean ± standard deviation Serum triglyceride, PUFA, LCPUFA: non-normally distributed, median (25,75 percentiles) Patients T1 vs patients T2: Wilcoxon signed rank test Patients T1 and T2 vs healthy controls T1: Mann–Whitney test Healthy control T1 vs healthy controls T2: Wilcoxon signed rank test PUFA = omega-3 + omega-6 polyunsaturated fatty acids in red blood cells LCPUFA = omega-3 + omega-6 PUFA with 20 or more carbon atoms in red blood cells Membrane lipids Membrane lipids were significantly lower in the patient group compared to healthy controls at T1, both for PUFA (p < 0.001) and LCPUFA (P < 0.001) (Table 2). At T2, there was no difference in the levels of membrane lipids between patients and healthy controls. The patients with schizophrenia had significantly lower levels of membrane lipids at T1 than at T2, both for PUFA (p < 0.001) and LCPUFA (p < 0.001) (Table 2). As reported earlier, the distribution of PUFA was bimodal at T1, defining a low and a high PUFA group, and normal at T2 [9]. The mean levels of membrane lipids increased only from T1 to T2 in patients belonging to the low PUFA group at T1. PUFA levels at T1 and T2 were not significantly associated. For healthy controls, there was no significant difference between membrane lipid levels at T1 and T2. Relationship between lipid levels and clinical characteristics Serum lipids In general, the associations between serum lipids and symptoms and functioning were stronger at T2 than T1 (Table 3). There were no significant relationship between the lipid levels and clinical variables at T1. As reported earlier, there was a significant correlation between cholesterol and GAF-S at T2 (p = 0.02), and significant correlations between triglyceride and GAF-S (p = 0.001), GAF-F (p = 0.01) and PANSS positive (p = 0.04). The associations at T2 indicate that higher serum lipids were related to poorer functioning and more severe symptoms.Table 3 Association (rs) between lipid levels in serum and cell membranes and clinical characteristics in patients S-triglyceride S-cholesterol RBC PUFA RBC LCPUFA T1 T2 T1 T2 T1 T2 T1 T2 PANSS total 0.06 0.26 0.01 0.15 0.08 0.14 0.09 0.31* PANSS negative 0.18 0.26 0.19 0.11 0.21 0.32* 0.20 0.52*** PANSS positive −0.04 0.28* −0.03 0.22 −0.02 0.01 −0.07 0.11 GAF-S 0.13 −0.48*** 0.26 −0.30* −0.24 −0.22 −0.26 −0.32* GAF-F 0.07 −0.32* 0.24 −0.23 −0.19 −0.21 −0.23 −0.29* *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 Spearman’s correlation coefficient is reported PANSS positive and negative syndrome scale GAF global assessment of functioning, S symptoms, F functioning) PANSS Positive positive component; items P1 + P3 + P5 + P6 + G9 PANSS Negative negative component; items N1 + N2 + N3 + N4 + N6 + G7 + G8 + G16 PUFA = omega-3 + omega- 6 polyunsaturated fatty acids in red blood cells (RBC) LCPUFA = omega-3 + omega-6 PUFA with 20 or more carbon atoms in red blood cells (RBC) We also investigated associations between serum lipid levels at T1, and symptom levels at T2. The results are reported in Table 4. Triglyceride at T1 was significantly correlated to PANSS total (p = 0.04), GAF-S (p = 0.003) and GAF-F (p = 0.006) at T2. For cholesterol at T1 there was significant correlation to GAF-S at T2 (p = 0.05). All these associations indicate that higher serum lipids at T1 were related to poorer functioning and more severe symptoms at T2.Table 4 Association (rs) between serum lipid levels in serum at T1 and clinical characteristics at T2 S-triglyceride T1 S-cholesterol T1 PANSS positive T2 0.21 0.06 PANSS negative T2 0.27 0.15 PANSS total T2 0.28* 0.11 GAF-S T2 −0.41** −0.28* GAF-F T2 −0.39** −0.26 *0.01 ≤ P < 0.05, **P < 0.01 Spearman’s correlation coefficient is reported There was no significant association between levels of PUFA (polyunsaturated fatty acids) / LCPUFA (long chain polyunsaturated fatty acids) at T1 and symptoms at T2 PANSS positive and negative syndrome scale GAF global assessment of functioning, S symptoms, F functioning) PANSS positive items P1 + P3 + P5 + P6 + G9 PANSS negative items N1 + N2 + N3 + N4 + N6 + G7 + G8 + G16 Membrane lipids As for serum lipids, the associations between membrane lipids and symptoms and functioning were stronger at T2 than T1 (Table 3). There were no significant relationships between these lipids and clinical variables at T1. At T2, there were significant correlations between LCPUFA and PANSS negative (p = 0.001), PANSS total (p = 0.02), GAF-S (p = 0.02) and GAF-F (p = 0.04). PUFA at T2 was correlated to PANSS negative (p = 0.02). All associations at T2 indicate that higher concentrations of membrane lipids are linked to poorer functioning and more severe symptoms. Analyses of relationship between membrane lipid levels at T1 and symptom levels at T2 indicate no significant associations. Discussion The main findings of the present study were repeated higher levels of serum triglyceride in schizophrenia patients than in healthy controls during a 5 year period, while membrane lipid levels (PUFA) were lower in the acute stage of the disease (T1). There were no significant associations between lipid levels and symptoms in the acute stage (T1), while at the chronic stage (T2) both serum and membrane lipid levels were associated with symptoms and functioning. Higher serum lipid levels in the acute stage (T1) were also associated with more severe symptoms in the chronic stage (T2). These findings suggest serum lipid abnormalities as a putative core disease mechanism related to disease traits, while membrane lipids seem to fluctuate in different disease phases. This may be related to changes in neuroinflammatory and oxidative processes which are reported to contribute to disease progression and underlie symptom severity [34, 35]. We found higher levels of serum triglyceride in the patient group both at an acute stage (T1) and at follow-up 5 years later (T2). In longitudinal studies, dyslipidemia in patients with schizophrenia has primarily been studied as a side effects of antipsychotic medication [36]. Some studies have shown that dyslipidemia and other metabolic risk factors may be present in early stages of the disease, before treatment is initiated [37, 38]. Antipsychotics have been shown to up-regulate the expression of cholesterol transport proteins [39]. Further, the degree of dyslipidemia may be predictive of the effect of treatment [40]. In the present study, a large proportion of patients did not receive antipsychotic treatment when lipids were measured (47 % at T1, 20 % at T2). Thus, the sustained higher levels of serum triglyceride, suggest that dyslipidemia may be associated with the disease itself, and not only a result of medication. Further, smoking, gender, antipsychotic medication, and dietary factors does not explain the levels of membrane lipids [9, 26]. The notion that triglyceride levels may be a disease trait is in accordance with recent findings of polygenic overlap between blood lipid levels and schizophrenia, suggesting similar molecular genetic factors [41]. Higher levels of serum lipids were associated with more severe psychiatric symptoms in the chronic phase of the disease. We are not aware of similar findings. Other studies have shown that an increase in serum lipids is related to a reduction of symptoms among patients during treatment with antipsychotic medication [28, 40]. Our findings may be due to more severely ill patients having a more unhealthy lifestyle or being treated with higher doses of antipsychotics, which both can entail higher serum lipid levels. Earlier we have shown that serum lipid levels were not significantly higher among the patients using antipsychotic medication at T2 [26]. The present association between serum lipids and symptoms and functioning became stronger as the disease progressed to a more stable phase. This may reflect increased oxidative stress during the acute psychotic state, disrupting normal relationship between serum lipids and psychiatric symptoms [42, 43]. Among patients, triglyceride levels, and to a lesser degree cholesterol levels, at T1 were associated with symptom levels at T2. Serum lipid levels seem to be related not only to present symptoms, weight and medication, but also to the disease itself. These findings may indicate that disturbed serum lipid levels is a disease trait. This also raises the possibility that serum lipids to some degree may be used as a biomarker in schizophrenia. To the best of our knowledge, the current study has the longest follow-up period of membrane lipids (PUFA) in schizophrenia yet reported. While long-term PUFA levels in the healthy control group were stable, the levels changed in the patient group. However, PUFA levels increased only in the patients with low PUFA levels at T1, while patients with higher PUFA had stable values. The bimodal distribution of PUFA and LCPUFA reported earlier at T1 [9] was not found at T2. Earlier reports on the differences in PUFA levels have shown discrepancies. Several studies have shown lower levels among patients [9, 11] while others reported no such difference [37]. Meta-analyses have shown lower levels of LCPUFA in schizophrenia patients than in healthy controls [10]. Lower levels of membrane lipids have been shown both among drug-naïve patients and in patients treated with antipsychotic drugs [38]. Discrepancies between studies can reflect that PUFA levels were measured at different stages during the course of the disease. During episodes with higher symptom intensity, levels of membrane lipids may be influenced by neuroinflammation, oxidative stress and lipid peroxidation [44, 45]. Treatment with atypical antipsychotic drugs may have a normalizing effect on the phospholipid composition of cell membranes, especially for drug-naïve patients [38]. In the present study, the PUFA levels in the chronic phase of schizophrenia were not lower than in healthy controls, possibly connected to long-term effects of antipsychotic drugs on PUFA [4] or to the remission from an acute psychotic episode [9]. Levels of membrane lipids (PUFA) were associated to symptoms at the follow-up stage (T2) but not at the acute stage (T1). At follow-up, higher levels of PUFA, especially LCPUFA, were associated with higher symptom intensity and poorer functioning [26]. Our findings at T1 may reflect lipid peroxidation, to which LCPUFA, such as DHA (docosahexaenoic acid) and EPA (eicosapentaenoic acid), are especially prone, and may support a role of oxidative stress and free radicals in schizophrenia [46]. Studies of fibroblasts from patients with schizophrenia have shown decreased lysolipid levels and disrupted extracellular matrix when exposed to oxidative stress [47]. Antioxidant defenses are regulators of immunological pathways [16], Schizophrenia is among several neuropsychiatric disorders for which neuroinflammatory processes have been suggested to play a role [44, 48]. Markers of increased inflammation have been found in post-mortem studies [49], in neuroimaging studies [50], in cerebrospinal fluid [51] and peripherally in blood [52] from schizophrenia patients, and disease severity has also been associated with inflammatory markers [53]. The relationship between lipid metabolism and inflammation is established through a series of experimental [54] and clinical studies [55], including shared genetic risk [56]. There is a link between inflammation and both serum lipids [57] and membrane lipids [58]. Inflammation may affect lipid levels through effects on arachidonic acid (ARA) and other phospholipids [59, 60]. PUFA are not only important components of neuronal cell membranes, but also play an important role in regulation of inflammation through the formation of eicosanoids [61]. Inflammation and oxidative stress may play a role in disease progression through lipid peroxidation and cholesterol oxidation, leading to neuronal cell death [62, 63]. These mechanisms may change during the course of the disease. This may explain the conflicting findings regarding membrane lipids and symptoms in different stages of schizophrenia. The unstable character of PUFA may explain why membrane lipids at T1 did not predict symptoms at T2, in contrast to serum lipids. Taken together, this indicates that, in contrast to serum lipids, the levels of membrane lipids are associated with present psychotic symptoms, and not future or past symptom severity. The current study has some limitations. T1 differ from T2, both in terms of duration of illness and acute versus chronic stage. Thus, the effect of duration of illness and the stage may both explain why levels and correlations differ between T1 and T2. The relative importance of each of the characteristics and the influence of confounding factors may not be established by the present design. The weight of the subjects was not obtained at T2, and thus obesity as a factor connected to dyslipidemia cannot be adjusted for. However, it is unlikely that weight affects the associations with symptom levels. Our screening protocol ensured a very small likelihood that any of the participants had familial dyslipidemia. However, diet, dietary supplements other than fatty acids, and other lifestyle factors that may influence the lipid levels, and family history of dyslipidemia were not controlled for. Further, at T1 there were 16 different raters, as opposed to one single rater at T2. More raters will introduce statistical noise, weakening relationships between lipids and symptoms, which could explain lack of association at T1. The control group was not matched for smoking habits and other lifestyle factors, and not for socioeconomic factors such as education level. Though these parameters may influence the lipid levels, matching them against a group of schizophrenia patients would give a sample not representative of the general population. The relative importance of disease and medication cannot be clarified with the present design. The design of the study is not fit to make conclusions about causality. Since this is a naturalistic study, confounding factors not accounted for in the present analysis can affect the lipid levels and clinical symptoms, and the findings need to be replicated in independent studies. Conclusion To conclude, serum lipid levels were elevated among patients both in the acute and chronic stage, while membrane lipids were low in the acute phase of schizophrenia. Serum lipid levels, both in the acute and the chronic stages of schizophrenia, correlated with chronic symptom levels, while membrane lipids showed a more mixed relationship to symptom levels. These findings suggest that higher serum lipids may be a disease trait of schizophrenia, and a possible biomarker, and provide new insight into the role of lipid profiles in schizophrenia pathophysiology. Abbreviations ARAArachidonic acid DHADocosahexaenoic acid EPAEicosapentaenoic acid GAFGlobal assessment of functioning LCPUFALong chain polyunsaturated fatty acids, omega-3 + omega-6, with 20 or more carbon atoms, in red blood cells MINIMini international neuropsychiatric interview PANSSPositive and negative syndrome scale PUFAPolyunsaturated fatty acids, omega-3 +, omega-6 polyunsaturated fatty acids in red blood cells RBCRed blood cells SCIDStructural clinical interview for DSM-IV Acknowledgements We thank the patients who took part in the study and Center for Psychopharmacology, Diakonhjemmet Hospital, who contributed to the data collection. Funding This work was principally supported by the Eastern Norway Regional Health Authority (project # 2005132) and Diakonhjemmet Hospital. We received minor grants from Josef and Haldis Andresen’s legacy, and Emil Stray’s legacy, Norway. OAA is funded by Research Council of Norway (223273) and KG Jebsen Foundation. Availability of data and materials Due to issues regarding confidentiality and ethics, data cannot be shared. Authors’ contributions DKS and HB participated in the design and coordination of the study, in the acquisition and analysis of the data and drafted the manuscript. HR and OAA participated in analysis of the data and drafted the manuscript. All authors have read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate The study was approved by the Regional Committee for Medical Research Ethics. All participants gave their written informed consent. ==== Refs References 1. Paton C Esop R Young C Taylor D Obesity, dyslipidaemias and smoking in an inpatient population treated with antipsychotic drugs Acta Psychiatr Scand 2004 110 4 299 305 10.1111/j.1600-0447.2004.00372.x 15352932 2. Saari K Jokelainen J Veijola J Koponen H Jones PB Savolainen M Serum lipids in schizophrenia and other functional psychoses: a general population northern Finland 1966 birth cohort survey Acta Psychiatr Scand 2004 110 4 279 285 10.1111/j.1600-0447.2004.00358.x 15352929 3. Newcomer JW Second-generation (atypical) antipsychotics and metabolic effects: a comprehensive literature review CNS Drugs 2005 19 Suppl 1 1 93 15998156 4. Kaddurah-Daouk R McEvoy J Baillie R Zhu H Yao K Nimgaonkar VL Impaired plasmalogens in patients with schizophrenia Psychiatry Res 2012 198 3 347 352 10.1016/j.psychres.2012.02.019 22513041 5. Schwarz E Prabakaran S Whitfield P Major H Leweke FM Koethe D High throughput lipidomic profiling of schizophrenia and bipolar disorder brain tissue reveals alterations of free fatty acids, phosphatidylcholines, and ceramides J Proteome Res 2008 7 10 4266 4277 10.1021/pr800188y 18778095 6. Thakore JH Metabolic disturbance in first-episode schizophrenia Br J Psychiatry Suppl 2004 47 S76 S79 10.1192/bjp.184.47.s76 15056598 7. Verma SK Subramaniam M Liew A Poon LY Metabolic risk factors in drug-naive patients with first-episode psychosis J Clin Psychiatry 2009 70 7 997 1000 10.4088/JCP.08m04508 19497246 8. Mahadik SP Evans DR Is schizophrenia a metabolic brain disorder? Membrane phospholipid dysregulation and its therapeutic implications Psychiatr Clin North Am 2003 26 1 85 102 10.1016/S0193-953X(02)00033-3 12683261 9. Bentsen H Solberg DK Refsum H Gran JM Bohmer T Torjesen PA Bimodal distribution of polyunsaturated fatty acids in schizophrenia suggests two endophenotypes of the disorder Biol Psychiatry 2011 70 1 97 105 10.1016/j.biopsych.2011.02.011 21546001 10. Hoen WP Lijmer JG Duran M Wanders RJ Van Beveren NJ De HL Red blood cell polyunsaturated fatty acids measured in red blood cells and schizophrenia: A meta-analysis Psychiatry Res 2013 207 1–2 1 12 10.1016/j.psychres.2012.09.041 23068078 11. Khan MM Evans DR Gunna V Scheffer RE Parikh VV Mahadik SP Reduced erythrocyte membrane essential fatty acids and increased lipid peroxides in schizophrenia at the never-medicated first-episode of psychosis and after years of treatment with antipsychotics Schizophr Res 2002 58 1 1 10 10.1016/S0920-9964(01)00334-6 12363384 12. Montesinos-Rueda L Canete-Crespillo J Palma-Sevillano C Gine-Serven E Erythrocyte membrane polyunsaturated fatty acid (pufa) levels in a sample of patients with schizophrenia and relation with clinical and progression variables Actas Esp Psiquiatr 2015 43 5 170 176 26320895 13. van der Kemp WJ Klomp DW Kahn RS Luijten PR Hulshoff Pol HE A meta-analysis of the polyunsaturated fatty acid composition of erythrocyte membranes in schizophrenia Schizophr Res 2012 141 2–3 153 161 10.1016/j.schres.2012.08.014 22981812 14. Woods AG Sokolowska I Taurines R Gerlach M Dudley E Thome J Potential biomarkers in psychiatry: focus on the cholesterol system J Cell Mol Med 2012 16 6 1184 1195 10.1111/j.1582-4934.2012.01543.x 22304330 15. Weber-Fahr W Englisch S Esser A Tunc-Skarka N Meyer-Lindenberg A Ende G Altered phospholipid metabolism in schizophrenia: a phosphorus 31 nuclear magnetic resonance spectroscopy study Psychiatry Res 2013 214 3 365 373 10.1016/j.pscychresns.2013.06.011 24045051 16. Anderson G Maes M Berk M Schizophrenia is primed for an increased expression of depression through activation of immuno-inflammatory, oxidative and nitrosative stress, and tryptophan catabolite pathways Prog Neuropsychopharmacol Biol Psychiatry 2013 42 101 114 10.1016/j.pnpbp.2012.07.016 22930036 17. Grima G Benz B Parpura V Cuenod M Do KQ Dopamine-induced oxidative stress in neurons with glutathione deficit: implication for schizophrenia Schizophr Res 2003 62 3 213 224 10.1016/S0920-9964(02)00405-X 12837517 18. Horrobin DF Glen AI Vaddadi K The membrane hypothesis of schizophrenia Schizophr Res 1994 13 3 195 207 10.1016/0920-9964(94)90043-4 7841132 19. Nuss P Tessier C Ferreri F De HM Peuskens J Trugnan G Abnormal transbilayer distribution of phospholipids in red blood cell membranes in schizophrenia Psychiatry Res 2009 169 2 91 96 10.1016/j.psychres.2009.01.009 19646766 20. Dietrich-Muszalska A Kontek B Lipid peroxidation in patients with schizophrenia Psychiatry Clin Neurosci 2010 64 5 469 475 10.1111/j.1440-1819.2010.02132.x 20923426 21. Wu JQ Kosten TR Zhang XY Free radicals, antioxidant defense systems, and schizophrenia Prog Neuropsychopharmacol Biol Psychiatry 2013 46 200 206 10.1016/j.pnpbp.2013.02.015 23470289 22. Le LS Simard G Martinez MC Andriantsitohaina R Oxidative stress and metabolic pathologies: from an adipocentric point of view Oxid Med Cell Longev 2014 2014 908539 25143800 23. Podrez EA Anti-oxidant properties of high-density lipoprotein and atherosclerosis Clin Exp Pharmacol Physiol 2010 37 7 719 725 10.1111/j.1440-1681.2010.05380.x 20374263 24. Roberts CK Sindhu KK Oxidative stress and metabolic syndrome Life Sci 2009 84 21–22 705 712 10.1016/j.lfs.2009.02.026 19281826 25. Mahadik SP Mukherjee S Correnti EE Kelkar HS Wakade CG Costa RM Plasma membrane phospholipid and cholesterol distribution of skin fibroblasts from drug-naive patients at the onset of psychosis Schizophr Res 1994 13 3 239 247 10.1016/0920-9964(94)90048-5 7841137 26. Solberg DK Bentsen H Refsum H Andreassen OA Association between serum lipids and membrane fatty acids and clinical characteristics in patients with schizophrenia Acta Psychiatr Scand 2015 132 4 293 300 10.1111/acps.12388 25597473 27. Du Bois TM Deng C Huang XF Membrane phospholipid composition, alterations in neurotransmitter systems and schizophrenia Prog Neuropsychopharmacol Biol Psychiatry 2005 29 6 878 888 10.1016/j.pnpbp.2005.04.034 16005134 28. Procyshyn RM Wasan KM Thornton AE Barr AM Chen EY Pomarol-Clotet E Changes in serum lipids, independent of weight, are associated with changes in symptoms during long-term clozapine treatment J Psychiatry Neurosci 2007 32 5 331 338 17823649 29. Sumiyoshi T Higuchi Y Matsui M Itoh H Uehara T Itoh T Membrane fatty acid levels as a predictor of treatment response in chronic schizophrenia Psychiatry Res 2011 186 1 23 27 10.1016/j.psychres.2010.07.049 20800904 30. Bentsen H Osnes K Refsum H Solberg DK Bohmer T A randomized placebo-controlled trial of an omega-3 fatty acid and vitamins E + C in schizophrenia Transl Psychiatry 2013 3 10.1038/tp.2013.110 24346133 31. Van der Gaag M Hoffman T Remijsen M Hijman R De HL Van MB The five-factor model of the Positive and Negative Syndrome Scale II: a ten-fold cross-validation of a revised model Schizophr Res 2006 85 1–3 280 287 10.1016/j.schres.2006.03.021 16730429 32. Pedersen G Karterud S The symptom and function dimensions of the Global Assessment of Functioning (GAF) scale Compr Psychiatry 2012 53 3 292 298 10.1016/j.comppsych.2011.04.007 21632038 33. Bentsen H Solberg DK Refsum H Bohmer T Clinical and biochemical validation of two endophenotypes of schizophrenia defined by levels of polyunsaturated fatty acids in red blood cells Prostaglandins Leukot Essent Fatty Acids 2012 87 1 35 41 10.1016/j.plefa.2012.05.005 22705264 34. Fernandes BS Steiner J Bernstein HG Dodd S Pasco JA Dean OM C-reactive protein is increased in schizophrenia but is not altered by antipsychotics: meta-analysis and implications Mol Psychiatry 2016 21 4 554 564 10.1038/mp.2015.87 26169974 35. Goff DC, Romero K, Paul J, Mercedes Perez-Rodriguez M, Crandall D, Potkin SG. Biomarkers for drug development in early psychosis: Current issues and promising directions. Eur Neuropsychopharmacol. 2016;19. 36. Mukundan A Faulkner G Cohn T Remington G Antipsychotic switching for people with schizophrenia who have neuroleptic-induced weight or metabolic problems Cochrane Database Syst Rev 2010 12 21154372 37. Fenton WS Hibbeln J Knable M Essential fatty acids, lipid membrane abnormalities, and the diagnosis and treatment of schizophrenia Biol Psychiatry 2000 47 1 8 21 10.1016/S0006-3223(99)00092-X 10650444 38. McEvoy J Baillie RA Zhu H Buckley P Keshavan MS Nasrallah HA Lipidomics reveals early metabolic changes in subjects with schizophrenia: effects of atypical antipsychotics PLoS One 2013 8 7 10.1371/journal.pone.0068717 23894336 39. Assies J Mocking RJ Lok A Ruhe HG Pouwer F Schene AH Effects of oxidative stress on fatty acid-and one-carbon-metabolism in psychiatric and cardiovascular disease comorbidity Acta Psychiatr Scand 2014 130 3 163 180 10.1111/acps.12265 24649967 40. Krakowski M Czobor P Cholesterol and cognition in schizophrenia: a double-blind study of patients randomized to clozapine, olanzapine and haloperidol Schizophr Res 2011 130 1–3 27 33 10.1016/j.schres.2011.04.005 21549568 41. Andreassen OA Djurovic S Thompson WK Schork AJ Kendler KS O’Donovan MC Improved detection of common variants associated with schizophrenia by leveraging pleiotropy with cardiovascular-disease risk factors Am J Hum Genet 2013 92 2 197 209 10.1016/j.ajhg.2013.01.001 23375658 42. Salim S Oxidative stress and psychological disorders Curr Neuropharmacol 2014 12 2 140 147 10.2174/1570159X11666131120230309 24669208 43. Vidovic B Stefanovic A Milovanovic S Ethordevic B Kotur-Stevuljevic J Ivanisevic J Associations of oxidative stress status parameters with traditional cardiovascular disease risk factors in patients with schizophrenia Scand J Clin Lab Invest 2014 74 3 184 191 10.3109/00365513.2013.873947 24446766 44. Muller N Weidinger E Leitner B Schwarz MJ The role of inflammation in schizophrenia Front Neurosci 2015 9 372 10.3389/fnins.2015.00372 26539073 45. Bitanihirwe BK Woo TU Oxidative stress in schizophrenia: an integrated approach Neurosci Biobehav Rev 2011 35 3 878 893 10.1016/j.neubiorev.2010.10.008 20974172 46. Tsaluchidu S Cocchi M Tonello L Puri BK Fatty acids and oxidative stress in psychiatric disorders BMC Psychiatry 2008 8 1 S5 10.1186/1471-244X-8-S1-S5 18433515 47. Fournier M Ferrari C Baumann PS Polari A Monin A Bellier-Teichmann T Impaired metabolic reactivity to oxidative stress in early psychosis patients Schizophr Bull 2014 40 5 973 983 10.1093/schbul/sbu053 24687046 48. Khandaker GM Cousins L Deakin J Lennox BR Yolken R Jones PB Inflammation and immunity in schizophrenia: implications for pathophysiology and treatment Lancet Psychiatry 2015 2 3 258 270 10.1016/S2215-0366(14)00122-9 26359903 49. Trepanier MO, Hopperton KE, Mizrahi R, Mechawar N, Bazinet RP. Postmortem evidence of cerebral inflammation in schizophrenia: a systematic review. Mol Psychiatry. 2016;7. 50. Najjar S Pearlman DM Neuroinflammation and white matter pathology in schizophrenia: systematic review Schizophr Res 2015 161 1 102 112 10.1016/j.schres.2014.04.041 24948485 51. Coughlin JM Wang Y Ambinder EB Ward RE Minn I Vranesic M In vivo markers of inflammatory response in recent-onset schizophrenia: a combined study using [(11) C] DPA-713 PET and analysis of CSF and plasma Transl Psychiatry 2016 6 10.1038/tp.2016.40 27070405 52. Morch RH Dieset I Faerden A Hope S Aas M Nerhus M Inflammatory evidence for the psychosis continuum model Psychoneuroendocrinology 2016 67 189 197 10.1016/j.psyneuen.2016.02.011 26923849 53. Hope S Ueland T Steen NE Dieset I Lorentzen S Berg AO Interleukin 1 receptor antagonist and soluble tumor necrosis factor receptor 1 are associated with general severity and psychotic symptoms in schizophrenia and bipolar disorder Schizophr Res 2013 145 1–3 36 42 10.1016/j.schres.2012.12.023 23403415 54. Siasos G Tousoulis D Oikonomou E Zaromitidou M Stefanadis C Papavassiliou AG Inflammatory markers in hyperlipidemia: from experimental models to clinical practice Curr Pharm Des 2011 17 37 4132 4146 10.2174/138161211798764780 22204374 55. Fritsche KL The science of fatty acids and inflammation Adv Nutr 2015 6 3 293S 301S 10.3945/an.114.006940 25979502 56. Andreassen OA Desikan RS Wang Y Thompson WK Schork AJ Zuber V Abundant genetic overlap between blood lipids and immune-mediated diseases indicates shared molecular genetic mechanisms PLoS One 2015 10 4 10.1371/journal.pone.0123057 25853426 57. Dieset I Hope S Ueland T Bjella T Agartz I Melle I Cardiovascular risk factors during second generation antipsychotic treatment are associated with increased C-reactive protein Schizophr Res 2012 140 1–3 169 174 10.1016/j.schres.2012.06.040 22817875 58. Robichaud PP Surette ME Polyunsaturated fatty acid-phospholipid remodeling and inflammation Curr Opin Endocrinol Diabetes Obes 2015 22 2 112 118 10.1097/MED.0000000000000138 25692925 59. Rao JS Kim HW Harry GJ Rapoport SI Reese EA Increased neuroinflammatory and arachidonic acid cascade markers, and reduced synaptic proteins, in the postmortem frontal cortex from schizophrenia patients Schizophr Res 2013 147 1 24 31 10.1016/j.schres.2013.02.017 23566496 60. Yagami T Koma H Yamamoto Y Pathophysiological Roles of Cyclooxygenases and Prostaglandins in the Central Nervous System Mol Neurobiol 2015 6 28 26437 26456 61. Janssen CI Kiliaan AJ Long-chain polyunsaturated fatty acids (LCPUFA) from genesis to senescence: the influence of LCPUFA on neural development, aging, and neurodegeneration Prog Lipid Res 2014 53 1 17 10.1016/j.plipres.2013.10.002 24334113 62. Fernandez-Perez EJ Peters C Aguayo LG Membrane Damage Induced by Amyloid Beta and a Potential Link with Neuroinflammation Curr Pharm Des 2016 22 10 1295 1304 10.2174/138161282210160304111702 26972288 63. Gamba P Testa G Gargiulo S Staurenghi E Poli G Leonarduzzi G Oxidized cholesterol as the driving force behind the development of Alzheimer’s disease Front Aging Neurosci 2015 7 119 10.3389/fnagi.2015.00119 26150787

PMC005xxxxxx/PMC5000424.txt

==== Front J Cardiovasc Magn ResonJ Cardiovasc Magn ResonJournal of Cardiovascular Magnetic Resonance1097-66471532-429XBioMed Central London 26910.1186/s12968-016-0269-7ReviewPrinciples of cardiovascular magnetic resonance feature tracking and echocardiographic speckle tracking for informed clinical use Pedrizzetti Gianni 1Claus Piet 2Kilner Philip J. 3http://orcid.org/0000-0002-6044-950XNagel Eike eike.nagel@cardiac-imaging.org 41 Department of Engineering and Architecture, University of Trieste, Trieste, Italy 2 Department of Cardiovascular Diseases, Laboratory for Cardiovascular Imaging and Dynamics, KU Leuven, Leuven, Belgium 3 CMR Unit, Royal Brompton Hospital and Imperial College, London, UK 4 Institute for Experimental and Translational Cardiovascular Imaging, DZHK Centre for Cardiovascular Imaging, Interdisciplinary Cardiovascular Imaging, Internal Medicine III and Institute for Diagnostic and Interventional Radiology, University Hospital Frankfurt, Main, Germany 26 8 2016 26 8 2016 2016 18 1 5125 3 2016 27 7 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Tissue tracking technology of routinely acquired cardiovascular magnetic resonance (CMR) cine acquisitions has increased the apparent ease and availability of non-invasive assessments of myocardial deformation in clinical research and practice. Its widespread availability thanks to the fact that this technology can in principle be applied on images that are part of every CMR or echocardiographic protocol. However, the two modalities are based on very different methods of image acquisition and reconstruction, each with their respective strengths and limitations. The image tracking methods applied are not necessarily directly comparable between the modalities, or with those based on dedicated CMR acquisitions for strain measurement such as tagging or displacement encoding. Here we describe the principles underlying the image tracking methods for CMR and echocardiography, and the translation of the resulting tracking estimates into parameters suited to describe myocardial mechanics. Technical limitations are presented with the objective of suggesting potential solutions that may allow informed and appropriate use in clinical applications. Keywords Cardiac mechanicsFeature trackingStrainMyocardial deformationCardiovascular magnetic resonanceissue-copyright-statement© The Author(s) 2016 ==== Body Background Tissue tracking post processing for the calculation of myocardial deformation parameters can, in principle, be applied to the routinely acquired cross sectional images of cardiovascular magnetic resonance (CMR) or echocardiography studies. This makes the approach very appealing for clinical investigation and research, and increasing numbers of studies based on it are being published. Available software packages can be user friendly, readily delivering traces and measured outputs of deformation parameters. However, the methods used and their limitations may be insufficiently understood. Commercial distributors tend to not fully expose all methods used. This paper, written in collaboration between engineers and clinicians, sets out to explain the main principles and variants of methods applied. Exact implementation, however, may vary between packages, which may not be directly comparable with one another, or with other CMR based approaches to the quantification of myocardial deformation. Technology of tissue tracking The technology of tissue tracking falls in a general category of image post-processing methods known as optical flow [1, 2]. The approach is broadly comparable to particle image velocimetry [3, 4] used in fluid dynamics. As with optical flow methods, the underlying principle is based on the recognition of patterns of features or irregularities in the image to be tracked and following them in the successive images of a sequence. This ‘feature tracking’ approach can be applied to routine cine CMR acquisitions [5] and is attracting the interest of many users in research and clinics. A comparable technique, commonly referred to as speckle tracking echocardiography (STE), has been widely applied in echocardiography where ventricular myocardium typically shows a speckled appearance [6, 7]. Optical flow methods need to be modified and optimized with respect to a particular field of application, with adjustments for image quality, temporal resolution, speed and magnitude of the expected displacements. In general, a tracking method begins by identifying a relatively small window on one image and searching for the most comparable image pattern in a window of the same size in the subsequent frame as shown in Fig. 1. The displacement found between the two patterns is taken as the local displacement of the tissue. A larger interrogation window will be needed if a pattern is displaced beyond the limits of a small search window, so window size may need to increase with the time interval between frames. However, use of a large window may reduce the accuracy of results because pattern similarities are averaged over a large area. While a small interrogation window may be unsuitable for detecting large displacements, it can improve the accuracy of detection of small ones. However, if a window is too small, it can become less feasible to recognize successive patterns [2]. In cardiology, speckles or features need to have the size of a few pixels to be recognized, and the dimensions of the window need to be at least 8 x 8 pixels.Fig. 1 Basic principle of tissue tracking. Tracking the portions of tissue about a series of point (indicated in red on the pictures) is based on defining small square windows centered about such points on a first image (left picture) and searching the as-much-as-possible-similar grayscale pattern on the following image (right picture) in the vicinity of the original window. At the first step, shown here, the search windows have the same position in the pair of images and the feature that was at the center of the window of the first images is sought on the corresponding window on the second image (red dot) to provide an estimate of the local displacement. This procedure is usually repeated moving the second windows at the center of the new position and using a so-called coarse-to-fine approach, reducing progressively the window size. Large windows permit to recognize large gross displacements of the tissue, while the reduction of the window search area about the previous estimated targets allows to improve the accuracy and the locality of the estimation The temporal resolution of the modality used is important. If too low, larger displacements will necessitate larger search areas, and the local patterns could become less comparable, an effect known as image de-correlation [8]. On the other hand, a very high temporal resolution must be accompanied by a high spatial resolution, otherwise frame to frame displacements may become smaller than the dimensions of pixels and harder to detect [3]. A complete tracking method may implement a hierarchical sequence of tracking steps. Initially, it detects the inward or outward motion of the cavity-tissue interface. It does this by taking the pattern of signal in the vicinity of each pre-selected point, typically along the endocardial border, located manually in one frame, and effectively shifting the pattern around the vicinity until the most closely matching pattern is found in the neighboring frames. Preferably, the boundary points are located in a frame near end systole as borders can usually be tracked more reliably when displaced apart. Conversely, if adjacent points and their search areas move together and overlap during contraction, the tracked border may depart from the visible boundary. In long-axis views, the relatively large motion of the atrio-ventricular junctions are commonly detected first by searching for the longitudinal displacements of relatively large image patterns in this region. The entire border is then adjusted in relation to this motion, proportionally from the base to the apex, which is assumed fixed at this stage. In the following stages, the algorithm refines the previously computed wide motion and detects local motion reducing progressively the search zone with smaller windows up to a minimal window size [9, 10]. The tracking technology is also amenable of integration with other techniques, like edge detection or registration methods, which can be used to adjust a tracked geometry toward predefined physiological shapes or according to elastic properties [11–14]. Implementation of such adjustments should be always described, because they aim to artificially improve estimation of tissue motion, for example by driving the solution toward an expected geometry which avoids unrealistic shapes or imposing expected elastic properties, at the same time they force a prescribed behavior may differ from the actual one and alter the strain pattern. Tissue tracking was initially developed for two-dimensional (2D) images, but the technology can in principle be extended to track three-dimensional (3D) volumetric regions without conceptual differences. As a result, some 3D tissue tracking solutions are currently available although experience with them is still limited. When this extension is feasible, local 3D tissue features may be tracked simultaneously in all directions to derive all deformation parameters. This could theoretically reduce artifacts in deformation such as those that may result from through-plane of displacements of 3D structures. CMR feature tracking: strengths and limitations Cine CMR is well suited for feature tracking (CMR-FT) by virtue of its relatively unrestricted access to large fields of view and its relatively high signal to noise and contrast to noise ratios. It provides the most accurate and reproducible assessments of global atrial and ventricular volumes and function available [15, 16]. Today’s standard steady state free precession (SSFP) techniques give good contrast between the myocardium [17] and the blood pool at a spatial resolution of about 1–2 mm in-plane and 6–10 mm through-plane acquired during breath holds of approximately 6–8 s. A limitation is the temporal resolution, which may not be able to resolve short-lived phases of cardiac motion. A standard CMR protocol covers the complete left and right ventricle with a stack of short axis images with additional long-axis views (usually 2-, 3- and 4-chamber views). CMR does not appear to be able to distinguish features within the compact myocardium of the LV, presumably due to the relatively large dimensions of voxels and the relative homogeneity of water content and tissue properties within them. Tissue tracking in CMR is therefore most effective around endocardial borders, most of which are trabeculated. The epicardium can also be distinguished, although its clarity depends on the image properties of overlying structures. It should also be borne in mind that cine acquisitions are periodic. A cine loop representing an effectively averaged cardiac cycle is typically reconstructed from ECG gated data acquired over several cycles in a breath hold and delivering 25 to 50 reconstructed phases per heartbeat. The frame rate depends on heart rate and various acquisition parameters. Since MR acquisitions obtain data over several heart beats minor beat-to-beat differences are smoothed out which, in combination with suboptimal temporal resolution, will obscure rapid isovolumic phases and might lead to underestimation of displacement and strain values. Clinical potential of CMR-FT has been recently described [18, 19]. However, it must be borne in mind that different methods have different strengths and limitations as they rely on different strategies for optimal tracking results: some use pure two-dimensional search windows of different sizes; others also include tracking by local M-mode representation along various directions that can be more efficient and accurate when motion is predominantly along one direction (like the longitudinal displacement of atrio-ventricular junctions). Some solutions focus on the endocardium first, which offers a more distinct interface with the cavity, and then on the epicardium. In any case, through-plane motion is a particular concern in this technique because CMR-FT is unable to track features that move out of plane in subsequent tracking frames. 3D tissue tracking can in principle be applied to CMR. However, the stack of short axis cines typically acquired for ventricular volume calculation is not well suited for this as the effective resolution is too low in the through-plane, long axis direction. Nevertheless, 3D acquisitions with comparable resolution in all three orthogonal directions are technically feasible. Although these have yet to be widely implemented, they can be achieved by using relatively long, navigated acquisitions and fast compressed sensing techniques. Speckle tracking echocardiography: strengths and limitations Speckle tracking echocardiography (STE) was the first example of tissue tracking in cardiac imaging [20–22]. Images with a good spatial resolution (pixel size about 0.3 mm) combined with a frame rate above 60 Hz yield a theoretical accuracy in the estimation of strain within a few percent. In STE, the patterns to be tracked are provided by speckles, which are relatively stable acoustic markers that can be identified in the 2D images and originate from the back-scatter of ultrasound from myocardial tissue. Even though they are more evident at the endocardial border they are also present inside the myocardium which can be tracked directly. Speckles are sites of positive interference with the ultrasound wave. They depend on the orientation of the scattering sites and the incident ultrasound field. The highly organized microstructure of compact myocardium consists of a syncytium of myofibers with a specific orientation, entangled with different types of collagen and organized in populations of locally oriented micro-laminar arrays known as sheets or sheetlets [23, 24]. We are not aware that the origin of speckles as a result of interference of scattering has been conclusively determined, but it is clear that this tissue organization, which includes locally oriented microstructural arrays, could potentially explain the differentiated back scattering of ultrasound by compact myocardium. The importance of the orientation of the scattering structures, such as fibers, sheets or collagen, with respect to the incident ultrasound beam, has been investigated [25]. Scattering structures are more echogenic when orientated perpendicular to the line of insonation than when they lie oblique or parallel to it. During the cardiac cycle fibers, sheets and collagen will deform and the orientation of the scattering with respect to the incident ultrasound beam will change, leading to changes in interference patterns and therefore speckles. Because the deformation is relatively slow and coherent, the speckle patterns change relatively slowly and can be tracked over several cycles. But this is a reason for decorrelation in time and the need of a relatively high frame rate in STE. Furthermore, echocardiography images have a lower signal-to-noise ratio than CMR. Some segments of the myocardium may not be adequately imaged and so require averaging of the measured values to fill these gaps. Indeed, a major limitation for the STE technique is that it is greatly dependent on the image quality, making it difficult in instances of poor echogenic windows, ultrasound dropouts and reverberations. In addition, because the quality is worse in the distal part of the ultrasound sector, more proximally located speckles enable better tracking. This leads to inconsistencies of accuracy and reproducibility, particularly relevant to segmental measurements and mitigated through the calculation of average, global strain [26]. Despite the suggested angle independency of STE, the best tracking quality can be achieved from speckles that move in the direction of the ultrasound beam [27, 28] where resolution is higher. Lateral resolution, perpendicular to the ultrasound beam, is lower and subjected to interpolation procedure across adjacent beams. Accordingly, apical views are more suitable for tracking speckles in the longitudinal direction than the radial direction, while parasternal short axis views give varied accuracy, according to segment, for radially or circumferentially directed strains. Averaging over the circumference may be preferable. Echocardiography has a relatively high temporal resolution; but the selection of frame rates represents a tradeoff in the clinical settings where increasing the sector width for a larger field of view is accompanied by a decrease in frame rate if spatial resolution is maintained [26]. More recently, with advances in matrix-array ultrasound transducers and the development of 3D echocardiographic techniques, 3D speckle tracking (3D-STE) has become feasible and several solutions are currently available. However, 3D echocardiographic images have substantially (at least 3 to 4 times) lower spatial and temporal resolution than their 2D counterparts, and image quality may not be adequate. Currently, 3D-STE assessments remain controversial [27]. Known pitfalls of the tissue tracking technology One basic assumption underlying tissue tracking in 2D images is that apparent in-plane displacements of boundaries or deformations of gray-scale distributions correspond to actual displacements or deformations of tissue structure. This is not necessarily the case. For example, through-plane displacement of a tapering, helically structured or otherwise obliquely angulated forms could be misinterpreted as in-plane deformation or displacement in a 2D image. On the other hand, it is also true that the through-plane motion is small in most regions, and first experiences of 3D tracking do not report systematic differences in this sense [29, 30]. It is particularly in basal short axis or off-axis long axis planes that the user must remain aware of this potential effect and consequent potential misinterpretations. Another assumption is that the myocardium is a coherent deformable tissue. Again this may not be true for all regions or across all spatial scales, for example in the trabeculated myocardial layer. Good spatial resolution is crucial to grasp the deformation of these small structures and problems arise with low-resolution images. In general, the use of relatively large interrogation windows in the tracking procedure helps to overcome this issue although it is not known whether effectively or only apparently. Use of tissue tracking technology implies the underlying assumption that the motion detected on the image sequence represents the motion of the actual tissue and that blood motion is not visible in the image. This is usually true in echocardiography because ultrasound wavelength is too long to detect blood cells, unless a contrast agent (made of hyper-echogenic resonating bubbles) is injected. However, blood motion can be visible on cine CMR with potential to affect tracking near the endocardial border. Blood may moves in a direction opposed to that of tissue (for example from base to apex during diastole when tissue moves toward the base) and may give rise to unrealistic results. CMR-FT therefore needs to be used with caution where adjoining intra-cavity blood flow is clearly visible. The most critical problem associated with any of the tissue tracking techniques appears to be suboptimal repeatability, even of the post processing, let alone between different studies and their operators [31–34]. The frame to frame tracking is based on the search for most probable pattern correspondence between regions in successive image frames. This delivers displacement values identified by being at the maximum of a correlation function. But these functions are generally smooth 2D surfaces (or 3D hypersurfaces in 3D tracking). Locating the maximum point along a smoothly bulging surface is not easy. It is comparable to asking 3 people to locate the top of a smooth hill: they may chose points of similar height but different location. Moreover, the actual position of such maxima can shift significantly when the surface is slightly modified by small variation of the search windows. Therefore, small differences of parameters or user choices during the tracking process can lead to different results. The presence of distinct features such as a sharp tissue-cavity interface or specifically delineated tags in a dedicated CMR acquisition increases the sharpness of the correlation functions and therefore reproducibility. The problem of poor repeatability may be reduced by tracking procedures that incorporate multiple steps with additional calculations aimed at improving the accuracy of pattern location. Nevertheless, the user must still check the border motion visually to avoid evident errors, and possibly repeat the calculation to verify the consistency of the results found. End-user software tools may incorporate smoothing and validation procedures to reduce variability and minimise unrealistic results. On the other hand, these artificial steps could mask clinically significant abnormalities, either of localised regions in the case of spatial smoothing, or of rapid transients in the case of temporal smoothing. Even under ideal conditions, the tissue tracking technology presents a technical limitation related to the pixel size. If a feature is displaced less than one pixel it may not appear to move. However, in practice, sub-pixel estimation can be achieved by testing larger, multi-pixel areas, although this may result in less accuracy and reproducibility. This limitation implies that an increase in the acquisition frame-rate improves the estimation of rapid displacements and large velocities but, at the same time reduces the accuracy in the evaluation of slow motion, if not accompanied by a corresponding increase in spatial resolution. Finally, it must be kept in mind that the tissue tracking technology is based on ‘estimations’ of tissue displacement rather than on exact formulae; therefore the presence of small errors is unavoidable during the evaluation of relative motion. When these inaccuracies are additive over successive frames, errors can accrue in the displacements calculated. This issue can be partially corrected taking into account the beat-to-beat periodic motion of the cardiac tissue and correcting the drift by assuming that the tissue returns to the same position by the following end-diastole. Some solutions enforce the constraint of periodicity during application of the tracking algorithm rather than by correcting afterwards. The estimation errors are usually uncorrelated between the different regions and the different time instants. This implies that global results are more accurate than local ones because they describe an average process where uncorrelated errors partly cancel out. Similarly, instantaneous measures (e.g. velocity or strain-rate) are less accurate than time-integrated parameters like displacement or strain. Calculations of deformation parameters from tissue tracking The attempted tracking of tissue motion from cine image frames is only a starting point for the quantification of cardiac deformation parameters. The tracked points are assumed to represent discrete tissue components that may move relative to given spatial coordinates. Relative to the most widely used ventricular coordinate system, they can move radially, which can usually be taken to be perpendicular to the local wall plane, and longitudinally or circumferentially, which mean along the wall as seen in either a long or short axis image plane, respectively. The tissue velocity vector can be computed as the frame-to-frame displacement divided by the time-interval. Bearing in mind that this is a time-derivative procedure (computed from the difference between displacement values at different instants), it is noisier than the displacement itself, reduces the accuracy, and requires a sufficiently high frame rate. The knowledge of more than one region of tissue motion allows the calculation of tissue strain between them. Strain can be expressed either as a decimal fraction such as 0.15 or as a percentage such as 15 %. Expression as percentage can give rise to confusion and potential misinterpretation when it comes to reporting, as percentages, the relative differences of measurements between groups. Specific percentage differences could then have one of two meanings: absolute differences of strain measurements, or proportional differences between groups, so it is important to describe and interpret such results very carefully. Strain records whether the length L of a piece of tissue gets smaller (shortening or thinning, known as negative strain) or larger (lengthening or thickening, known as positive strain) with respect to an initial end-diastolic length L0. Based on this, strain can be defined as: 1 St=L−L0L0; This is known as Lagrangian strain because it refers to an initial undeformed state [26]. In some applications a different definition, called Eulerian strain, has been used where the same difference in lengths is normalized with the final length L, instead of L0, in the denominator of (1); however use of this formula is not usual in cardiology. Strain can be computed taking a tissue segment of length L along any specified direction. When this segment is taken along the longitudinal direction it gives the longitudinal strain, or circumferential strain when it is taken along the circumference, or it is the radial strain when the length is taken over the thickness. Endocardial longitudinal and circumferential strains are computed when the segment of tissue length L is taken from the endocardial border. Endocardial strain values are those more frequently used in clinical studies because they better represent the functional purpose of myocardial contraction, reducing the endocardial surface of the cavity to eject the stroke volume, and as such correlate best with volumetric measurements like ejection fraction [35]. Likewise epicardial strain, which it typically less than endocardial, can be evaluated along the epicardial border, although this is rarely used. It has been argued that strain is of interest as a measure of muscular contraction [36, 37], and it has been measured as the average value of strain computed over the myocardial thickness. The description of displacements or strain must be expressed with respect to an initial, reference state, which is usually taken the end diastolic instant. In contrast, velocity and strain-rate should reflect an instantaneous activity irrespective of a starting point. From Lagrangian strain values the Lagrangian strain-rate can be computed as its time-derivative 2 SRL=dStdt=1L0dLdt; which means that, as for velocity, strain-rate is a differential quantity and its evaluation depends on sufficient temporal resolution. Its accuracy is lower than that of strain. However, looking at (2), it is apparent that the definition of Lagrangian strain-rate does depend on a reference length L0, a dependency that is disturbing when defining an instantaneous property. For this reason, Lagrangian strain-rate is not commonly used. Instead, a physically consistent definition for strain-rate is known as natural strain-rate 3 SR=1LdLdt; where the rate of shortening of a length L of tissue is measured relative to the actual length of tissue, independently from its previous deformation history. The Lagrangian strain (1) and the natural strain-rate (3) are those more frequently used in cardiology, usually without the Lagrangian/natural suffix. In that case, the relationship between strain and strain-rate becomes 4 SR=1St+1dStdt,St=exp∫t0tSRdt−1. This apparent complexity derives from the need of defining strain in the intuitive manner (1), while a natural definition of strain as the integral of the natural strain-rate, and commonly called natural strain 5 StN=∫t0tSRdt=logLL0, would have ensured an improved mathematical consistency. But the definition (5) is less intuitive and cannot be immediately compared to visual measures, like thickening from M-mode, and is not practical for clinical cardiology. 3D deformation parameters and future representation In principle, all deformation parameters can be computed from a single 3D acquisition by 3D tissue tracking. In 3D, three normal strain values (i.e., longitudinal, circumferential, and radial) can be computed at each point to describe the amount of shortening or stretch in the tissue. At the same time, the sliding between tissue layers at each point is described as shear (i.e., circumferential-longitudinal, longitudinal-radial and circumferential-radial shear values, respectively). An important shear component is the circumferential-longitudinal shear, which describes the torsional deformation of the LV. What is commonly referred as LV torsion is usually evaluated by the difference of rotation measured in the basal and apical short axis level (referred as LV twist), normalized with the distance between the two planes [38, 39]. This definition of LV torsion has the unit of degrees/cm, this is used in cardiology but it is not typical for other evaluations of shear that are generally defined as the ratio between two sides of a sheared parallelogram, giving a dimensionless measure representing the tangent of the shear angle. In general, the strain and shear together describe the complete deformation and the individual values are arranged in a 3x3 table which is defined the strain tensor 6 St=StlongStlong−circStlong−radStlong−circStcircStcirc−radStlong−radStcirc−radStrad; where, for completeness, it must be remarked that the strain tensor (6) is symmetric because a non-symmetric contribution would correspond to local rigid rotations that do not contribute to deformation. Before moving ahead, it is useful to remark that different definitions of strain can be employed in the strain tensor (6). Lagrangian, Eulerian and natural strain were previously introduced in the context of two-dimensional imaging and can be immediately extended to 3D; other definitions are often employed to better relate the finite deformation to the effective tissue stress depending on tissue material properties [40]. This brief discussion, however, is limited to strain description without entering into tissue stresses and elasticity, and applies independently from the specific definitions adopted. When considering 3D deformation in its entirety it is also possible to enforce the constraint of the conservation of mass or volume (neglecting the displacement of the intramyocardial blood volume). Contraction in longitudinal and circumferential directions is accompanied by radial thickening. In quantitative terms this means that the sum of the three strains values on the diagonal of (6) is equal to zero (this is true considering natural strain, otherwise the relationship is slightly more complicated); thus once two strain values are measured, the third can be recovered from first principles. In practice, tissue tracking in 3D is more easily performed at the endocardial level because it is aided by the echogenicity, or signal contrast, of the tissue-cavity interface, while the spatial resolution does not allow adequate resolution of variations across the thickness comprised only of a few pixels. In that case, the circumferential-longitudinal (torsional) shear is the only shear measured and the value of radial strain can theoretically be calculated based on the principle of conservation of mass. It remains unclear how all of these 3D strain components can be interlinked and represented for convenient clinical assessment of myocardial deformation. One approach, initially developed through CMR technologies [41–43], and recently introduced in 3D echocardiography [44–46], entails identification of the principal directions of deformation. Principal directions are those orthogonal directions defined such that there is no shear across them. Therefore, the same 3D deformation, described in (6) with reference to the longitudinal circumferential and radial directions, can be equivalently described by only three principal strains along the principal directions without shear. CMR results showed that the greatest systolic principal strain is positive and directed predominantly in the radial direction, while the other two are predominantly in the wall plane and both negative, so delivering reduction of the LV cavity, with one contractile strain much larger than the other. Tissue tracking reliability and reproducibility Tissue tracking software can give measurements of LV deformations rapidly and easily. However, given the limitations of the technique, users must be aware that it is not a perfect measurement tool. First, 2D tissue tracking analysis should ideally be performed in defined planes through the LV where the tissue is well visualised, through-plane motion is minimal and image quality is high, with sufficient spatial and temporal resolution. Even then, results still have potential for variability. Therefore, multiple tracing, with visual checks for effective tracking, can help to reduce the risk of erroneous results. In particular, in long axis views, the tracking of the annulus must be checked visually as errors here can corrupt other parameters. Secondly, it must be understood that different variables and indicators of mechanical function present different reliability. As a general rule integral variables, like displacement and strain, are more reliable than differential ones like velocity and strain-rate, which require more cautious interpretation. Moreover, when the time interval between frames is large, velocities and strain-rate values cannot reproduce transient phenomena, such the spikes at iso-volumetric phases, that last no more than a few of such intervals. Among displacements, measurements of rotation can be less reliable because of the small displacements along the line of the wall relative to those of the inward, radially directed movement of the endocardial boundary. However, radial strain is usually less accurate than longitudinal and circumferential strain because the distance between endocardium and epicardium is small, there may be systolic elimination of visible blood spaces between trabeculae that exaggerates apparent shift of the endocardial boundary (Fig. 2), and there may be overlap of adjacent interrogation windows during calculation that makes it hard to resolve different displacements between adjacent regions. Table 1 gives an overview of the relative accuracies of different parameters of cardiac deformation.Fig. 2 Possible effects on apparent strain of through-plane tissue displacements and trabecular appearances. Panel a This end-diastolic four chamber cine frame shows a basal septal bulge (*) whose long axis displacement causes it to move apically (arrow) to its end systolic position (b). Here it has moved into the short axis plane marked by the pale line. The septum in that short axis plane will therefore appear to have thickened more than it really did so that excess radial strain could be measured by tissue boundary tracking. Conversely, a more basal short axis slice might underestimate local radial strain due to tapering of the septum near the atrio-ventricular junction. Panel c Short axis images typically show trabeculations inside the LV free wall. Panel d this resin cast of human heart cavities shows the typical right handed helical alignments of free wall trabecular indentations. Systolic long axis displacement of these oblique trabecular structures (arrow) could give a false impression of trabecular displacement in the plane indicated by the white bar. A further issue is that trabeculations tend to thicken and move together in systole. This can exclude intervening blood, particularly in a hypertrophied ventricle with good function (panel e, from the same cine as c). If they merge to appear as part of the LV wall, it could result in over-estimation of radial and circumferential strain, if these were based on attempting to track the apparent endocardial boundary Table 1 Parameters based on instantaneous and/or local results are less reliable than those based on a proper combination of a large number of results More reliable (time-integral parameters) Less reliable (instantaneous parameters) Notes Displacements Velocities Radial motion more accurate than tangential (e.g. rotation) because the tissue-cavity interfaces is a better feature that those found along the tissue. Strain Strain-rate Values along the borders (longitudinal and circumferential strain) are more accurate than radial ones because the latter is a difference between close structures. Properties of time-curve profiles (phase of harmonics, principal components) Instantaneous values (peak, time to peak) Instantaneous measures present unavoidable estimation errors which are smoothed-out when a larger number of measures is combined together. More reliable (spatial-integral values) Less reliable (local values) Global Strain Segmental strain Local measures present unavoidable estimation errors which are smoothed-out when a larger number of measures is combined together. Measures built by numerous spatial values Indicators built by values at single points Thirdly, with variable results, the application of smoothing can result in either regional or instantaneous values being less reliable. Currently, global measures give more reliable results. They include volumes, areas, global strain, average inward motion and possibly mean rotation. Regional differences may be better described in terms of mean dispersion (standard deviation) or other integral measures that are based on appropriate combinations of numerous values. A particular case is the visualisation and quantification of dyssynchronous motion [47]. Complete deformation curves through the course of the cycle can be taken into account when comparing the amplitude or timing of movements rather than confining measurements of point values such as a single peak or time to peak. When visually reviewing time curves, we naturally compare whole curves and can learn to interpret differences. For quantification, statistical comparisons of the curves, for example by manifold learning, enables abnormalities to be detected from the complete trace and could potentially increase robustness of the technique [48]. Dedicated CMR strain acquisitions CMR has a long-standing tradition of dedicated sequences and post-processing techniques for non-invasive measurement of myocardial deformation. CMR-based myocardial line tagging [49] was probably the first technique to assess regional myocardial deformation non-invasively. The original linear tagging approach was then augmented by tagging two orthogonally intersecting sets of lines mark rectangular grids in a 2D image by Spatial Modulation of Magnetization (SPAMM or C-SPAMM) techniques. These have been considered the gold standard for myocardial strain, although their limitations include the possible obscuring by tag lines of some endocardial borders, suboptimal temporal resolution and the additional, dedicated acquisition time. The tag lines represent the intersections with the image plane of sets of tag planes orientated orthogonal to it. For this reason, tagging is not subjected to the strain artifacts illustrated in Fig. 2 that result from through-plane displacements of tapering or obliquely orientated structures. The methods of quantifying deformation from tagged images are essentially similar to those of feature tracking. However, tags represent imposed features that are more regularly and clearly defined. They can therefore be tracked with more ease than natural features, enabling higher reproducibility with potential for regional differentiation of strain values [50]. The challenge of post-processing has been further reduced by the Harmonic Phase (HARP) algorithm, using the first harmonic peaks of the image’s k-space which has been carefully validated [51]. This potentially delivers information relating to all tissue areas, not just tag intersection points which allows some improvement of effective spatial resolution and the technique has been widely applied to assessments of circumferential strains. An alternative way to assess cardiac motion with CMR is ‘Displacement Encoding with Stimulated Echos’ (DENSE) [52]. Tissue displacement is measured at the pixel level in three dimensions (2 in-plane and 1 trough plane). This technique has no visible tags, can be performed with high spatial resolution and provides black blood images with a good endocardial border definition. Post-processing is relatively easy and displacement is calculated on a pixel level. However, clinical research with this technique has been limited. Strain-encoded CMR (SENC) was introduced as an extension to HARP, in order to measure strain directed orthogonal to the image plane, for example longitudinal strain from a short axis acquisition [53], although its use has so far been limited. Speckle tracking and feature tracking experience and prospects More experience is available with STE echocardiography to assess virtually any cardiac chamber. LV longitudinal, circumferential and radial strains and strain-rates in addition to LV rotational mechanics were calculated using both 2D-STE and 3D-STE. Global longitudinal strain (GLS) averaged from the apical views is the most robust and reproducible of the LV deformation parameters [54–57]. In CMR-FT the most consistent parameters was GCS, closely followed by GLS [32, 58]. In contrast, variations in GRS between studies were large. Segmental strain values show significantly greater variability, at levels which may be unacceptable for clinical use [18, 19]. Overall GCS compared best with measurements by CMR tagging with HARP post-processing [5]; GLS was compared to SENC [59] showing small bias and fair agreement for both RV and LV. Several studies recently compared results from STE and CMR-FT with good agreement in values of GLS and GCS [60–63]. Circumferential strain may present improved reproducibility in CMR-FT than in STE presumably for the higher quality of CMR short axis images. The high echogenicity of the fibrous annulus enables more accurate results for longitudinal strain by STE. Agreement is lower for radial strain, which is probably less reliable for both modalities as previously discussed. However, comparability is still inconclusive in pathologic conditions and non-satisfactory for regional deformation [18]. Recent echocardiographic studies demonstrated that GLS is a sensitive marker of LV dysfunction [64], and therefore GLS is now considered a useful diagnostic and prognostic tool [65]. GLS is well estimated by STE because it depends principally from the motion of the mitral annulus, which is highly echogenic; moreover, the use of averaged strain overcame the limitations of non-visible segments in echography. In CMR long-axis views, the annulus shows a complex structure which deforms and rotates during the heartbeat possibly with a small through-plane component. Its tracking can sometime be elusive and care may be needed to verify the motion detected by CMR-FT to ensure that GLS value is accurate. In contrast, literature is less conclusive about circumferential strain. The evaluation of GCS in echocardiography is less widely used as it requires additional short-axis recordings, STE literature reports considerable variability which might be caused by echoes from other structures lying in short-axis views. Further studies are needed to assess its clinical value. CMR short-axis cines, on the other hand, are typically of good quality with few problems related to image artifacts so preliminary studies suggest good feasibility of the tracking approach [19]. Longitudinal strain can also be relevant clinically as it is associated with the apical displacement of the mitral annulus. Circumferential strain tends to be more segmentally variable, reflecting local inward motion and thickening. Thus, CMR-FT could potentially distinguish the relative advantages of global or regional circumferential strain measurements. The clinical relevance of deformation imaging is discussed in [18] in relation to different pathological conditions. An important source of variability in STE was attributed to differences between different vendor machine and software [66]. The existence of software-related variability induced scientific societies (American Society of Echocardiography, ASE, and European Association of Cardiovascular Imaging, EACVI) to call for a standardization of deformation imaging. A first objective has been that of providing a unique common definition of strain parameters agreed by most vendors [26] and explicitly of stating whether results apply to the endocardium or to the entire myocardial muscle. Another important objective has been the comparison of results obtained from the available software on the same groups of subjects in order to verify intra-vendor and inter-vendor reproducibility [34]. This effort is driving the product development of imaging companies to ensure a comparable technological ground where differences are attributable to the accuracy of the proprietary approach and the proposed parameters or visualization options. As an initial result, GLS has been recently included in the guidelines for the evaluation of LV diastolic function by echocardiography [67]. A similar issue is anticipated in CMR-FT where different vendors are offering independently developed solutions whose differences are reflected in the numeric values [19]. These are also likely to require studies of comparability, and hopefully progress towards the optimization and standardization protocols as well as the harmonization between STE and CRM-FT. A potential difficulty is associated with the reluctance of the distributors of commercial post-processing software to explain all details of the methods used, especially in the ongoing process of competitive development. This is understandable, and methods may anyway be complex and in the process of improvement. On the positive side, a commercial operation brings with it the benefits of dedicated, user-friendly product development, distribution and support. However, it may also be worth considering the potential advantages a non-commercial, open-source and research-funded approach to software development. However the field progresses, further refinement of software, more informed use of it, and carefully conducted studies are all likely to be needed for CMR-FT to become reliable and widely accepted for clinical application. Conclusions The automated tracking and analysis of tissue motion in cine images is a relatively new and convenient tool which, if used with awareness of its limitations, can provide the clinician with quantitative information and visually informative temporal traces that together can become a valuable support to the diagnostic process. However, tissue tracking is not an exact measurement method and requires knowledge of its underlying methods for informed interpretation. With current technology, global parameters are those that show sufficient reproducibility to be potentially useful as clinical parameters, with measurements of longitudinal and circumferential deformations more reliable than radial ones. Identification and standardization of the most effective available solutions would be desirable to reduce inter-vendor variability and facilitate comparative research. Acknowledgements GP was supported by the University of Trieste and research was partially funded from the Italian Government under grant PRIN N. 2012HMR7CF_002. PC was supported by KU Leuven under grant nr. PF/10/014. PJK was supported by the NIHR Cardiovascular Biomedical Research Unit at the Royal Brompton and Harefield NHS Foundation Trust and Imperial College London. EN was supported by the German Centre for Cardiovascular Research (DZHK). Authors’ contributions All authors contributed equally to the manuscript ideation, to its writing and revision. All authors read and approved the final manuscript. Competing interests GP is shareholder of AMID srl (Sulmona, Italy), a company which is partner of Tomtec Imaging Systems Gmbh (Unterschleissheim, Germany) and Medis Medical Imaging Systems bv (Leiden, The Netherlands). EN receives support in kind from TomTec Imaging Systems GmbH (Unterschleissheim, Germany) and from Circle Cardiovascular Imaging Inc. (Calgary, Canada). PJK and PC declare no competing interests. ==== Refs References 1. Singh A Optic flow computation: a unified perspective 1991 Los Alamitos IEEE Computer Society Press 2. Barron JL Fleet DJ Beauchemin SS Performance of optical flow techniques Int J Comput Vis 1994 12 43 77 10.1007/BF01420984 3. Adrian RJ Particle-Imaging Techniques for Experimental Fluid Mechanics Annu Rev Fluid Mech 1991 23 261 304 10.1146/annurev.fl.23.010191.001401 4. Willert CE Gharib M Digital particle image velocimetry Exp Fluids 1991 10 181 193 10.1007/BF00190388 5. Hor KN Gottliebson WM Carson C Wash E Cnota J Fleck R Comparison of Magnetic Resonance Feature Tracking for Strain Calculation With Harmonic Phase Imaging Analysis JACC Cardiovasc Imaging 2010 3 144 151 10.1016/j.jcmg.2009.11.006 20159640 6. Bohs LN Trahey GE A novel method for angle independent ultrasonic imaging of blood flow and tissue motion IEEE Trans Biomed Eng 1991 38 280 286 10.1109/10.133210 2066142 7. Kaluzynski K Chen X Emelianov SY Skovoroda AR O’Donnell M Strain rate imaging using two-dimensional speckle tracking IEEE Trans Ultrason Ferroelectr Freq Control 2001 48 1111 1123 10.1109/58.935730 11477770 8. Liang T Yung L Yu W On feature motion decorrelation in ultrasound speckle tracking IEEE Trans Med Imaging 2013 32 435 448 10.1109/TMI.2012.2230016 23204278 9. Pirat B Khoury DS Hartley CJ Tiller L Rao L Schulz DG A Novel Feature-Tracking Echocardiographic Method for the Quantitation of Regional Myocardial Function. Validation in an Animal Model of Ischemia-Reperfusion J Am Coll Cardiol 2008 51 651 659 10.1016/j.jacc.2007.10.029 18261685 10. Hor KN, Baumann R, Pedrizzetti G, Tonti G, Gottliebson WM, Taylor M, Benson W, Mazur W. Magnetic resonance derived myocardial strain assessment using feature Tracking. JoVE. 2011;48. doi:10.3791/2356. 11. Bistoquet A Oshinski J Skrinjar O Left Ventricular Deformation Recovery From Cine MRI Using an Incompressible Model IEEE Trans Med Imaging 2007 26 1136 1153 10.1109/TMI.2007.903693 17896588 12. Bistoquet A Oshinski J Skrimjar O Myocardial deformation recovery from cine MRI using a nearly incompressible biventricular model Med Image Anal 2008 12 69 85 10.1016/j.media.2007.10.009 18234539 13. Yue Y Clark JW Khoury DS Speckle tracking in intracardiac echocardiography for the assessment of myocardial deformation IEEE Trans Biomed Eng 2009 56 416 425 10.1109/TBME.2008.2006036 19272903 14. Heyde B Jasaityte R Barbosa D Robesyn V Bouchez S Wouters P Elastic image registration versus speckle tracking for 2-d myocardial motion estimation: A direct comparison in vivo IEEE Trans Med Imaging 2013 32 449 459 10.1109/TMI.2012.2230114 23204281 15. Pennell DJ Cardiovascular magnetic resonance: Twenty-first century solutions in cardiology Clin Med (Northfield Il) 2003 3 273 278 10.7861/clinmedicine.3-3-273 16. Chuang ML Hibberd MG Salton CJ Beaudin RA Riley MF Parker RA Importance of imaging method over imaging modality in noninvasive determination of left ventricular volumes and ejection fraction: Assessment by two- and three-dimensional echocardiography and magnetic resonance imaging J Am Coll Cardiol 2000 35 477 484 10.1016/S0735-1097(99)00551-3 10676697 17. Thiele H Paetsch I Schnackenburg B Bornstedt A Grebe O Wellnhofer E Improved accuracy of quantitative assessment of left ventricular volume and ejection fraction by geometric models with steady-state free precession J Cardiovasc Magn Reson 2002 4 327 339 10.1081/JCMR-120013298 12234104 18. Claus P Omar AMS Pedrizzetti G Sengupta PP Nagel E Tissue Tracking Technology for Assessing Cardiac MechanicsPrinciples, Normal Values, and Clinical Applications JACC Cardiovasc Imaging 2015 8 1444 1460 10.1016/j.jcmg.2015.11.001 26699113 19. Schuster A Hor KN Kowallick JT Beerbaum P Kutty S Cardiovascular Magnetic Resonance Myocardial Feature Tracking: Concepts and Clinical Applications Circ Cardiovasc Imaging 2016 9 e004077 10.1161/CIRCIMAGING.115.004077 27009468 20. Malpica N Santos A Zuluaga MA Ledesma MJ Pérez E García-Fernández MA Tracking of regions-of-interest in myocardial contrast echocardiography Ultrasound Med Biol 2004 30 303 309 10.1016/j.ultrasmedbio.2003.11.007 15063512 21. Leitman M Lysyansky P Sidenko S Shir V Peleg E Binenbaum M Two-dimensional strain--a novel software for real-time quantitative echocardiographic assessment of myocardial function J Am Soc Echocardiogr 2004 17 1021 1029 10.1016/j.echo.2004.06.019 15452466 22. Vannan MA Pedrizzetti G Li P Gurudevan S Houle H Main J Effect of cardiac resynchronization therapy on longitudinal and circumferential left ventricular mechanics by velocity vector imaging: Description and initial clinical application of a novel method using high-frame rate B-mode echocardiographic images Echocardiography 2005 22 826 830 10.1111/j.1540-8175.2005.00172.x 16343165 23. LeGrice IJ Smaill BH Chai LZ Edgar SG Gavin JB Hunter PJ Laminar structure of the heart: ventricular myocyte arrangement and connective tissue architecture in the dog Am J Physiol 1995 269 H571 H582 7653621 24. Hales PW Schneider JE Burton RAB Wright BJ Bollensdorff C Kohl P Histo-anatomical structure of the living isolated rat heart in two contraction states assessed by diffusion tensor MRI Prog Biophys Mol Biol 2012 110 319 330 10.1016/j.pbiomolbio.2012.07.014 23043978 25. Crosby J Hergum T Remme EW Torp H The effect of including myocardial anisotropy in simulated ultrasound images of the heart IEEE Trans Ultrason Ferroelectr Freq Control 2009 56 326 333 10.1109/TUFFC.2009.1041 19251519 26. Voigt J-U Pedrizzetti G Lysyansky P Marwick TH Houle H Baumann R Definitions for a common standard for 2D speckle tracking echocardiography: consensus document of the EACVI/ASE/Industry Task Force to standardize deformation imaging Eur Heart J Cardiovasc Imaging 2015 16 1 11 10.1093/ehjci/jeu184 25525063 27. Mor-Avi V Lang RM Badano LP Belohlavek M Cardim NM Derumeaux G Current and evolving echocardiographic techniques for the quantitative evaluation of cardiac mechanics: ASE/EAE consensus statement on methodology and indications endorsed by the Japanese Society of Echocardiography J Am Soc Echocardiogr 2011 24 277 313 10.1016/j.echo.2011.01.015 21338865 28. Langeland S Wouters PF Claus P Leather HA Bijnens B Sutherland GR Experimental assessment of a new research tool for the estimation of two-dimensional myocardial strain Ultrasound Med Biol 2006 32 1509 1513 10.1016/j.ultrasmedbio.2006.06.021 17045871 29. Saito K Okura H Watanabe N Hayashida A Obase K Imai K Comprehensive evaluation of left ventricular strain using speckle tracking echocardiography in normal adults: comparison of three-dimensional and two-dimensional approaches J Am Soc Echocardiogr 2009 22 1025 1030 10.1016/j.echo.2009.05.021 19556106 30. Yodwut C Weinert L Klas B Lang RM Mor-Avi V Effects of Frame Rate on Three-Dimensional Speckle-Tracking--Based Measurements of Myocardial Deformation J Am Soc Echocardiogr 2012 25 978 985 10.1016/j.echo.2012.06.001 22766029 31. Cheng S Larson MG McCabe EL Osypiuk E Lehman BT Stanchev P Reproducibility of speckle-tracking-based strain measures of left ventricular function in a community-based study J Am Soc Echocardiogr 2013 26 1258 1266 10.1016/j.echo.2013.07.002 23953701 32. Morton G Schuster A Jogiya R Kutty S Beerbaum P Nagel E Inter-study reproducibility of cardiovascular magnetic resonance myocardial feature tracking J Cardiovasc Magn Reson 2012 14 43 10.1186/1532-429X-14-43 22721175 33. Schuster A Morton G Hussain ST Jogiya R Kutty S Asrress KN The intra-observer reproducibility of cardiovascular magnetic resonance myocardial feature tracking strain assessment is independent of field strength Eur J Radiol 2013 82 296 301 10.1016/j.ejrad.2012.11.012 23246014 34. Farsalinos KE Daraban AM Ünlü S Thomas JD Badano LP Voigt J-U Head-to-Head Comparison of Global Longitudinal Strain Measurements among Nine Different Vendors J Am Soc Echocardiogr 2015 28 1171 1181 10.1016/j.echo.2015.06.011 26209911 35. Pedrizzetti G Mangual J Tonti G On the geometrical relationship between global longitudinal strain and ejection fraction in the evaluation of cardiac contraction J Biomech 2014 47 746 749 10.1016/j.jbiomech.2013.12.016 24411099 36. Abraham TP Nishimura RA Myocardial Strain : Can We Finally Measure Contractility ?* J Am Coll Cardiol 2001 37 1 4 10.1016/S0735-1097(00)01173-6 11153722 37. Gorcsan J III Tanaka H Echocardiographic Assessment of Myocardial Strain J Am Coll Cardiol 2011 58 1401 1413 10.1016/j.jacc.2011.06.038 21939821 38. Goffinet C Chenot F Robert A Pouleur A-C le Polain de Waroux J-B Vancrayenest D Assessment of subendocardial vs. subepicardial left ventricular rotation and twist using two-dimensional speckle tracking echocardiography: comparison with tagged cardiac magnetic resonance Eur Heart J 2009 30 608 617 10.1093/eurheartj/ehn511 19019994 39. Kowallick JT Morton G Lamata P Jogiya R Kutty S Lotz J Inter-study reproducibility of left ventricular torsion and torsion rate quantification using MR myocardial feature tracking J Magn Reson Imaging 2016 43 128 137 10.1002/jmri.24979 26114731 40. Green AE Adkins JE Large elastic deformations 1960 Oxford Clarendon 41. Moore CC Lugo-Olivieri CH McVeigh ER Zerhouni EA Three-dimensional systolic strain patterns in the normal human left ventricle: characterization with tagged MR imaging Radiology 2000 214 453 466 10.1148/radiology.214.2.r00fe17453 10671594 42. Hess AT Zhong X Spottiswoode BS Epstein FH Meintjes EM Myocardial 3D strain calculation by combining cine displacement encoding with stimulated echoes (DENSE) and cine strain encoding (SENC) imaging Magn Reson Med 2009 62 77 84 10.1002/mrm.21984 19322795 43. Zhong X Spottiswoode BS Meyer CH Kramer CM Epstein FH Imaging three-dimensional myocardial mechanics using navigator-gated volumetric spiral cine DENSE MRI Magn Reson Med 2010 64 1089 1097 10.1002/mrm.22503 20574967 44. Pedrizzetti G Kraigher-Krainer E De Luca A Caracciolo G Mangual JO Shah A Functional strain-line pattern in the human left ventricle Phys Rev Lett 2012 109 048103 10.1103/PhysRevLett.109.048103 23006111 45. Piras P Evangelista A Gabriele S Nardinocchi P Teresi L Torromeo C 4D-Analysis of Left Ventricular Heart Cycle Using Procrustes Motion Analysis PLoS One 2014 9 e86896 10.1371/journal.pone.0086896 24466282 46. Pedrizzetti G Sengupta S Caracciolo G Park CS Amaki M Goliasch G Three-Dimensional Principal Strain Analysis for Characterizing Subclinical Changes in Left Ventricular Function J Am Soc Echocardiogr 2014 27 1041 1050 10.1016/j.echo.2014.05.014 24998517 47. Onishi T Saha SK Ludwig DR Onishi T Marek JJ Cavalcante JL Feature tracking measurement of dyssynchrony from cardiovascular magnetic resonance cine acquisitions: comparison with echocardiographic speckle tracking J Cardiovasc Magn Reson 2013 15 95 10.1186/1532-429X-15-95 24134158 48. Duchateau N De Craene M Piella G Frangi AF Constrained manifold learning for the characterization of pathological deviations from normality Med Image Anal 2012 16 1532 1549 10.1016/j.media.2012.07.003 22906821 49. Zerhouni EA Parish DM Rogers WJ Yang A Shapiro EP Human heart: tagging with MR imaging--a method for noninvasive assessment of myocardial motion Radiology 1988 169 59 63 10.1148/radiology.169.1.3420283 3420283 50. Wu L Germans T Güçlü A Heymans MW Allaart CP van Rossum AC Feature tracking compared with tissue tagging measurements of segmental strain by cardiovascular magnetic resonance J Cardiovasc Magn Reson 2014 16 10 10.1186/1532-429X-16-10 24450803 51. Osman NF McVeigh ER Prince JL Imaging heart motion using harmonic phase MRI IEEE Trans Med Imaging 2000 19 186 202 10.1109/42.845177 10875703 52. Aletras AH Ding S Balaban RS Wen H DENSE: displacement encoding with stimulated echoes in cardiac functional MRI J Magn Reson 1999 137 247 252 10.1006/jmre.1998.1676 10053155 53. Osman NF Sampath S Atalar E Prince JL Imaging longitudinal cardiac strain on short-axis images using strain-encoded MRI Magn Reson Med 2001 46 324 334 10.1002/mrm.1195 11477637 54. Amundsen BH Helle-Valle T Edvardsen T Torp H Crosby J Lyseggen E Noninvasive myocardial strain measurement by speckle tracking echocardiography: Validation against sonomicrometry and tagged magnetic resonance imaging J Am Coll Cardiol 2006 47 789 793 10.1016/j.jacc.2005.10.040 16487846 55. Bansal M Cho GY Chan J Leano R Haluska BA Marwick TH Feasibility and Accuracy of Different Techniques of Two-Dimensional Speckle Based Strain and Validation With Harmonic Phase Magnetic Resonance Imaging J Am Soc Echocardiogr 2008 21 1318 1325 10.1016/j.echo.2008.09.021 19041575 56. Seo Y Ishizu T Enomoto Y Sugimori H Yamamoto M Machino T Validation of 3-dimensional speckle tracking imaging to quantify Regional Myocardial Deformation Circ Cardiovasc Imaging 2009 2 451 459 10.1161/CIRCIMAGING.109.858480 19920043 57. Kaku K Takeuchi M Tsang W Takigiku K Yasukochi S Patel AR Age-related normal range of left ventricular strain and torsion using three-dimensional speckle-tracking echocardiography J Am Soc Echocardiogr 2014 27 55 64 10.1016/j.echo.2013.10.002 24238753 58. Schuster A Stahnke VC Unterberg-Buchwald C Kowallick JT Lamata P Steinmetz M Cardiovascular magnetic resonance feature-tracking assessment of myocardial mechanics: Intervendor agreement and considerations regarding reproducibility Clin Radiol 2015 70 989 998 10.1016/j.crad.2015.05.006 26139384 59. Ohyama Y Ambale-Venkatesh B Chamera E Shehata ML Corona-Villalobos CP Zimmerman SL Comparison of strain measurement from multimodality tissue tracking with strain-encoding MRI and harmonic phase MRI in pulmonary hypertension Int J Cardiol 2015 182 342 348 10.1016/j.ijcard.2015.01.016 25590963 60. Azam S Desjardins CL Schluchter M Liner A Stelzer JE Yu X Comparison of velocity vector imaging echocardiography with magnetic resonance imaging in mouse models of cardiomyopathy Circ Cardiovasc Imaging 2012 5 776 781 10.1161/CIRCIMAGING.111.972406 22977126 61. Amaki M Savino J Ain DL Sanz J Pedrizzetti G Kulkarni H Diagnostic concordance of echocardiography and cardiac magnetic resonance-based tissue tracking for differentiating constrictive pericarditis from restrictive cardiomyopathy Circ Cardiovasc Imaging 2014 7 819 827 10.1161/CIRCIMAGING.114.002103 25107553 62. Padiyath A Gribben P Abraham JR Li L Rangamani S Schuster A Echocardiography and cardiac magnetic resonance-based feature tracking in the assessment of myocardial mechanics in tetralogy of fallot: An intermodality comparison Echocardiography 2013 30 203 210 10.1111/echo.12016 23167248 63. Kempny A Fernández-Jiménez R Orwat S Schuler P Bunck AC Maintz D Quantification of biventricular myocardial function using cardiac magnetic resonance feature tracking, endocardial border delineation and echocardiographic speckle tracking in patients with repaired tetralogy of fallot and healthy controls J Cardiovasc Magn Reson 2012 14 32 10.1186/1532-429X-14-32 22650308 64. Smiseth OA Torp H Opdahl A Haugaa KH Urheim S Myocardial strain imaging: how useful is it in clinical decision making? Eur Heart J 2016 37 1196 1207 10.1093/eurheartj/ehv529 26508168 65. Russo C Jin Z Elkind MSV Rundek T Homma S Sacco RL Prevalence and prognostic value of subclinical left ventricular systolic dysfunction by global longitudinal strain in a community-based cohort Eur J Heart Fail 2014 16 1301 1309 10.1002/ejhf.154 25211239 66. Lang RM Badano LP Mor-Avi V Afilalo J Armstrong A Ernande L Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging J Am Soc Echocardiogr 2015 28 1 39 10.1016/j.echo.2014.10.003 25559473 67. Nagueh SF Smiseth OA Appleton CP Byrd BF Dokainish H Edvardsen T Recommendations for the Evaluation of Left Ventricular Diastolic Function by Echocardiography: An Update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging J Am Soc Echocardiogr 2016 29 277 314 10.1016/j.echo.2016.01.011 27037982

PMC005xxxxxx/PMC5000425.txt

==== Front BMC Public HealthBMC Public HealthBMC Public Health1471-2458BioMed Central London 339610.1186/s12889-016-3396-zResearch ArticleWhy do marital partners of people living with HIV not test for HIV? A qualitative study in Lusaka, Zambia Musheke Maurice mushekem@yahoo.com 1Merten Sonja sonja.merten@unibas.ch 23Bond Virginia GBond@zambart.org.zm 451 Population Council, Private Bag RW319x, Lusaka, Zambia 2 Swiss Tropical and Public Health Institute, Socinstrasse 57, CH-4002 Basel, Switzerland 3 University of Basel, Petersplatz 1, CH-4003 Basel, Switzerland 4 Zambart Project, University of Zambia, P.O. Box 50697, Lusaka, Zambia 5 Department of Global Health and Development, Faculty of Public Health and Policy, London School of Hygiene and Tropical Medicine, Keppel Street, London, WC1E 7HT UK 25 8 2016 25 8 2016 2016 16 1 88210 11 2015 28 7 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Knowledge of HIV status is crucial for HIV prevention and management in marital relationships. Yet some marital partners of people living with HIV decline HIV testing despite knowing the HIV-positive status of their partners. To date, little research has explored the reasons for this. Methods An exploratory qualitative study was undertaken in Lusaka, Zambia, between March 2010 and September 2011, nested within a larger ethnographic study. In-depth interviews were held with individuals who knew the HIV-positive status of their marital partners but never sought HIV testing (n = 30) and HIV service providers of a public sector clinic (n = 10). A focus group discussion was also conducted with eight (8) lay HIV counsellors. Data was transcribed, coded and managed using ATLAS.ti and analysed using latent content analysis. Results The overarching barrier to uptake of HIV testing was study participants’ perception of their physical health, reinforced by uptake of herbal remedies and conventional non-HIV medication to mitigate perceived HIV-related symptoms. They indicated willingness to test for HIV if they noticed a decline in physical health and other alternative forms of care became ineffective. Also, some study participants viewed themselves as already infected with HIV on account of the HIV-positive status of their marital partners, with some opting for faith healing to get ‘cured’. Other barriers were the perceived psychological burden of living with HIV, modulated by lay belief that knowledge of HIV-positive status led to rapid physical deterioration of health. Perceived inability to sustain uptake of life-long treatment – influenced by a negative attitude towards treatment – further undermined uptake of HIV testing. Self-stigma, which manifested itself through fear of blame and a need to maintain moral credibility in marital relationships, also undermined uptake of HIV testing. Conclusions Improving uptake of HIV testing requires a multi-pronged approach that addresses self-stigma, lay risk perceptions, negative treatment and health beliefs and the perceived psychological burden of living with HIV. Strengthening couple HIV testing services, including addressing conflict and addressing gendered power relationships are also warranted to facilitate joint knowledge, acceptance and management of HIV status in marital relationships. Keywords HIVHIV testingMarital partnerAntiretroviral therapyCouple counsellingZambiahttp://dx.doi.org/10.13039/501100001711Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung (CH)PDFMP3-123185Merten Sonja issue-copyright-statement© The Author(s) 2016 ==== Body Background A 2016 Joint United Nations Programme on AIDS (UNAIDS) report indicates that half of all people living with HIV (PLHIV) are unaware of their HIV status [1]. In sub-Saharan Africa (SSA), which has a generalized HIV epidemic, most HIV infections occur in marital or cohabiting relationships [2–4], for instance, 50–65 % in Swaziland, 35–62 % in Lesotho and 44 % in Kenya [4]. According to data from 27 cohorts totaling 13,061 sero-discordant couples in SSA, and Demographic and Health Survey (DHS) data on 1,145 sero-discordant couples in 14 countries, the proportion of HIV-positive women in stable heterosexual sero-discordant relationships was 47 % [5]. Despite the progressive roll-out of different HIV testing initiatives in many settings of SSA as a crucial HIV prevention strategy, barriers to HIV testing have persisted. These barriers include stigma and discrimination [5, 6], self-perception of being at less risk of infection [7–9], perceived inability by service providers to maintain confidentiality [8, 10, 11], lack of symptoms or deterioration of health [12–14] and women’s lack of control over decisions related to HIV testing [15–17]. Zambia has an estimated HIV prevalence of 13.3 % in the adult population aged 15–49 years [18]. The HIV prevalence peaked in the late 1990s before levelling off and declining to current rates [19]. For marital partners, the need to test as an HIV prevention and management strategy is even more critical. Empirical data on urban Zambia shows that at least 60 % of new heterosexual HIV infections occur within marriages or cohabiting relationships [20], and among married/cohabiting partners, discordance rate is about 11 % [18]. HIV testing by marital partners is thus critical to ensure: increased uptake and adherence to treatment [5]; increased uptake and adherence to HIV treatment for own health (which in turn decreases drug resistance, morbidity and mortality); adoption of risk reduction sexual behaviour [5, 21, 22]; decreased stigma and normalization [5]; and uptake of treatment for prevention of mother-to-child transmission of HIV by women [5, 23, 24]. While a recent systematic review has synthesised the factors influencing uptake of HIV testing in SSA [25], there is still a dearth of information on specifically why individuals aware of the HIV-positive status of their marital partners do not seek HIV testing. Understanding these barriers is critical in the prevention of HIV transmission and management of HIV within marital relationships. Therefore, to contribute to the body of knowledge on barriers to uptake of HIV testing, this study reports the reasons for non-uptake of HIV testing by marital partners of People Living with HIV (PLHIV) who knew the HIV-positive status of their partners. Methods Study design This was an exploratory qualitative study nested within a larger 18-month ethnographic study on factors influencing uptake of HIV testing, non-initiation of and retention in ART care. This study design was suitable for identifying and eliciting in-depth insight into factors hindering uptake of HIV testing by individuals who knew the HIV-positive status of their marital partners. In this study, marital partnership referred to a man and women who were officially married whether under statutory or customary law, and were living together as husband and wife. Study setting The study was conducted in a low-income, high-density urban setting, located about 10 km south of the Business District in Lusaka, the capital city of Zambia. Based on field observations, common physical features of the study setting are a crowded mix of formal and informal housing structures, shops and market stalls, one main tarred road with adjoining dusty, unlabeled roads that become water-logged and muddy during the rainy season, and poor solid waste disposal facilities. The majority of the residents have developed both strong kin and non-kin social network relationships. Most of the households are large, in part due to the negative impact of HIV, with some individuals growing up as orphans under the care of extended family members. Although not all family members live together, they still maintain reciprocal social and economic support ties. Other social network relationships are a product of religious affiliations and occupational and social lifestyle activities. Public social amenities are non-existent in the area. Social life often revolves around spending time in the bars, night clubs and make-shift drinking places; these places often serve as one source of sexual network relationships. Economically, the living conditions of local residents are mixed. Some people are formally employed in government and the private sector. The majority of the people earn their living in the informal sector of the economy, mostly as traders selling fruits, vegetables, meat products, fish, charcoal and second-hand clothes in the city centre markets, other markets within Lusaka, and in the open-air local markets. The unemployment situation is further exacerbated by rural–urban migration, as people move into the city in search of job opportunities and a better life. Health services are mainly accessed from a public health centre. The health centre has an out-patient unit, an in-patient unit with female and male admission wards, ‘opt-in’ HIV counselling and testing (HCT) unit, an Antiretroviral Therapy (ART) unit, a Maternal and Child Health (MCH) unit which provides antenatal and postnatal health services, and a Tuberculosis (TB) screening and treatment unit. In addition, mobile HCT services in the area are periodically provided by non-governmental HIV service providers. HCT and ART services are provided free of charge. Free couple HCT services are provided at the MCH unit and the opt-in HCT unit of the health centre. The health centre also previously housed a couple HCT project implemented by a local organisation called Zambia Emory HIV research project (ZEHRP). By March 2010, when this qualitative study started, the public health centre had more than 5,000 people on ART and more than 5,000 on pre-ART. There is also a plethora of privately owned clinics and drug stores. Other health service providers in the area include herbalists, traditional doctors and faith healers, some of whom advertise their services, including ‘cure’ of HIV and sexually transmitted infections (STIs). Tied to healing is Christianity, which is the dominant religion in the area, with a myriad of charismatic evangelical Pentecostal churches, some of which provide faith healing sessions for people suffering from different health conditions, including HIV. Participant selection Marital partners of PLHIV were identified, contacted and recruited through their marital partners receiving ART care at a local health centre. A two stage-recruitment strategy was used to recruit this group of study participants. First, ART clinic staff purposively identified PLHIV who disclosed their status to their marital partners but whose partners opted not to seek HIV testing. Second, PLHIV were then asked to recruit, on behalf of the study, their partners for interviews. Only those spouses who agreed to participate were contacted by the research team to schedule time and location for interviews. Using snowball and opportunistic sampling techniques, health care workers (nurses and lay HCT counsellors) who were involved in the delivery of HCT and ART services were identified and recruited from the various units of the public health centre. Data collection and analysis Data was collected between March 2010 and September 2011 as part of the first author’s doctoral studies in epidemiology. To ensure consistency in the administration of the research tools, all interviews were conducted by the first author, a social scientist with extensive experience in designing and conducting qualitative research. Thirty-eight (38) PLHIV and receiving ART at the local public health center were approached to recruit their marital partners for the study. Thirty-four (34) agreed to talk to and recruit their spouses for the study, out of which thirty (30) participated in open ended, face-to-face, audio-recorded in-depth interviews. The first author lived in the study setting for the entire period of data collection (18 months) and this enabled him to win the trust and confidence of the study participants, thereby enabling them to open up and share their perspectives. In addition, in-depth interviews were held with health care workers involved in the provision of HCT and ART services (n = 10). One focus group discussion (FGD) was also conducted with the health facility-based lay HCT counsellors (n = 8). Five (5) of the HCT counsellors were women and the rest were men. No repeat interviews or FGDs were conducted with study participants. The main research question asked was: “What are the reasons for not seeking HIV testing despite knowing the HIV-positive status of your/their spouses?” Interviews with health care workers and FGD with HCT counsellors were conducted in English while interviews with spouses of PLHIV were conducted in Nyanja - the local language mainly spoken in the area. The in-depth interviews lasted between 30 and 45 min and the FGD with HIV counsellors lasted about an hour. Data collection and preliminary data analysis was a cyclical process. The data collection tools were first piloted and fine-tuned. During actual data collection, interview data informed ensuing interviews and data collection was ended when emerging data became repetitive. All interviews conducted in local language were translated and all interviews were transcribed verbatim. The transcripts were then entered into, and organised and managed using, ATLAS.ti version 6. The data was then coded inductively. The first author developed the coding framework and coded all the data, which were reviewed and approved by the other authors. Team meetings were used to discuss and resolve differences regarding the coding framework and the codes and themes generated. Qualitative latent content analysis [26] was used to analyse and interpret the data. Latent content analysis involves an analysis of the relationship aspects of the textual data and an interpretation of the underlying meaning of the text, referred to as the latent content [26]. All interview and FGD transcripts constituted our unit of analysis. Unit of analysis refers to all words and phrases of the interview and FGD transcripts [26]. They were read several times to create a sense of the whole data [26, 27]. Within-case and across-case analysis [28] of the interview transcripts was undertaken to inductively generate concepts across the individual interviews. For each interview transcript, we conducted within-case analysis and retrieved and coded reasons for not seeking HIV testing despite knowing the HIV-positive status of a marital partner. Thereafter, we conducted across-case analysis by comparing and contrasting participants’ reasons. Similar codes were then put together to form categories. A category is therefore a group of content that shares a commonality; it is a thread throughout the codes [26]. Themes were then developed by interpreting categories for their underlying meaning. The themes are, therefore, the expression of the latent content (underlying meaning) of the textual data [26]. For instance, codes such as ‘feeling healthy’, ‘not sick’, ‘nothing wrong in the body’ were categorised as ‘state of physical health’ as described in the results section of the paper. The theme generated from this is ‘lay wellness and illness beliefs.’ These themes are described in the discussion section of the paper. Three reference points were used to identify emergent themes: recurrence, repetition and forcefulness of ideas within the interview data [29]. Through this analytical strategy, we were able to identify themes that cut across study participants but were still grounded in individual perspectives [28]. Lastly, we selected interview and group discussion excerpts that best illustrated these themes. Arising from the rapport and relationships created with study participants during ethnographic fieldwork, emerging findings were shared with study participants. Ethical considerations The study was approved by the Ethics Committee in Basel (Ethik-Kommission beider Basel) and the University of Zambia School of Humanities and Social Sciences Research Ethics Committee, as part of the research project ‘Improving equity of access to HIV care and treatment in Zambia’. Written informed consent was obtained from all study participants. To ensure confidentiality, interviews and FGD with health centre staff took place in private spaces of the health facility while interviews with marital partners of PLHIV took place at locations of their choice. Some preferred to be interviewed at home, in the absence of their spouses; others preferred to be interviewed at neutral locations such as private spaces at public health centre and homes study participants’ friends and relatives. To protect the identity of study participants, all identifying information was excluded from the interview transcripts. Study participants were given a reimbursement of ZMK50.00 (about US$10 at the time) as compensation for their time. In addition and when appropriate, study participants were provided with refreshments and given transport reimbursement of ZMK50.00 (about US$10 at the time) when interviews were conducted outside their homes. Results Characteristics of study participants Half of the marital partners of PLHIV interviewed were women; more than half were aged ≥35 years with the oldest being a 51-year old man. The majority of the study participants were in informal employment; all those in formal employment (n = 5) were men. More than half had been married for at least 4 years (Table 1). Two-thirds (n = 20) had known the HIV-positive status of their spouses for at least 2 years and all the marital partners of PLHIV were on ART.Table 1 Characteristics of marital partners of PLHIV Characteristic Number of Respondents Age (Years)  18–24 4  25–34 9  35–44 10  >44 7 Sex  Male 15  Female 15 Source of livelihood  Formal employment 5  Informal employment 18  Not working/dependant 7 Duration of marital relationship  <12 months 0  1–<2 years 1  2–<3 years 7  3–<4 years 4  4–<5 years 10  ≥5 years 8 Duration since known HIV status of partner  6–<12 months 4  1–<2 years 6  2–<3 years 9  3–<4 years 9  ≥4 years 2 Treatment status of HIV+ marital partner  On ART 30  Not on ART 0 The health care workers interviewed belonged to two categories - professional nurses and lay HCT counsellors. The nurses were full-time government employees while lay HCT counsellors worked as volunteers. The lay HCT counsellors, some of whom were living with HIV, worked as HIV testing and treatment supporters. Their responsibilities included providing pre and post HIV counselling, HIV treatment preparedness and adherence counselling, and tracing treatment defaulters (HIV patients not reporting for clinical appointment or pick-up of medication) and bringing them back into care. The nurses and counsellors were all drawn from the HCT, ART, Mother and Child Health (MCH) and TB units of the public health centre. Reasons for non-uptake of HIV testing State of physical health Participants’ own perception of their physical health was the overarching barrier to uptake of HIV testing. They measured good health in terms of functional ability and not the clinical presence of HIV infection. Therefore, poor health was viewed as the sine qua non for seeking HIV testing. Asked when they would consider going for HIV testing, one study participant bluntly said: “[I will test] when I get sick, I mean as in being bed-ridden. Not when I am still strong, I do not see any need to test and I pray that I will not get to that stage” (30-year old woman). Another study participant replied: “She [wife] went to the clinic; she tested and was given her results. I saw the results as well.... But me, I am just ok.…Because I rarely get sick, I do not see any need to go and test. So it is very difficult to just go and test when you feel that there is nothing wrong in your body.” (47-year old man) The reluctance to test when still physically strong was echoed by health care providers. A nurse at the ART clinic noted that: “I think from what I have observed, somebody might know their HIV status but because they still have energy to run around and do their work, they would not care about coming for HIV testing. Most of them wait until they have serious opportunistic infections. That is when we see them rushing here to the clinic.” For some men, the perception of being physically healthy could be attributed to the notion of masculinity. They needed to show that they were strong and not reliant on HIV testing and treatment to sustain their physical health. As one male study participant explained: “When you go and test, when you start taking the drugs, it means you have given up as a man that you are not strong enough on your own; you become drug-dependent, you see what I mean?” (25-year old man). Ironically, although some male respondents showed unwillingness to test to show that they could live healthy and normal lives without medication, they still adopted self-care practices of using herbal remedies and conventional non-HIV medication to deal with symptoms of opportunities infections, such as periodic episodes of diarrhoea, rash, sore throats and coughing. Fear of blame by marital partner Non-uptake of HIV testing was also used to fend off blame or accusations of being responsible for HIV infection. This pattern of blame was most often levelled and articulated by men towards women. It was also often the case in distrustful relationships, in which some study participants blamed their HIV-positive partners for being responsible for HIV infection. Therefore, by not seeking HIV testing, some study participants pointed out that they were able to maintain moral credibility in the marital relationship, that they were not responsible for HIV infection. While a few men suspected their female partners to be responsible for HIV infection, the majority of them acknowledged being the possible source of HIV infection due to extra-marital relationships. Therefore, not testing for HIV was meant to avoid confirming their partners’ suspicions - marital infidelity. One male study participant said: “If my wife knows that I am HIV-positive, she is going to say “yes, it is because you had other sexual partners”. Sometimes the best way to avoid problems with women is not to test. They cannot blame you for HIV because you have not tested even if deep down your heart, you know that you may be the one who contracted HIV and then infected your wife” (32-year old man). For women, and due to power imbalance in marital relationships, they feared that seeking HIV testing could prompt their partners to shift the blame for HIV infection onto them. One female study participant said: “Sometimes we women fear that if we went for testing, men would take advantage of the situation and shift the blame on us. So to avoid being blamed, some women opt not to go and test.” (45-year old woman). These fears were confirmed by lay HCT counsellors and professional health care workers. One lay HCT counsellor said: “Usually the ones that behave badly are men; when the wife is HIV-positive or even when both are positive, they (men) deny it or claim that they are not responsible for the infection. So women become reluctant to test for HIV for fear of being blamed” (Female VCT counsellor, FGD). Perception of already being infected with HIV The majority of the spouses of PLHIV interviewed, across gender, used the HIV-positive status of their partners as a representation of their own HIV status. While they acknowledged the existence of discordance and the possibility of being in a discordant relationship, the chances were described as remote, in part because of the length of time they had lived with their spouses and the number of times they had unprotected sex with their partners. Testing for HIV was therefore viewed as unnecessary when the likelihood of being infected was already high. One study participant indicated that: “I was sure that I was also going to get the same results. So there was no need for me to go for the test. You know we have had unprotected sex and one can get HIV through unprotected sex. Since my husband is HIV positive, I am also HIV-positive” (20-year old woman). Similarly, a 47-year old man said: “I also believe that I am HIV-positive because my wife has already tested HIV-positive”. Perceived psychological burden of living with HIV While some study participants acknowledged that the HIV-positive status of their partners meant that they too could be infected, paradoxically, the perceived psychological burden of confirming that they also had HIV undermined the uptake of HIV testing. They preferred to live without knowing their HIV status. They perceived knowing one’s HIV status as having a deleterious mental health effect, which in turn was perceived as hastening deterioration of physical health. Some respondents narrated how their HIV-positive spouses had exhibited despair, became irritable, non-sociable and some adopted a fatalistic attitude towards life on account of their HIV-positive status. They worried that this could trigger adoption of similar behaviour if they sought HIV testing. When asked why he had not sought HIV testing, a 47-year old man explained that: “When you are just feeling ok, to go and test and be found HIV-positive would just bring psychological problems. You start thinking too much about your health and your life and how your new status will change your life.” For some study participants, because HIV is not yet curable, knowing one’s HIV-positive status was synonymous with death itself. A 29-year old woman interviewed said: “You know when you have HIV, even death comes to your door step because the disease cannot be cured. They say that there are drugs, but those ARVs do not cure HIV. So, it is better that I don’t know that I will die very soon.” Paradoxically, this was the case despite the availability of life-prolonging ART and the acknowledgement that people on treatment were living longer and healthy lives. Lack of self-efficacy Since HIV testing is a pre-requisite for initiating HIV treatment, lack of self-efficacy, namely the inability to sustain positive treatment behaviour, dissuaded some study participants from testing. The lack of self-efficacy was attributed to three main reasons. First, some study participants felt that they would not manage to strictly adhere to the HIV treatment regimen, a requisite for achieving optimal viral suppression and living a normal and healthy life. Some narrated their experiences of failing to adhere to treatment for other health conditions such as malaria, which were short-term in nature, and therefore doubted their ability to stick to the HIV treatment regimen for the rest of their lives. They disliked the taste and smell of the drugs, a factor they said could impact on their ability to take their medication every day. One study participant explained: “The problem with ARVs is that once you start taking ARVs, you take them for life. But herbs [alternative medication], you take when you feel like; and you can decide not to take for say 3 months. ARVs are very demanding” (30 year old woman). The influence of perceived inability to adhere to treatment on uptake of HIV testing was echoed by health workers. During the FGD, one of the male lay HCT counsellors said: “Sometimes people worry about taking the drugs for the rest of their lives. So they fear to test because they are not ready to start treatment in case they are found HIV-positive.” Second, some study participants reported the challenges of integrating treatment into their day-to-day livelihood activities. This was particularly the case with respondents who were in self-employment, such as taxi drivers, construction workers and small-scale cross boarder traders, most of whom spent long hours away from their home. They feared that that they would default on accessing and taking their medication. As one 31-year old male study participant who worked as a taxi driver remarked: “So, I thought that with my kind of work, because I work as a taxi driver, I wondered if I was going to manage to be driving if they say that those drugs make one feel drowsy. How am I going to manage to drive the car? You cannot be drowsy while driving, you will cause an accident, and I also cannot stop working because that is my only source of livelihood.” Thirdly, lack of self-efficacy was attributed to the side effects of HIV medication. The most common side effects mentioned were nausea, rash, feeling drowsy, changes in the texture of the skin and numbness, tingling or burning sensations. Fear of lipodystrophy - fat redistribution due to medication - was also a common theme. Asked why she had not sought HIV testing, a 29-year old female study participant answered: “Me, I said, I will not go and test because I do not want to start ARVs. Because I have seen a lot of people taking ARVs face a lot of problems like having swollen legs, complaining of headache, they develop “ichifungalashi” (body numbness) that they cannot do anything.” Local descriptions such as ‘kusintha kwa thupi’ (physical changes to the body), ‘kuonongeka mawonekedwe’ (loss of physical appearance of the body) and‘kutupa’ (to look swollen) were widely used to describe the fear of these treatment-induced body changes. More so, observing the struggles of their spouses as they experienced and confronted treatment-related side effects did little to bolster their interest in and embolden them to seek HIV testing - the first step to get into treatment and care. This made them question the value of testing if they would not start treatment in the first place. The negative attitude towards medication was reinforced by some deeply held community beliefs that HIV medication was insidiously harmful, and death of some PLHIV was attributed to the medication itself. A study participants whose aunt and two young sisters died shortly after being put on HIV medication attributed their death to the effects of the medication. A 42-year old male study participant explained: “Then I also hear that those drugs destroy the liver, and that is why people keep on going to the clinic to check their liver. What kind of drugs are those? On one hand they say they prolong your life, but at the same time the same drugs are destroying your liver”. A 30-year old female study participant revealed that a friend’s husband went blind after being started on treatment and “He was told that the drugs were too strong and the veins collapsed.” Herbal remedies and non-HIV medication to mitigate HIV-related symptoms As an alternative to HIV testing and starting treatment if found HIV-positive, all the spouses of PLHIV interviewed reported seeking recourse to alternative care - herbal remedies and conventional non-HIV medication to mitigate HIV-related symptoms. Forever Aloe Vera gel - a refined herbal product from South Africa - and tembusha (a local herbal plant), and a plethora of herbal remedies generally known as ‘immune boosters’ (including Chinese herbs) were the most widely used herbal remedies. The cost of these herbal remedies varied, with the most expensive being Aloe Vera gel, at US$15 per 1 l container. Due to high demand for herbal remedies, advertisements for different herbal products had become widespread; in the daily newspapers, through pamphlets handed out to motorists and pedestrians, posters stuck on trees and walls of private and public buildings. Some participants also reported treating opportunistic infections using conventional non-HIV medication such as co-trimoxazole prophylaxis used by their HIV-positive marital partners. This in turn undermined uptake of HIV testing. Two study participants narrated this self-care practice. One said: “I saw that people that have HIV but have not yet qualified to start ARVs take septrin. So, because I suspect that I have HIV, I just take septrin….Sometime back, when he (husband) was on septrin, we used to share the medication. So that is when I started organizing it on my own” (45-year old woman). Another stated that: “Even if I have not tested to confirm my HIV status, at least I take herbs in order to boost my immune system. At the moment, I am using tembusha. I make a 2.5 litre herbal solution and I take a glass in the morning, in the afternoon and in the evening” (23-year old woman). In addition, some study participants reported opting for faith healing instead of seeking HIV testing. This option has become common practice, in part, due to the growth of the evangelical Christian movement in Zambia. Some participants reported buying ‘anointing water’ - bottled water purported to have healing properties - from local churches and churches outside Zambia. Some ‘anointing water’ from Nigeria reportedly cost as much as US$100 for a 100 ml bottle. Others reported touching their television sets during Christian healing programmes on television in the hope of getting cured of HIV. Some acknowledged the power of prayer and faith in God in addressing health conditions, including incurable infections such as HIV. When asked about prayer and healing, one study participant acknowledged that: “Yes, and it depends on your faith, like there is this friend of mine who went to [Dr…], she went there and she was prayed for....So people in the community would prefer going to churches to be prayed for than to come to the clinic… for HIV testing.” (28-year old woman). Self-stigma Anticipated stigma - fear of ‘othering’, social isolation, gossip, labelling and public shaming - was not reported as a barrier to HIV testing. This was attributed to perceived low level of stigma in the community. This view was shaped by the perceived experiences of their HIV-positive partners. Instead, self-stigma (feeling of guilt and shame for having a stigmatising condition) and as already described above, moralisation and blaming behaviour within marital relationships, negatively affected uptake of HIV testing. A 31-year old man explained that: “In the community there, stigma is no longer there. People would just sympathise with you. What is just remaining is self-stigma. You just feel guilty and ashamed that you have HIV. I avoid this feeling by not going for testing.” This view was also shared by professional health care workers and lay HCT counsellors. When asked whether fear of enacted stigma prevented spouses of PLHIV from seeking HIV testing, a nurse based at the ART clinic disagreed that: “No, stigma is very low…. You know some of our clients when they come here for treatment, at times their phones will ring and they would say ‘my friend, I am at the clinic, I am getting my ARVs.’ You know such statements and openness can give you an impression that people have now taken it [HIV] to be normal….What is killing people is the self-stigma that they have.” The barriers to uptake of HIV testing described above are not mutually exclusive; they are interrelated and some coalesce to undermine HIV testing behaviour. For instance, while lay self-assessment of risk of HIV infection reduces motivation to test, perceived state of physical health, sometimes due to the use of herbal remedies and non-HIV conventional medication to mitigate HIV-related symptoms, inhibits uptake of HIV testing (Fig. 1). Similarly, lack of self-efficacy, which is sometimes driven by negative attitude towards HIV-medication, leads to uptake of alternative health care and this in turn dissuades people from seeking HIV testing. Also, while self-stigma, which manifests itself in desire to maintain moral credibility and absolve oneself from blame of HIV infection, and perceived psychological burden of living with HIV undermine uptake of HIV testing, these barriers may be modulated by the state of physical health, with study participants indicating willingness to test and initiate treatment if their health deteriorated and herbal remedies and conventional non-HIV medication were no longer effective in sustaining physical health (Fig. 1).Fig. 1 Conceptual model of relationship of barriers to uptake of HIV testing Discussion Our study found that five inextricably linked barriers explained why marital partners of PLHIV did not seek HIV testing despite knowing the HIV-positive status of their partners. These were: quest to maintain moral credibility, influenced by notion of masculinity and gendered power relationships; self-stigma; use of partner HIV status as a proxy for own HIV status; lay HIV treatment and health seeking practices; and lay self-assessment of wellness and illness. These are discussed in detail below. These findings corroborate previous studies and have implications for scaling-up HIV testing and treatment, a critical strategy for achieving the UNAIDS 90–90–90 treatment target-where 90 % of people living with HIV know their status; 90 % of people who know their HIV status are accessing treatment; and 90 % of people on treatment have suppressed viral loads [1]. Maintenance of moral credibility, influenced by notion of masculinity and gendered power relationships One of our findings was that non-uptake of HIV testing was done to maintain high moral standing within marital relationships. This was particularly so in distrustful relationships. For men, non-uptake of HIV testing was meant to avoid any admission of guilt for HIV infection, thereby maintaining moral credibility within the marital relationship. For women, non-uptake of HIV testing was intended to reflect being viewed as supporting and accepting the behaviour of the HIV-positive partner that could have led to HIV infection. These findings corroborate those of Larsson and colleagues [30] who have reported that unstable and distrustful nature of marital relationships undermined uptake of HIV testing. These findings have implications for HIV prevention. Given the fragility of marital relationships, strengthening the promotion of couple HIV testing is critical to mediate blaming attitude and may improve couples’ ability to manage HIV in marital relationships. This is especially so given the reported high rate of new infections in steady, long-term partnerships [31]. Relatedly, our findings suggest that the quest to maintain moral credibility in marital relationships is inextricably linked to the notion of masculinity - male notion of power, influence and being in control, and gender inequality - subordination of women in male–female relationships. For men, particularly when still physically healthy, non-uptake of HIV testing may therefore be used as a strategy to maintain social status, reputation and power as seeking HIV testing may be perceived as a sign of weakness which could consequently diminish influence, control and power over their marital partners. These findings corroborates other studies in Uganda [32], Zimbabwe [33] and Tanzania [34] which encapsulate how masculinity acts as a barrier to men’s use of HIV services. Siu and colleagues have described two different forms of masculinity – respectable masculinity which promotes positive health-seeking behaviour and reputation masculinity which undermines it [35]. In our study, reputation masculinity was found to be at play in influencing non-uptake of HIV testing among men. Thus, improving uptake of HIV testing by men requires encouraging the positive aspects of masculinity such as being socially and economically responsible and de-constructing the notion of reputation masculinity that undermine positive health-seeking behaviour, such as the sense of being strong and in control. As a manifestation of gendered power relationships, for men, blaming women for HIV infection sometimes served other ends in relationships - to reinforce male control in marital relationships as exemplified by women’s fear that seeking HIV testing would lead to men shifting blame on them for HIV infection. A previous study in Uganda also reached this conclusion [36]. Similarly, Skovdal and colleagues [37] have reported that masculinity interferes with women’s uptake of HIV services. Our findings, therefore, call for the promotion of couple HIV testing to ensure joint knowledge and management of HIV status within marital relationships. Self-stigma The impact of stigma on uptake of HIV testing has been reported in other studies [6, 11, 17, 32, 38]. In this study, fear of enacted stigma [39] was not found to be a barrier to HIV testing, even after probing for its influence. This was attributed to the experiences of study participants’ HIV-positive partners who were reported as not having experienced stigma. Instead, self-stigma was found to undermine uptake of HIV testing as reflected by the quest to avoid carrying the burden of shame associated with having HIV and moralising and blaming pattern described above. This finding about enacted stigma experiences should however be interpreted with caution. It might be limited to this particular group of marital partners of people living with HIV. Interviewing HIV-positive partners would have elicited insights into whether indeed they had not experienced stigma. Notwithstanding this limitation, counselling efforts aimed at creating social acceptance of HIV infection at individual level as well as in marital relationships are still warranted. The findings also echo the call by UNAIDS for all countries to implement the 2012 World Health Organisation (WHO) guidance on couples HIV testing and counselling in order to reach the targets set in the United Nations 2011 Political Declarations on AIDS [40]. As a caveat, couple HIV testing should be implemented with caution taking into account marital partners’ ‘lived’ experiences as couple HIV counselling and testing could aggravate already fragile marital relationships as previously reported in our study on experiences with couple testing in Zambia [41]. Partner HIV status used as a proxy for own HIV status The finding that marital partners use the HIV-positive status of their partners as a marker for HIV infection is consistent with evidence reported elsewhere [6, 42]. What this suggests is that while having an HIV-infected partner elevates the risk of HIV infection, this does not ipso facto lead to uptake of HIV testing, as those at risk assume that they are already infected. This is despite the existence of HIV discordance in marital relationships. This finding underscores a need for more HIV awareness campaigns on the existence and possibility of HIV discordance in marital relationships, and the importance of couple HIV counselling and testing to facilitate joint knowledge of HIV status and management of HIV condition. Optimising uptake of HIV testing also requires addressing the disjunction between perceived and actual risk of HIV infection as previously reported in other studies in Zambia [43], Malawi [44, 45] and Nigeria [46]. Lay HIV treatment beliefs and health-seeking practices Previous studies have reported how lack of self-efficacy dissuades individuals living with HIV from initiation of treatment [47–52]. Our findings suggest that lack of self-efficacy also extends to those who do not know their HIV status, thus inhibiting uptake of HIV testing. This is despite the widely reported positive impact of availability of ART on uptake of HIV testing [6, 12, 38, 53, 54]. Therefore, improving uptake of HIV testing also requires addressing lack of self-efficacy, which is in part influenced by negative HIV treatment and health-related beliefs. This finding should also be treated with caution. Given the rapid upswing in the numbers of PLHIV on treatment in Zambia since this study was carried out, these negative beliefs may have shifted with longer experience of ART. However, sensitisation campaigns against the perceived efficacy of herbal remedies and faith healing - which dissuade people from seeking HIV testing and initiate treatment - are still warranted given the reported widespread use of these practices. The perception that treatment was insidiously harmful and had deleterious effects in the long-term was a recurring theme in this study, resulting in recourse to alternative forms of care - faith healing and herbal remedies - with the latter being reported as effective as antiretroviral treatment while hoping to get cured through the former. These treatment beliefs were couched in the experiences of patients started on early generation of antiretroviral treatment, whose adverse effects were reportedly severe and physically debilitating. What this suggests is that while a new generation of patient-tolerant antiretroviral drugs are now increasingly available, negative attitude towards treatment still persist. This does not only result in HIV-patient non-initiation of and attrition from antiretroviral treatment as reported in previous studies [52, 55]; it also undermines uptake of HIV testing - the first step in the ART care trajectory. In Zambia, the fear of HIV medication needs to be contextualised. In 2007, the pharmaceutical company Roche announced that some batches of viracept (an ARV used in second-line treatment) had been accidentally contaminated with mesylate, which can cause cancer and genetic mutation [56]. While the drug was immediately discontinued, this created panic among people on treatment and reinforced the perception that HIV medication was harmful. Similarly, Zambia has progressively revised its treatment guidelines and the 2013 guidelines recommended the accelerated phasing out of stavudine (d4T) and zidovudine (AZT) in first-line combined antiretroviral therapy (cART) regimens and replacing them with efficacious drugs with better overall toxicity profile [57]. This was in line with the 2009 WHO recommendations to progressive phasing out d4T due to adverse effects such as disfiguring, unpleasant and potentially life threatening toxicity effects [58]. In our study, some respondents viewed the periodic changes to treatment guidelines and drug substitutions as emblematic of the dangers of the HIV medication, culminating in reluctance to seek HIV testing. Therefore, sensitisations to assuage these lay negative treatment beliefs and the availability of new patient-tolerant drugs as opposed to people seeking faith healing and scientifically unproven herbal remedies are warranted. Lay self-assessment of wellness and illness We found that the overarching barrier to uptake of HIV testing was individuals’ perception of their corporeal health. Perceived good physical health dissuaded individuals from testing despite being at heightened risk of HIV infection. While study participants described a myriad of other factors inimical to uptake of HIV testing, and also reported resorting to alternative care, HIV testing was deferred until physical health had deteriorated. These findings are consistent with previous studies [9, 11, 38, 59]. This suggests that HIV testing is perceived narrowly as a gateway into treatment and care and not as a critical HIV prevention strategy. These findings corroborate those by Jürgensen and colleagues [6] who found that HIV testing was largely used as a diagnostic tool to access health care and not as an HIV prevention mechanism. Our findings suggest that the risk of declining health was counter-balanced by uptake of herbal remedies and conventional non-HIV medication to “boost” the immune system and to deal with episodic non-severe HIV-related symptoms. This undermined access to HIV testing and possible entry into ART care. In view of the drive towards the ‘test and treat HIV’ model to HIV prevention, treatment and care as suggested by Granich and colleagues [60], our findings suggest that without addressing them, lay wellness and illness beliefs cast serious aspersions on the viability of universal HIV testing and immediate treatment as an HIV prevention strategy. As a corollary, we found that fear of psychological burden associated with knowing one’s HIV status and its perceived negative impact on physical health dissuaded individuals from seeking HIV testing. This is consistent with previous studies [6, 13, 14, 38, 61, 62]. Our findings suggest that not seeking HIV testing despite acknowledging the possibility of being infected is used as a psychological buffer against the perceived mental burden of living with an incurable infection. This implies that despite the increasingly wider availability of life-saving ART, HIV still exudes fear of death in view of its incurable nature, and this fear may be exacerbated by the memories of suffering and death of people infected with HIV, including those on treatment. Thus, sensitization activities on the benefits of HIV testing regardless of physical health and addressing lay health beliefs that knowledge of HIV-positive status leads to rapid deterioration of health remain crucial in improving uptake of HIV testing. The complex interplay of these findings makes two points saliently clear: an individual may not face a single barrier to uptake of HIV testing. Similarly, HIV testing behaviour is not a linear, sequential process - that knowing the HIV-positive status of a partner would ipso facto lead to uptake of HIV testing. These findings, therefore, point to a need for a multi-pronged, context-specific and individualised approach to addressing multiple, inter-linked and sometimes mutually reinforcing factors that undermine uptake of HIV testing. Limitations of the study The study participants were recruited through their spouses receiving ART care at a local public sector clinic, and only those that agreed to be interviewed participated in the study. The findings may therefore not be representative of other individuals who refused to participate in the study. Consequently, this recruitment strategy could have led to some clustering of shared ideas and views. Interviewing the HIV-positive spouses could have provided more insights on non-testing behaviour through comparability of marital partners’ perspectives. Future studies should explore this further. A more general limitation concerns the generalisability of the findings. This study was conducted in a low-income setting with a small sample of respondents and aimed at gaining breadth and in-depth insights, in contrast to the purpose of quantitative research which aims to describe the frequency of such views. Therefore, similar studies are therefore warranted in other settings, including a larger, gender disaggregated sample size, for additional insights, such as the influence of other characteristics like age, duration of marital relationship and income status on uptake of HIV testing. Additionally, time has elapsed since the study was undertaken and since then, the number of PLHIV on treatment has greatly increased and alongside community experience of ART [63]. However, as our findings show, non-uptake of HIV testing still persists as demonstrated by some spouses who choose not to test for HIV despite knowing that the partner is living with HIV. Notwithstanding the limitations, the strength of this study was the diverse representation of our study participants in terms of key demographic characteristics - age, gender and economic livelihood - and knowledge of treatment status of marital partner. Interviews with health care providers also provided additional insights, including for triangulation of data. The findings could apply to similar settings in urban areas in the country and provide useful insights that can inform policy and practice to improve uptake of HIV testing. Conclusions HIV testing is an important strategy to combat HIV transmission and to facilitate entry into HIV care. However, our study has shown that knowing the HIV-positive status of a marital partner does not always lead to uptake of HIV testing. Instead, HIV testing behaviour is undermined by a complex and dynamic range of factors, which sometimes interact and coalesce. This study has found that individuals reach lay conclusions of already being infected on account of the HIV-positive status of their marital partners, thus viewing HIV testing as unnecessary. Testing is also not done to maintain moral credibility within marital relationship and to avoid legitimizing partner behaviour which could have led to HIV infection. Not knowing one’s HIV status is also aimed at creating a distance from HIV, thus acting as a buffer against the perceived psychological burden of living with HIV. While free HIV treatment has become widely available, its perceived negative effects and use of herbal remedies and conventional non-HIV medication to mitigate HIV-related symptoms and faith healing hamper uptake of testing. All these barriers appear to be modulated by the state of corporeal health, with individuals planning to test only after their health had deteriorated. Therefore, HIV care and prevention efforts should aim at addressing lay interpretations of risk of HIV infection, notion of masculinity, health and treatment beliefs, and promote the preventive and treatment benefits of early diagnosis of HIV. The reluctance by individuals to test despite knowing the HIV-positive status of their marital partners also calls for strengthening the promotion of couple HIV counselling and testing to ensure joint knowledge and management of HIV in marital relationships. Abbreviations ART, antiretroviral therapy; DHS, Demographic and Health Survey; HCT, HIV counselling and testing; HIV, Human Immunodeficiency Virus; MCH, Maternal and Child Health; PLHIV, people living with HIV; SSA, sub-Saharan Africa; STI, sexually transmitted infections; TB, Tuberculosis; UNAIDS, Joint United Nations Programme on AIDS; WHO, World Health Organisation; ZEHRP, Zambia Emory HIV Research Project Acknowledgements We thank the health workers of the local public sector clinic who helped to identify PLHIV whose marital partners had not tested for HIV. We also thank the study participants for agreeing to participate in the study. Funding This research was funded by Swiss National Science Foundation (Grant Number PDFMP3–123185). The funding agency played no role in the design of the study, data collection, analysis and interpretation of the data, and in the writing of, and decision to submit, the manuscript. Availability of data and materials The authors declare that the data supporting the findings of this study are available within the article. Authors’ contributions MM conceptualized the study, conducted data collection and analysis and wrote the draft manuscript. VB and SM contributed towards the conceptualization of the study, provided input in the analysis, interpretation of the findings and drafting of the manuscript. All authors have given final approval of the version to be published. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate The study was approved by the Ethics Committee in Basel (Ethik-Kommission beider Basel) and the University of Zambia School of Humanities and Social Sciences Research Ethics Committee, as part of the research project ‘Improving equity of access to HIV care and treatment in Zambia’. Administrative approval was also obtained from the Zambian Ministry of Health. Written informed consent was obtained from all study participants. Participation was voluntary and confidential. ==== Refs References 1. UNAIDS. Towards Universal Access: 2016-2016 Strategy: On the Fast-Track to end AIDS. Geneva: UNAIDS. Available at: http://www.unaids.org/sites/default/files/media_asset/20151027_UNAIDS_PCB37_15_18_EN_rev1.pdf. Accessed 31 May 2016. 2. Carpenter LM Kamali A Ruberantwari A Malamba SS Whiteworth JA Rates of HIV-1 transmission within marriage in rural Uganda in relation to the HIV sero-status of the partners AIDS 1999 13 1083 1089 10.1097/00002030-199906180-00012 10397539 3. Hugonnet S Mosha F Todd J Mugeye K Klokke A Ndeke L Ross D Grosskurth H Hayes R Incidence of HIV infection in stable sexual partnerships: A retrospective cohort study of 1802 couples in Mwanza region, Tanzania J Acquir Immune Defic Syndr 2002 30 1 73 80 10.1097/00042560-200205010-00010 12048366 4. Malamba S Mermin JH Bunnell RE Mubangizi J Kalule J Marum E Hu DJ Wangalwa S Smith D Downing R Couples at risk. HIV-1 concordance and discordance among sexual partners receiving voluntary counselling and testing in Uganda J Acquir Immune Defic Syndr 2005 39 5 576 580 16044010 5. WHO Guidance on couple HIV testing and counselling including antiretroviral therapy for treatment and prevention in sero-discordant couples: Recommendations for a public health approach 2012 Geneva WHO 6. Jürgensen M Tuba M Fylkesnes K Blystad A The burden of knowing: balancing benefits and barriers in HIV testing decisions. A qualitative study from Zambia BMC Health Serv Res 2012 12 2 10.1186/1472-6963-12-2 22222028 7. De Paoli MM Manongi R Klepp K Factors influencing acceptability of voluntary counselling and HIV-testing among pregnant women in Northern Tanzania AIDS Care 2004 16 4 411 425 10.1080/09540120410001683358 15203410 8. Bhagwanjee A Petersen I Akintola O George G Bridging the gap between VCT and HIV/AIDS treatment uptake: perspectives from a mining-sector workplace in South Africa Afri J AIDS Res 2008 7 3 271 279 10.2989/AJAR.2008.7.3.4.651 9. Chirawu P Langhaug L Mavhu W Pascoe S Dirawo J Cowan F Acceptability and challenges of implementing voluntary counselling and testing (VCT) in rural Zimbabwe: Evidence from the Regai Dzive Shiri Project AIDS Care 2010 22 1 81 88 10.1080/09540120903012577 20390484 10. Fylkesness K Siziya S A randomised trial on acceptability of voluntary HIV counselling and testing Trop Med Int Health 2004 9 5 566 572 10.1111/j.1365-3156.2004.01231.x 15117300 11. Namakhoma I Bongololo G Bello G Nyirenda L Phoya A Phiri S Negotiating multiple barriers: health workers’ access to counselling, testing and treatment in Malawi AIDS Care 2010 22 Suppl 1 68 76 10.1080/09540121003607241 20680862 12. Day JH Miyamura K Grant AD Leeuw A Munsamy J Baggaley R Attitudes to HIV voluntary counselling and testing among mineworkers in South Africa: Will availability of antiretroviral therapy encourage testing? AIDS Care 2003 15 5 665 72 10.1080/0954012030001595140 12959817 13. McPhail CL Pettifor A Coates T Rees H “You must do the test to know your status”: Attitude to HIV voluntary counselling and testing for adolescents among South Africa Youth and Parents Health Educ Behav 2008 35 1 87 104 10.1177/1090198106286442 16870815 14. Angotti N Bula A Gaydosh L Kimchi EZ Thornton RL Yeatman SE Increasing the acceptability of HIV counselling and testing with three C’s: Convenience, confidentiality and credibility Soc Sci Med 2009 68 2263 2270 10.1016/j.socscimed.2009.02.041 19375208 15. Urassa P Gosling R Pool R Reyburn H Attitudes to voluntary counselling and testing prior to the offer of nevirapine to prevent vertical transmission of HIV in northern Tanzania AIDS Care 2005 17 7 842 52 10.1080/09540120500038231 16120501 16. Taegtmeyer M Kilonzo N Mung’ala L Morgan G Theobald S Using gender analysis to build voluntary counselling and testing responses in Kenya Trans R Soc Trop Med Hyg 2006 100 4 305 11 10.1016/j.trstmh.2005.06.001 16214191 17. Mabunda G Voluntary counselling and testing: Knowledge and practices in a rural South African village J Transcult Nurs 2006 17 1 23 29 10.1177/1043659605281978 16410433 18. Central Statistical Office (CSO), Ministry of Health (MOH) and ICF International Zambia Demographic and Health Survey 2013-2014 2014 Rockville CSO, Ministry of Health, and ICF International 19. Ministry of Health/National AIDS Council Zambia Country Report: Multi-sectoral AIDS response monitoring and evaluation biennial report 2008-2009. Submitted to the United Nations General Assembly Special Session on AIDS 2010 Lusaka MoH/NAC 20. Dunkle KL Stephenson R Karita E Chomba E Kayitenkore K Vwalika C New heterosexually transmitted HIV infections in married or cohabiting couples in urban Zambia and Rwanda: an analysis of survey and clinical data Lancet 2008 371 9631 2183 2191 10.1016/S0140-6736(08)60953-8 18586173 21. Roth DL Stewart KE Clay OJ van der Straten A Karita E Allen S Sexual practices of HIV discordant and concordant couples in Rwanda: Effects of a testing and counselling programme for men Int J STD AIDS 2001 12 181 188 10.1258/0956462011916992 11231872 22. Allen S Meizen-Derr J Kautzman M Zulu I Trask S Fideli U Sexual behaviour of HIV discordant couples after HIV counselling and testing AIDS 2003 17 5 733 740 10.1097/00002030-200303280-00012 12646797 23. Farquhar C Kiarie JN Richardson BA Kabura MN John FN Nduati RW Antenatal couple counselling increases uptake of interventions to prevent HIV-1 transmission J Acquir Immune Defic Syndr 2004 37 1620 1626 10.1097/00126334-200412150-00016 15577420 24. Conkling M Shutes EL Karita E Chomba E Tichacek A Sinkala M Couples’ voluntary counselling and testing and nevirapine use in antenatal clinics in two African capitals: a prospective cohort study J Int AIDS Soc 2010 13 10 10.1186/1758-2652-13-10 20230628 25. Musheke M Bond V Merten S A systematic review of qualitative findings on factors enabling and deterring uptake of HIV testing in Sub-Saharan Africa BMC Public Health 2013 13 220 10.1186/1471-2458-13-220 23497196 26. Graneheim UH Lundman B Qualitative content analysis in nursing research: Concepts, procedures and measures to achieve trustworthiness Nurse Educ Today 2004 24 105 112 10.1016/j.nedt.2003.10.001 14769454 27. Burnard P A method of analysing interview transcripts in qualitative research Nurse Educ Today 1991 11 461 466 10.1016/0260-6917(91)90009-Y 1775125 28. Ayres L Kavanaugh K Knafl KA Within-case and across-case approaches to qualitative data analysis Qual Health Res 2003 13 871 883 10.1177/1049732303013006008 12891720 29. Overcash JA Narrative research: A viable methodology for clinical nursing Nurs Forum 2004 39 1 15 22 10.1111/j.0029-6473.2004.00015.x 15098321 30. Larsson EC Thorson A Nsabagasani X Namusoko S Popenoe R Ekström AM Mistrust in marriage-Reasons why men do not accept couple HIV testing during antenatal care-a qualitative study in eastern Uganda BMC Public Health 2010 10 769 10.1186/1471-2458-10-769 21167040 31. Joint United Nations Programme on HIV/AIDS: Report on the Global AIDS Epidemic. Geneva: UNAIDS; 2010. Available at: www.unaids.org/documents/20101123_globalreport_em.pdf. Accessed on April 2 2013. 32. Siu GE Wight D Seeley JA Masculinity, social context and HIV testing: an ethnographic study of men in Busia district, rural eastern Uganda BMC Public Health 2014 14 33 10.1186/1471-2458-14-33 24417763 33. Skovdal M Campbell C Madanhire C Mupambireyi Z Nyamukapa C Greyson S Masculinity as a barrier to men’s use of HIV services in Zimbabwe Global Health 2011 7 1 13 10.1186/1744-8603-7-13 21575149 34. Nyamhanga TM Muhondwa EPY Shayo R Masculine attitudes of superiority deter men from accessing antiretroviral therapy in Dar es Salaam Glob Health Action 2013 6 21812 24152373 35. Siu GE Seeley J Wight D Dividuality, masculine respectability and reputation: How masculinity affects men’s uptake of HIV treatment in rural eastern Uganda Soc Sci Med 2013 89 45 52 10.1016/j.socscimed.2013.04.025 23726215 36. Pool R Nyanzi S Whitworth JAG Attitude to voluntary counselling and testing for HIV among pregnant women in rural south-west Uganda AIDS Care 2001 13 5 605 15 10.1080/09540120120063232 11571007 37. Skovdal M Campbell C Nyamukapa C Greyson S When masculinity interferes with women’s treatment of HIV infection: a qualitative study about adherence to antiretroviral therapy in Zimbabwe JIAS 2011 14 29 10.1186/1758-2652-14-29 38. Simpson A Christian identity and men’s attitudes to antiretroviral therapy in Zambia Afr J AIDS Res 2010 9 4 397 405 10.2989/16085906.2010.545650 25875888 39. Roura M Urassa M Busza J Mbata D Wringe A Zaba B Scaling up stigma? The effects of antiretroviral roll-out on stigma and HIV testing. Early evidence from rural Tanzania Sex Transm Infect 2009 85 4 308 12 10.1136/sti.2008.033183 19036776 40. UNAIDS welcomes new guidelines which give additional HIV prevention option to discordant couples. UNAIDS, April 2012. Available at: http://www.unaids.org/sites/default/files/web_story/20120419_PS_coupletesting_en_0.pdf. Accessed 19 Aug 2016. 41. Musheke M Bond V Merten S Couple experiences of provider-initiated couple HIV testing in an antenatal clinic in Lusaka, Zambia: Lessons for policy and practice BMC Health Serv Res 2013 13 97 10.1186/1472-6963-13-97 23496926 42. Bwambale FM Ssali SN Byaruhanga S Kalyango JN Karamagi CAS Voluntary HIV counselling and testing among men in rural western Uganda: Implications for HIV prevention BMC Public Health 2008 8 263 10.1186/1471-2458-8-263 18664301 43. Stringer EM Sinkala M Kumwenda R Chapman V Mwale A Vermund SH Personal risk perception, HIV Knowledge and risk avoidance behaviour and their relationships to actual HIV sero-status in an urban African obstetric population J Acquir Immune Defic Syndr 2004 35 60 66 10.1097/00126334-200401010-00009 14707794 44. Bignami-Van AS Chao LW Anglewicz P Chilongozi D Bula A The validity of self-reported likelihood of HIV infection among the general population in rural Malawi Sex Transm Infect 2006 83 35 40 10.1136/sti.2006.020545 16790561 45. Anglewicz P Kohler HP Over-estimating HIV infection: The construction and accuracy of subjective probabilities of HIV infection in rural Malawi Demogr Research 2009 20 6 65 96 10.4054/DemRes.2009.20.6 46. Fagbamigbe AF Akinyemi JO Adedokun BO Bamgboye EA Gender variation in self-reported likelihood of HIV infection in comparison with HIV test results in rural and urban Nigeria AIDS Res Ther 2011 8 44 10.1186/1742-6405-8-44 22185294 47. Gold RS Hinchy J Batrouney CG The reasons behind decisions not to take up antiretroviral therapy among Australians infected with HIV Int J STDs AIDS 2000 11 6 361 370 10.1258/0956462001916065 48. Gold RS Ridge DT “I will start treatment when I think time is right”: HIV positive gay men talk about their decision not to access antiretroviral therapy AIDS Care 2001 13 6 693 708 10.1080/09540120120076869 11720640 49. Gellaitry G Cooper V Davies C Fisher M Leake M Patients’ perception of information about HAART: Impact on treatment decisions AIDS Care 2005 17 3 367 376 10.1080/09540120512331314367 15832885 50. Alfonso V Bermbach N Geller J Montaner JSG Individual variability in barriers affecting people’s decision to take HAART: A qualitative study identifying barriers to being on HAART AIDS Patient Care STDS 2006 20 848 857 10.1089/apc.2006.20.848 17192150 51. Fox MP Mazimba A Seidenberg P Crooks D Sikateyo B Rosen S Barriers to initiation of antiretroviral treatment in rural and urban Zambia: A cross-sectional study of cost, stigma and perception of treatment J Int AIDS Soc 2010 13 8 10.1186/1758-2652-13-8 20205930 52. Musheke M Bond V Merten S Deterrents to HIV-patient initiation of antiretroviral therapy in urban Lusaka, Zambia: A qualitative study AIDS Patient Care STDS 2013 27 4 231 241 10.1089/apc.2012.0341 23530573 53. Warwick Z The influence of antiretroviral therapy on the uptake of HIV testing in Tutume, Botswana Int J STD AIDS 2006 17 7 479 81 10.1258/095646206777689189 16820079 54. Phakathi Z Van Rooyen H Fritz K Richter L The influence of antiretroviral treatment on willingness to test: A qualitative study in rural KwaZulu-Natal, South Africa Afr J AIDS Res 2011 10 2 173 180 10.2989/16085906.2011.593381 25859740 55. Musheke M Bond V Merten S Individual and contextual factors influencing patient attrition from antiretroviral therapy care in an urban community of Lusaka, Zambia JIAS 2012 15 Suppl 1 17366 56. Schumaker LL Bond VA Antiretroviral therapy in Zambia: Colours, ‘spoiling’, ‘talk’ and the meaning of antiretrovirals Soc Sci Med 2008 67 2126 2134 10.1016/j.socscimed.2008.09.006 18952338 57. Ministry of Health Zambia Consolidated Guidelines for Treatment and Prevention of HIV Infection 2013 Lusaka MoH 58. WHO Rapid Advice: Antiretroviral therapy for HIV infection in adults and adolescents 2009 Geneva WHO 59. Oshi SN Ezugwu FO Oshi DC Dimkpa U Korie FC Okperi BO Does self-perception of risk of HIV infection make the youth to reduce risky behaviour and seek voluntary counselling and testing services? A case study of Nigerian Youth J Soc Sci 2007 14 2 195 203 60. Granich RM Gilks CF Dye C De Cock KM Williams BG Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: A mathematical model Lancet 2009 373 48 57 10.1016/S0140-6736(08)61697-9 19038438 61. Råssjö EB Darj E Konde-Lule J Olsson P Responses to VCT for HIV among young people in Kampala Uganda Afr J AIDS Res 2007 6 3 215 222 10.2989/16085900709490417 25866167 62. Obermeyer CM Sankara A Bastien V Parsons M Gender and HIV testing in Burkina Faso: An exploratory study Soc Sci Med 2009 69 6 877 84 10.1016/j.socscimed.2009.07.003 19631435 63. National AIDS Council Zambia country report monitoring the declaration of commitment on HIV and AIDS and the universal access 2015 Lusaka NAC

PMC005xxxxxx/PMC5000426.txt

==== Front BMC Evol BiolBMC Evol. BiolBMC Evolutionary Biology1471-2148BioMed Central London 73610.1186/s12862-016-0736-7Research ArticleSocially cued developmental plasticity in web-building spiders http://orcid.org/0000-0002-6334-9934Neumann Rainer (049)4042838 7894epeira@web.de Schneider Jutta M. (049)4042838 3878jutta.schneider@uni-hamburg.de Zoologisches Institut, Universität Hamburg, Martin-Luther-King-Platz 3, 20146 Hamburg, Germany 26 8 2016 26 8 2016 2016 16 1 17029 2 2016 8 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Socially cued anticipatory plasticity (SCAP) has been proposed as a widespread mechanism of adaptive life-history shifts in semelparous species with extreme male mating investment. Such mating systems evolved several times independently in spiders and male reproductive success should critically depend on timely maturation and rapid location of a receptive and, ideally, virgin female. We experimentally investigated socially cued anticipatory plasticity in two sympatric, closely related Nephila species that share many components of their mating systems, but differ in the degree to which male reproductive success depends on mating with virgin females. Juveniles of both species were reared either in the presence or absence of virgin female silk cues. We predicted strong selection on socially cued plasticity in N. fenestrata in which males follow a highly specialized terminal investment strategy, but expected a weaker plastic response in N. senegalensis in which males lost the ability to monopolize females. Results Contrary to our predictions, N. fenestrata males presented with virgin female silk cues did not mature earlier than siblings reared isolated from such cues. Males in N. senegalensis, however, showed a significant response to female cues and matured several days earlier than control males. Plastic adjustment of maturation had no effect on male size. Conclusions Our results indicate that a strong benefit of mating with virgins due to first male sperm priority does not necessarily promote socially cued anticipatory plasticity. We emphasize the bidirectional mode of developmental responses and suggest that this form of plasticity may not only yield benefits through accelerated maturation, but also by avoiding costs of precipitate maturation in the absence of female cues. Keywords Adaptive plasticityEnvironmental uncertaintyDensity cuesInformation useMale-male competitionNephilidaeSSDhttp://dx.doi.org/10.13039/501100001659Deutsche ForschungsgemeinschaftSCHN561/9-1Schneider Jutta M. issue-copyright-statement© The Author(s) 2016 ==== Body Background In most organisms, genetically identical individuals develop markedly different phenotypes when exposed to different environments [1–6] and such plastic modifications of morphology, physiology, life-history or behavior have been frequently shown to be adaptive, yielding increased fitness returns under specific conditions [7–9]. Juvenile development, maturation, and the period of reproduction in many animal species follow recurrent seasonal gradients [10, 11], thus it is crucial to adjust one’s own reproductive period to the opposite sex, particularly in semelparous species experiencing only a single reproductive episode. Often thermal threshold values [12] or photoperiod changes are used as indicators of large-scale seasonal progression [13]. However, fluctuations of external conditions may alter the density and structure of a population [14–16] and sex-specific differences in developmental rates or mortality will further add temporal demographic variation. Such local differences are difficult to predict from large scale cues and plastic adjustment of life-history traits based on local information might be advantageous [17]. Recent studies have highlighted the role of social cues in adaptive life-history shifts, for example, in response to the density of conspecifics [18, 19]. Since accelerated or delayed juvenile development in response to conspecific cues precedes its fitness-relevant effect at the stage of maturity, these mechanisms have been termed ‘socially cued anticipatory plasticity’ (SCAP; [20]). Socially cued developmental tactics are hypothesized to be more common than currently appreciated [20], but plasticity can also involve fitness costs [21, 22] that may constrain the evolution of such traits. Moreover, an established cue from the social environment that benefits plastically responding individuals in a single species may not be similarly relevant in related species, as the value of particular cues will strongly relate to specific features of the mating system under study [20]. Many spiders show extreme reversed sexual size dimorphism (SSD; [23, 24]) and specialized male strategies to maximize and protect paternity, such as self-sacrifice [25–27], genital plugging [28–30] or remote copulation [31]. In such species, males generally benefit most from locating a virgin female [32–34], which may impose selection on males to mature earlier. The transition to the reproductive stage, however, is critical because maturing males lose the ability to capture prey on their own. Since adult males are restricted to feeding opportunistically on the females’ prey [35, 36], precipitate maturation may also be unfavorable. Spider males perceive conspecific females using one or more sensory modalities, but most species lack acute vision and have to rely on mechanical or chemical signals that indicate the presence of receptive females [37, 38]. Female sex pheromones, both volatile and incorporated in female silk, have been shown to serve this function in orb-web spiders [39–41], but only one study on the Australian red-back spider Latrodectus hasselti provides evidence that female pheromones induce adaptive developmental plasticity in males [42]. Like most animals [43], males in this species have to trade-off developmental time against growth, as fast-maturing males stay relatively small, but intense contest competition shifts fitness payoffs to larger males [44]. Specialized male mating strategies that allow maximizing paternity with a single or very few females and constitute a very high mating effort have evolved at least four times independently in different spider families [45, 46]; providing ideal model systems to investigate whether such characteristics generally promote socially cued anticipatory plasticity. Differences in specific mating traits may affect selection on such mechanisms and a comparative approach may help to relate the magnitude of plastic responses to the associated adaptive value. The golden-silk spider genus Nephila is an established model system in sexual selection research, which has been used to study, for example, female-biased SSD [24], sexually selected life-history traits [47, 48], male-male competition [49–52], and intrasexual size variation [53, 54]. Especially male size varies greatly in some species [55, 56]. Socially cued developmental plasticity may increase male size variation; even more so as the capacity for plastic modifications may differ between genotypes [9]. Nephila females build large orb-webs, whereas males cease web-building after reaching maturity to search for females [57]. Population densities change in the course of the season [58–60], but in addition, local environmental conditions cause strong between-year variation in some species [61]. Socially cued plasticity could serve to adjust male development to female availability and male-male competition, but experimental work is required to determine whether predefined cues induce the expected developmental modifications [20]. We examined the capacity of males to optimize the timing of maturation in response to female silk cues using two sympatric Nephila species that are exposed to almost identical abiotic cues of seasonal changes in their natural habitat, N. fenestrata and N. senegalensis. Both species are generally similar in their reproductive biology, but differ in certain aspects of male mating strategies. N. fenestrata males follow a terminal investment strategy aimed at monopolizing a single female by means of mate plugging through copulatory organ breakage [28], whereas in N. senegalensis, males do not produce mating plugs and each male is able to fertilize up to four females [62]. Males in this species adopt flexible mating tactics including male mate choice and polygyny, which reduce the imbalance in reproductive success between males that encounter a virgin or a non-virgin female first [55]. Hence, although males in both species prefer mating with virgins [55, 63, 64], life-time fitness in N. fenestrata more strongly depends on locating an unmated female. These differences affect the value of prospective mates and are expected to generate dissimilar selection on socially cued anticipatory plasticity; an assumption in line with a field study on two other orb-web spider species suggesting anticipatory plastic responses to female densities in the monogynous N. plumipes, but not in Argiope keyserlingi, in which males are usually bigynous [17]. We reared juvenile N. fenestrata and N. senegalensis under standardized conditions in climate-control chambers, presenting spiders in the experimental treatment with virgin female silk. We expected the highly specialized, terminally investing N. fenestrata males to accelerate their development in the presence of virgin female silk cues. Males were supposed to mature earlier, but at smaller size, than siblings in the control treatment without virgin female cues. In N. senegalensis, males are often polygynous and depend less on locating a virgin female, hence we predicted a weaker developmental response in this species. Methods Study animals Spiders used in this study were F2 offspring descending from females that were collected at Mawana Game Reserve, Zululand District, KwaZulu-Natal, South Africa in 2012 (permit OP 990/2012 from EZEMVELO KZN WILDLIFE PERMITS OFFICE). All families of study animals were derived from mating virgin individuals from different maternal lines. Six family lineages were used in each species comprising 23.7 ± 5.6 individuals per family in N. senegalensis and 25.5 ± 5.4 individuals per family in N. fenestrata. We reared hatchlings communally at first, and separated them after approximately two additional molts to maintain them in 200 ml plastic cups, which were turned upside down. Spiders were kept under standardized conditions in our main laboratory [63] before being transferred to the experimental rooms. The experiments took place at the Zoological Institute, University of Hamburg, between May 23 and August 25, 2013. Experimental setup and treatments Spiders were kept in two climate-control chambers, measuring approximately 1.9 m × 4.3 m × 2.4 m each, in order to control temperature, relative humidity, light-dark cycle, and light intensity. Climate-control chambers (Weiss Umwelttechnik GmbH, model type WK 21’/5–40) featured identical technical specifications. Both devices were contemporaneously installed, calibrated and put into operation by the manufacturer’s expert staff in 2012. For each of our study species, we established an experimental treatment in which adult virgin females’ silk was introduced to the spiders’ rearing cups (referred to as the Female cues treatment). Thereby, we presented the study animals with potential contact pheromones or any properties of silk that may indicate the presence of adult females. In a control treatment, spiders were reared isolated from adult virgin female silk cues (referred to as the No cues treatment). To exclude long distance perception of female silk cues in the control treatment, we arranged the Female cues treatments for both of our study species simultaneously in one climate chamber and used the second chamber for both No cues treatments. Each climate chamber was equipped with six bottom shelves and six top shelves. A tubular fluorescent daylight lamp was mounted above each shelf with a distance of 60 cm. We placed up to twenty-seven rearing cups on each shelf with an equal distance of approximately 15 cm between cups. Prior to the transfer of study animals to the experimental treatments, we adjusted climate-control chambers to provide identical conditions of temperature, relative humidity, and light regime. Temperature and humidity were regulated corresponding to periods of artificial daytime and night-time throughout the experiment; i.e., temperature was set to 26 °C during lighting periods and 21 °C during dark periods. We set daytime humidity to 50 % and night-time humidity to 70 %, respectively. These conditions fit well within the range in both species’ habitats. In the beginning of the experiment, we used a 14:12 h light-dark cycle and reduced the daily lighting duration by 10 min each week to simulate a decrease in day length, which both of our study species experience during summer and autumn in their habitats of origin. Transfer of study animals to climate-control chambers Nephila fenestrata study animals were transferred to the climate chambers on May 23; N. senegalensis were transferred on May 26/27. We used a split brood design and allocated equal numbers of randomly chosen individuals from each family lineage to each treatment. After the transfer had been completed, we checked all study animals for presence and condition on the following day and replaced a small number of spiders that had died or vanished from the rearing cups. No study animals were replaced at a later date. Maintenance and monitoring schedule The regular monitoring of study animals began on May 29 (defined as the start of the experiment) with the following numbers of study animals: N. fenestrata: Female cues treatment: n = 156; No cues treatment: n = 157; N. senegalensis: Female cues treatment: n = 162; No cues treatment: n = 162. Spiders were fed Drosophila flies twice a week on a regular schedule. In the initial stage of the experiment when the spiders were still very small, we used flies that had been killed at −80 °C. When all spiders had reached a minimum body length of approximately 5 mm, we supplemented the diet with live insects. This food supply allowed the spiders ad libitum feeding. Water was offered on 6 days per week. At this stage, we checked the animals’ condition four times a week and recorded any cases of death as well as spiders that had vanished from their rearing cups (missing spiders likely dropped from rearing cups during feeding or cleaning of shelves). Introduction of female silk cues As a consequence of female-biased SSD, Nephila females take longer to mature than males, so that early maturing males become adults in populations devoid of adult females (protandry; [57, 60]). As our goal in this study was to simulate the beginning of the mating season, we presented males with adult virgin female cues not from the start, but after a period of development in the absence of such cues. In the Female cues treatment, we introduced the first set of silk cues to the rearing cups on days 22/23 from the start of the experiment for N. fenestrata and on days 22–24 for N. senegalensis (all subsequent sets of silk cues were introduced within one day). We used plastic expansion bolts to present silk samples to the study animals. For this purpose, the expansions of each piece were spread, resulting in a Y-shaped object, which we put up in vertical position using a base of potting clay. These silk fixtures measured 5.5 cm in height. For acquiring silk cues, we used female webs the spiders had built into 40 cm × 40 cm-sized Perspex frames. Webs had usually been newly built in the previous night, but were at most two days old. Females were removed from their webs and the frames were taken to the female cues chamber. We then twisted a few silk threads from the web’s moistened catching spiral around the upper expansions of each silk fixture and used fine scissors to dissect the threads from the web. One silk fixture was placed under each rearing cup, so that the spider inside could easily access the silk threads, especially with its pedipalps and forelegs, bearing the most important sensory organs to perceive physical and chemical cues [65, 66]. Fresh silk cues were introduced on a weekly schedule (on days 29, 36, 43, 50, 57, 64, 71, 78, and 85 from the start of the experiment). On the previous day, we removed all silk fixtures from the rearing cups and cleaned the shelves in the experimental rooms. Each object was cleaned of silk with alcohol and air-dried prior to reuse. In order to standardize experimental conditions, we placed identical objects free of silk under the rearing cups in the No cues treatments. Silk cues were acquired from twenty-four adult virgin female N. fenestrata (up to four per turn) and thirty-three N. senegalensis (up to six per turn). Females originated from eleven family lineages in N. fenestrata and twelve family lineages in N. senegalensis. Average female adult age (days passed from date of maturity) at the time of web production was 13 days (range: 2–30 days) in N. fenestrata and 11.5 days (range: 2–29 days) in N. senegalensis. How many times a male received fresh silk cues depended on individual developmental durations. Those males in the Female cues treatments that were used in our analysis received fresh cues 5.4 ± 0.1 times in N. fenestrata and 6.4 ± 0.1 times in N. senegalensis (range in both species: 3-8 times). Individual silk cues were obtained from a female unrelated to the cues-receiving male (48 % of cues in N. fenestrata and 45 % of cues in N. senegalensis) or from a female that had one parental lineage in common with the cues-receiving male (52 % of cues in N. fenestrata and 53 % of cues in N. senegalensis). In < 1 % of cues in N. fenestrata and 2 % of cues in N. senegalensis, we could not avoid using silk from females that had both parental lineages in common with the cues-receiving male. No male received cues from related females only. With the first implementation of female silk cues, we adjusted the monitoring of study animals and checked the individual state of development on six days per week. For each male, we recorded the duration of development from the start of the experiment to maturity and the duration of the subadult instar (i.e., the last developmental stage; subadult males can easily be detected by the swollen palp tarsi indicating the ongoing transformation into copulatory organs). Juvenile females were immediately removed from the study when they were clearly discernible (body length ≥ approximately 12 mm, pedipalps unmodified). Statistical analyses We defined the start of the experiment as the first monitoring of study animals after being transferred to the climate-control chambers (May 29). In N. fenestrata, some males matured before the first introduction of female silk cues had been completed (June 21). These males were excluded from the analyses (predefined female cues chamber: n = 5; no cues chamber: n = 9). In each of our study species, we analyzed effects of our experimental treatment (Female cues/No female cues) on male development with separate linear mixed models performed in R 3.2.4 (R Development Core Team 2016). Dependent variables were (1) Duration of development from the start of the experiment, (2) Duration of subadult stage, (3) Adult size, and (4) Adult weight. The study animals’ family lineage was included as a random effect. We tested for statistical significance of Treatment using ANOVA model comparisons with χ2 tests between the full model and a model that had the variable removed. Using the same dependent variables, we conducted generalized linear models in JMP IN 7.0 (SAS Institute Inc., Carey, NC, USA) to test for an interaction between Treatment and Family lineage. Models were fitted with normal error structure and identity-link function. We removed the interaction term if it was non-significant (α = 0.05) while retaining both main effects in the final models. Developmental durations were log-transformed to improve model fit. Descriptive statistics are given as mean ± standard error. Within experiments, sample sizes may differ due to missing data. Results We performed linear mixed models to test effects of our experimental treatment on male development and growth. The models clearly revealed a significant influence of our treatment on the duration of development in N. senegalensis. Males in the Female cues treatment matured two to five days earlier, on average, than males in the No cues treatment and the mean duration of the subadult stage alone differed by one and a half to two days (ANOVA model comparisons: Duration of development from the start of the experiment: χ2 = 10.563, p = 0.001; Duration of subadult stage: χ2 = 29.724, p < 0.001; Table 1). However, shortened development did not translate into different male size or body mass (ANOVA model comparisons: Adult size: χ2 = 1.134, p = 0.287; Adult weight: χ2 = 2.586, p = 0.108; Table 1). Contrary to our predictions, in N. fenestrata, there were no significant differences in various life-history parameters between males presented with virgin female silk cues and those reared in the absence of such cues (ANOVA model comparisons: Duration of development from the start of the experiment: χ2 = 0.006, p = 0.939; Duration of subadult stage: χ2 = 1.632, p = 0.202, Adult size: χ2 = 0.528, p = 0.467; Adult weight: χ2 = 1.629, p = 0.202; Table 1).Table 1 Developmental parameters of male Nephila fenestrata and N. senegalensis reared in different experimental treatments N. fenestrata N. senegalensis No female cues Female cues n No female cues Female cues n Duration of development (start to maturity) [d] 57.36 ± 0.78 57.18 ± 0.86 153 68.83 ± 0.86 65.36 ± 0.63 142 Duration of subadult stage [d] 18.51 ± 0.18 18.26 ± 0.15 153 21.12 ± 0.25 19.38 ± 0.21 142 Adult size/patella-tibia [mm] 5.69 ± 0.09 5.65 ± 0.09 150 4.74 ± 0.09 4.62 ± 0.1 132 Adult weight [mg] 18 ± 0.58 18.22 ± 0.56 153 22.46 ± 0.66 21.06 ± 0.7 141 We ran additional generalized linear models to analyze potential family-specific variation of developmental plasticity. In N. senegalensis, the response toward a shortened development was present in all family lineages (Fig. 1). The interaction between Family lineage and Treatment, however, was always found to be non-significant at the 5 % level; although developmental responses varied considerably between families (Fig. 1, Table 2). Corroborating mixed model results, the GLMs showed that in N. fenestrata, only Family lineage predicted developmental durations, size, and weight, while Treatment had no effect (Table 2). In contrast, both Family lineage and Treatment significantly determined developmental durations in N. senegalensis (Table 2).Fig. 1 Duration of the subadult instar (i.e., the last developmental stage preceding maturity) in the presence or absence of virgin female silk cues compared between male Nephila senegalensis and N. fenestrata. Graphs illustrate mean developmental durations according to family lineages Table 2 Effects of family lineage and treatment on developmental parameters in Nephila fenestrata and N. senegalensis Explanatory variable Duration of development (start to maturity) Duration of subadult stage Adult size Adult weight χ 2 p df χ 2 p df χ 2 p df χ 2 p df N. fenestrata  Family lineage 50.65 <.0001 5 14.6 0.012 5 67.07 <.0001 5 61.86 <.0001 5  Treatment 0.002 0.961 1 2.08 0.15 1 0.73 0.391 1 2 0.158 1  Family lineage  *Treatment 2.101 0.835 5 2.332 0.807 5 1.588 0.903 5 1.704 0.888 5 N. senegalensis  Family lineage 21.89 0.0006 5 22.23 0.0005 5 37.56 <.0001 5 40.77 <.0001 5  Treatment 11.34 0.0008 1 31.01 <.0001 1 1.16 0.281 1 2.621 0.105 1  Family lineage  *Treatment 10.126 0.072 5 7.924 0.161 5 10.58 0.06 5 9.757 0.082 5 Likelihood-ratio tests and corresponding p-values derive from generalized linear models performed in JMP IN 7.0 (SAS Institute Inc., Carey, NC, USA). Non-significant interaction terms were removed from the final models. Developmental durations were log-transformed. Significant p-values are shown in bold Discussion Males in one of our study species, Nephila senegalensis, plastically adjusted development and matured significantly earlier in response to female silk cues than those reared isolated from such cues. However, we found no developmental response in N. fenestrata. While the plastic adjustment of maturation in N. senegalensis is in accordance with our predictions, we expected an even more distinct modification of development in the monogynous N. fenestrata males whose fitness strongly depend on locating a virgin female [28]. The absence of a plastic response in this species indicates that socially cued anticipatory plasticity is not a universal feature in species with strong developmental differences between the sexes and a high male mating effort. The plastic adjustment of maturation in N. senegalensis is best described as flexibility in the duration of the subadult instar and did not affect male adult size or mass. Males in this species are able to fertilize multiple females [62, 63] and differential mating investment has been identified as an integral part of a flexible mating strategy in this species [55]. In nature, individual males visit up to four females (Neumann & Schneider, unpublished observations); hence adjusted maturation in response to the perception of female cues may increase a male’s chance to locate a virgin female first, and to mate with further females in a period of low or moderate competitive conditions. Animals in general have to trade-off developmental duration against growth [43]. Increased food intake and delayed maturation will usually result in larger adult body size, which often is an important determinant of male reproductive success in mating systems involving contest competition [67]. This relationship also exists in some web-building spiders [44] and we expected males perceiving the presence of virgin females to mature at smaller size as a consequence of accelerated development. In contrast to a previous study on Australian red-back spiders [42], however, adjustment of maturation in N. senegalensis was not achieved by substantially abbreviating development. Rather, the timing of maturation was modified by differences in the duration of the subadult instar and there was no trade-off between adjustment of development and adult body size. Hence, we found no support for socially cued plasticity to contribute to the extreme male size variation observed in many Nephila species [47–49, 55]. While males adjusted the timing of maturation in the same direction across family lineages, we also observed considerable variation between lineages regarding the magnitude of plastic responses. Genotype-specific degrees of plasticity in response to an environmental trigger could contribute to phenotypic variation, but our study found little evidence for such interrelations in our model species. It is important to realize the bidirectional mode of a plastic response; hence not only the expression of a specific modification appropriate to requirements should be beneficial, but also the non-expression of the same modification in the absence of the corresponding trigger. What is to be gained from staying subadult for a male N. senegalensis in the absence of adult females? With sexual maturation, web-spider males undergo drastic changes in terms of morphology, physiology and life-style, solely targeted on reproduction [35]. Adult male spiders lose weight during mate search [68] but are no longer able to build capture webs [51, 57]. In order to maintain a sound physical condition, they depend on stealing prey from female webs [48, 57, 69]. Males maturing without the perspective of locating a female in a short time risk declining physical strength, whereas subadult males residing in their own webs stay relatively safe from predation and may continue feeding on self-captured prey. Another potential benefit of a delayed maturation may relate to sperm-limitation, which is a universal trait in nephilid spider males [70]. Male N. senegalensis produce their lifetime sperm supply in their subadult instar and spermatogenesis is terminated prior to adulthood [62]. Total sperm numbers vary considerably among males [68] and a prolonged subadult instar may allow males to increase sperm quantity to prevail in sperm-competition. Taken together, these arguments support the assumption that N. senegalensis males significantly benefit from shifting maturation until mating is about to take place, and not to mature when the probability of finding a female is low. However, most of these arguments apply to N. fenestrata as well and the absence of a plastic response to virgin female cues in this species is puzzling. Owing to mate plugging and copulatory organ breakage, male mating tactics are less flexible than in N. senegalensis, and male reproductive success critically depends on the ability of monopolizing a single female [28, 32]. To explain our findings, we might consider between-species differences regarding the value of developmental responses from both male and female perspective. Such differences might be linked to our study species’ ecology, as habitat requirements differ slightly between both species, which could affect the predictability of receptive females. Habitats in N. senegalensis range from humid areas to bush savannahs and habitat heterogeneity is reflected in varying population densities (Neumann & Schneider, unpublished observations). Predicting female presence is therefore challenging and selection may favor male ability to fine-tune maturation on a local scale. In contrast, N. fenestrata occurs in forested areas (sub N. pilipes; [71]) providing relatively constant temperature and humidity, and females typically form dense aggregations in preferred sites (Penney, unpublished observations). Given the rather narrow range of tolerated conditions, female presence may directly be indicated by abiotic large-scale cues and habitat quality, making socially cued anticipatory plasticity less needed in N. fenestrata. Finally, the presence and absence of socially cued plastic responses in the respective species could be explained from the female perspective. We cannot unambiguously relate the developmental response in male N. senegalensis to silk-borne pheromones, as the physical properties of adult females’ silk alone could indicate their presence, but females in various web-building spiders use specific chemical signals to attract males and secure a timely mating [72–74]. N. senegalensis females are polyandrous [62] and may use pheromone signals to repeatedly attract males. In N. fenestrata, however, there may be little need for females to advertise their presence, as males may easily locate them in their spatially-limited habitats; and also because female mating rates are much lower compared to N. senegalensis. Pheromone production itself may be costly [75–77] and attracting unwanted males could even decrease female fitness, if there are no significant benefits to be gained from multiple matings [78–80]. Additional research should investigate whether female N. fenestrata produce sex pheromones strategically; e.g., only under a high risk of remaining unmated [81]. Conclusions Our results suggest that a strong benefit of mating with virgins due to first male sperm priority does not necessarily promote socially cued anticipatory plasticity. Benefits and costs of using and providing information may differ between the sexes. Even if males, in principle, would benefit from plastic life-history shifts, they may sensorially rely on information provided by females. In such cases, the evolution of plasticity may depend on whether females benefit from providing cues, and future studies should take the female perspective into account. In addition, we suggest that the adaptive value of socially cued anticipatory plasticity might not be limited to males that adaptively accelerate development to mature in time, but males that delay maturation in the absence of female cues might also benefit by avoiding potential costs of precipitate maturation. We thank Julia Becker for her tremendous help in rearing, separating, and moving of study animals; Antje Hundertmark for generously lending females from her lab stock for silk production; Annalena Ritter, Miki Cartus, Tomma Dirks, and Angelika Taebel-Hellwig for additional assistance in the lab; Albert Driescher, Michele Mignini, Onno Preik, and the volunteers at Mawana, South Africa, for invaluable support in acquiring spiders for our lab stocks; Yael Lubin, Jasmin Ruch and two anonymous referees for helpful comments on the manuscript. Funding This work was supported by the German Science Foundation (DFG) (SCHN561/9-1 to J.M.S.). Availability of data and materials The data sets supporting the results of this article are available in Dryad: doi:10.5061/dryad.7117c [82] Authors’ contributions JMS conceived the study and provided funding and infrastructure. JMS and RN designed the experiment and wrote the manuscript. RN conducted the experiment, analyzed the data and crafted the figure. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate This study complies with all Federal and State laws of Germany. An animal ethics approval is not required. ==== Refs References 1. Smith-Gill SJ Developmental plasticity: developmental conversion versus phenotypic modulation Am Zool 1983 23 1 47 55 10.1093/icb/23.1.47 2. Pigliucci M Müller G Pigliucci M Phenotypic plasticity Evolution: The Extended Synthesis 2009 Cambridge MIT Press 355 378 3. Bradshaw AD Evolutionary significance of phenotypic plasticity in plants Adv Genet 1965 13 1 115 155 4. Liefting M Hoffmann AA Ellers J Plasticity versus environmental canalization: population differences in thermal responses along a latitudinal gradient in Drosophila serrata Evolution 2009 63 8 1954 1963 10.1111/j.1558-5646.2009.00683.x 19473402 5. Stearns SC. The role of develoment in the evolution of life histories. Life Sciences Research Report. 1982. p. 237–58. 6. Schmalhausen II Factors of Evolution 1949 7. Whitman DW, Agrawal AA. What is phenotypic plasticity and why is it important. Enfield: Phenotypic plasticity of insects; 2009. p. 1–63. 8. Stearns SC The Evolutionary Significance of Phenotypic Plasticity - Phenotypic sources of variation among organnisms can be described by developmental switches and reaction norms Bioscience 1989 39 7 436 445 10.2307/1311135 9. West-Eberhard MJ. Developmental plasticity and evolution. New York: Oxford University Press; 2003. 10. Helm B, Ben-Shlomo R, Sheriff MJ, Hut RA, Foster R, Barnes BM, Dominoni D. Annual rhythms that underlie phenology: biological time-keeping meets environmental change. Proc R Soc B Biol Sci. 2013;280(1765):20130016. doi:10.1098/rspb.2013.0016. 11. Foster RG Kreitzman L Seasons of Life - The biological rhythms that enable living things to thrive and survive 2009 New Haven Yale University Press 12. Caro SP Schaper SV Hut RA Ball GF Visser ME The Case of the Missing Mechanism: How Does Temperature Influence Seasonal Timing in Endotherms? PLoS Biol 2013 11 4 1 8 10.1371/journal.pbio.1001517 13. Hut RA, Paolucci S, Dor R, Kyriacou CP, Daan S. Latitudinal clines: an evolutionary view on biological rhythms. Proc R Soc B Biol. 2013;280(1765):20130433. doi:10.1098/rspb.2013.0433. 14. Angelstam P Lindstrom E Widen P Synchronous short-term population fluctuations of some birds and mammals in Fennoscandia - occurence and distribution Holarct Ecol 1985 8 4 285 298 15. Owen-Smith N Demography of a large herbivore, the Greater Kudu, Tragelaphus strepsiceros , in relation to rainfall J Anim Ecol 1990 59 3 893 913 10.2307/5021 16. Saether BE Environmental stochasticity and population dynamics of large herbivores: A search for mechanisms Trends Ecol Evol 1997 12 4 143 149 10.1016/S0169-5347(96)10068-9 21238011 17. Kasumovic MM Bruce MJ Herberstein ME Andrade MCB Evidence for developmental plasticity in response to demographic variation in nature Ecology 2009 90 8 2287 2296 10.1890/08-1540.1 19739390 18. Kasumovic MM Hall MD Try H Brooks RC The importance of listening: juvenile allocation shifts in response to acoustic cues of the social environment J Evol Biol 2011 24 6 1325 1334 10.1111/j.1420-9101.2011.02267.x 21457172 19. Bailey NW Gray B Zuk M Acoustic Experience Shapes Alternative Mating Tactics and Reproductive Investment in Male Field Crickets Curr Biol 2010 20 9 845 849 10.1016/j.cub.2010.02.063 20417103 20. Kasumovic MM Brooks RC It's All Who You Know: The Evolution Of Socially Cued Anticipatory Plasticity as A Mating Strategy Q Rev Biol 2011 86 3 181 197 10.1086/661119 21954701 21. Relyea RA Costs of phenotypic plasticity Am Nat 2002 159 3 272 282 10.1086/338540 18707379 22. DeWitt TJ Sih A Wilson DS Costs and limits of phenotypic plasticity Trends Ecol Evol 1998 13 2 77 81 10.1016/S0169-5347(97)01274-3 21238209 23. Foellmer MW Fairbairn DJ Competing dwarf males: sexual selection in an orb-weaving spider J Evol Biol 2005 18 3 629 641 10.1111/j.1420-9101.2005.00868.x 15842492 24. Kuntner M, Elgar MA. Evolution and maintenance of sexual size dimorphism: aligning phylogenetic and experimental evidence. Front Ecology Evol. 2014;2(26). doi:10.3389/fevo.2014.00026. 25. Welke KW, Schneider JM. Males of the orb-web spider Argiope bruennichi sacrifice themselves to unrelated females. Biol Lett. 2010;6(5):585–8. 26. Andrade MCB Sexual selection for male sacrifice in the Australian redback spider Science 1996 271 5245 70 72 10.1126/science.271.5245.70 27. Foellmer MW Fairbairn DJ Spontaneous male death during copulation in an orb-weaving spider Proc R Soc B Biol Sci 2003 270 S183 S185 10.1098/rsbl.2003.0042 28. Fromhage L Schneider JM Emasculation to plug up females: the significance of pedipalp damage in Nephila fenestrata Behav Ecol 2006 17 353 357 10.1093/beheco/arj037 29. Snow LSE Abdel-Mesih A Andrade MCB Broken copulatory organs are low-cost adaptations to sperm competition in redback spiders Ethology 2006 112 4 379 389 10.1111/j.1439-0310.2006.01163.x 30. Nessler SH Uhl G Schneider JM Genital damage in the orb-web spider Argiope bruennichi (Araneae: Araneidae) increases paternity success Behav Ecol Sociobiol 2007 18 174 181 10.1093/beheco/arl074 31. Li DQ Oh J Kralj-Fiser S Kuntner M Remote copulation: male adaptation to female cannibalism Biol Lett 2012 8 4 512 515 10.1098/rsbl.2011.1202 22298805 32. Fromhage L Elgar MA Schneider JM Faithful without care: the evolution of monogyny Evolution 2005 59 7 1400 1405 10.1111/j.0014-3820.2005.tb01790.x 16153026 33. Gaskett AC Herberstein ME Downes BJ Elgar MA Changes in male mate choice in a sexually cannibalistic orb-web spider (Araneae: Araneidae) Behaviour 2004 141 1197 1210 10.1163/1568539042729676 34. Huber BA Sexual selection research on spiders: progress and biases Biol Rev 2005 80 3 363 385 10.1017/S1464793104006700 16094804 35. Foelix RF Biologie der Spinnen 1979 Stuttgart Thieme 36. Foellmer MW Fairbairn DJ Selection on male size, leg length and condition during mate search in a sexually highly dimorphic orb-weaving spider Oecologia 2005 142 4 653 662 10.1007/s00442-004-1756-3 15619098 37. Trabalon M, Bagneres AG. Contact recognition pheromones in spiders and scorpions. Cambridge: Insect Hydrocarbons: Biology, Biochemistry, and Chemical Ecology; 2010. p. 344–74. 38. Pollard SD Macnab AM Jackson RR Communication with chemicals: pheromones and spiders 1987 Berlin, Heidelberg Springer 39. Nessler SH Uhl G Schneider JM Scent of a Woman - The Effect of Female Presence on Sexual Cannibalism in an Orb-Weaving Spider (Araneae: Araneidae) Ethology 2009 115 7 633 640 10.1111/j.1439-0310.2009.01646.x 40. Schulte KF Uhl G Schneider JM Mate choice in males with one-shot genitalia: limited importance of female fecundity Anim Behav 2010 80 4 699 706 10.1016/j.anbehav.2010.07.005 41. Chinta SP Goller S Lux J Funke S Uhl G Schidz S The Sex Pheromone of the Wasp Spider Argiope bruennichi Angew Chem Int Ed 2010 49 11 2033 2036 10.1002/anie.200906311 42. Kasumovic MM Andrade MCB Male development tracks rapidly shifting sexual versus natural selection pressures Curr Biol 2006 16 7 R242 R243 10.1016/j.cub.2006.03.017 16581497 43. Stearns SC Life history evolution: successes, limitations, and prospects Naturwissenschaften 2000 87 11 476 486 10.1007/s001140050763 11151666 44. Kasumovic MM Andrade MCB A change in competitive context reverses sexual selection on male size J Evol Biol 2009 22 2 324 333 10.1111/j.1420-9101.2008.01648.x 19032500 45. Schneider JM Fromhage L Monogynous mating strategies in spiders Animal behaviour: evolution and mechanisms 2010 Berlin Springer 441 464 46. Schneider JM. Sexual cannibalism as a manifestation of sexual conflict. Cold Spring Harbor Perspect Biol. 2014;6(11):369–84. 47. Rittschof CC Hilber SA Tudor MS St Mary CM Modeling male reproductive strategies and optimal mate number in an orb-web spider Behav Ecol 2012 23 1 1 10 10.1093/beheco/arr142 48. Higgins L Coddington J Goodnight C Kuntner M Testing ecological and developmental hypotheses of mean and variation in adult size in nephilid orb-weaving spiders Evol Ecol 2011 25 6 1289 1306 10.1007/s10682-011-9475-9 49. Schneider JM Herberstein ME De Crespigny FC Ramamurthy S Elgar MA Sperm competition and small size advantage for males of the golden orb-web spider Nephila edulis J Evol Biol 2000 13 6 939 946 10.1046/j.1420-9101.2000.00238.x 50. Rittschof CC Male density affects large-male advantage in the golden silk spider, Nephila clavipes Behav Ecol 2010 21 5 979 985 10.1093/beheco/arq099 51. Christenson TE Goist KCJ Costs and benefits of male-male competition in the orb weaving spider, Nephila clavipes Behav Ecol Sociobiol 1979 5 87 92 10.1007/BF00302697 52. Fromhage L Schneider JM Virgin doves and mated hawks: contest behaviour in a spider Anim Behav 2005 70 1099 1104 10.1016/j.anbehav.2005.02.020 53. Vollrath F. Sex-linked differences in the growth of Nephila clavipes. Proceedings of the 17th European Colloquium of Arachnology, Edinburgh 1997. 1998. p. 161–6. 54. Schneider JM Elgar MA The combined effects of pre- and post-insemination sexual selection on extreme variation in male body size Evol Ecol 2005 19 5 419 433 10.1007/s10682-005-8310-6 55. Neumann R Schneider JM Differential investment and size-related mating strategies facilitate extreme size variation in contesting male spiders Anim Behav 2015 101 107 115 10.1016/j.anbehav.2014.12.027 56. Elgar MA De Crespigny FEC Ramamurthy S Male copulation behaviour and the risk of sperm competition Anim Behav 2003 66 211 216 10.1006/anbe.2003.2189 57. Vollrath F Male body size and fitness in the web-building spider Nephila clavipes Z Tierpsychol - J Comp Ethol 1980 53 1 61 78 10.1111/j.1439-0310.1980.tb00733.x 58. Higgins LE Developmental plasticity and fecundity in the orb-weaving spider Nephila clavipes J Arachnol 1992 20 2 94 106 59. Miyashita T Growth, egg-production and population-density of the spider Nephila clavata in relation to food conditions in the field Res Popul Ecol 1986 28 1 135 149 10.1007/BF02515542 60. Miyashita T Male male competition and mating success in the orb-web spider, Nephila clavata , with reference to temporal factors Ecol Res 1993 8 1 93 102 10.1007/BF02348612 61. Higgins L The interaction of season length and development time alters size at maturity Oecologia 2000 122 1 51 59 10.1007/PL00008835 62. Schneider JM, Michalik P. One-shot genitalia are not an evolutionary dead end - Regained male polygamy in a sperm limited spider species. BMC Evol Biol. 2011;11:197. doi:10.1186/1471-2148-11-197. 63. Schneider JM Lucass C Brandler W Fromhage L Spider Males Adjust Mate Choice but Not Sperm Allocation to Cues of a Rival Ethology 2011 117 11 970 978 10.1111/j.1439-0310.2011.01960.x 64. Fromhage L Schneider JM Safer sex with feeding females: sexual conflict in a cannibalistic spider Behav Ecol 2005 16 2 377 382 10.1093/beheco/ari011 65. Foelix RF Chemosensitive hairs in spiders J Morphol 1970 132 313 334 10.1002/jmor.1051320306 5485925 66. Foelix RF Chu-Wang I The morphology of spider sensilla. II. Chemoreceptors Tissue Cell 1973 5 461 478 10.1016/S0040-8166(73)80038-2 4744682 67. Lindenfors P Gittleman JL Jones KE Sexual size dimorphism in mammals 2007 68. Ceballos L Jones TM Elgar MA Patterns of Sperm Transfer in the Golden Orb-Weaver Nephila edulis Ethology 2015 121 6 617 624 10.1111/eth.12375 69. Higgins L Goodnight C Developmental response to low diets by giant Nephila clavipes females (Araneae: Nephilidae) J Arachnol 2011 39 3 399 408 10.1636/B11-18.1 70. Michalik P, Rittschof CC. A comparative analysis of the morphology and evolution of permanent sperm depletion in spiders. PLoS ONE. 2011;6(1):e16014. doi:10.1371/journal.pone.0016014. 71. Robinson MH Robinson B Comparative studies of the courtship and mating behavior of tropical araneid spiders Pac Insects Monogr 1980 36 1 218 72. Gaskett AC Spider sex pheromones: emission, reception, structures, and functions Biol Rev 2007 82 1 26 48 10.1111/j.1469-185X.2006.00002.x 73. Schulz S Toft S Identification of a sex pheromone from a spider Science 1993 260 1635 1637 10.1126/science.260.5114.1635 17810206 74. Uhl G Elias DO Herberstein ME Communication Spider Behaviour: Versatility and Flexibility 2011 Cambridge Cambridge University Press 75. Harari AR Zahavi T Thiery D Fitness costs of pheromone production in signaling female moths Evolution 2011 65 6 1572 1582 10.1111/j.1558-5646.2011.01252.x 21644949 76. Delisle J Vincent C Modified pheromone communication associated with insecticidal resistance in the obliquebanded leafroller, Choristoneura rosaceana (Lepidoptera: Tortricidae) Chemoecology 2002 12 1 47 51 10.1007/s00049-002-8325-1 77. Byers JA A cost of alarm pheromone production in cotton aphids, Aphis gossypii Naturwissenschaften 2005 92 2 69 72 10.1007/s00114-004-0592-y 15586262 78. Arnqvist G Comparative evidence for the evolution of genitalia by sexual selection Nature 1998 393 6687 784 786 10.1038/31689 79. Boulton RA Shuker DM The costs and benefits of multiple mating in a mostly monandrous wasp Evolution 2015 69 4 939 949 10.1111/evo.12636 25756346 80. Umbers KDL Symonds MRE Kokko H The Mothematics of Female Pheromone Signaling: Strategies for Aging Virgins Am Nat 2015 185 3 417 432 10.1086/679614 25674695 81. Cory A, Schneider JM. Old maids have more appeal: effects of age and pheromone source on mate attraction in an orb-web spider. PeerJ. 2016;4:e1877. doi:10.7717/peerj.1877. 82. Neumann R Schneider JM Data from: Socially cued developmental plasticity in web-building spiders Dryad Digital Respository 2016

PMC005xxxxxx/PMC5000427.txt

==== Front BMC Pregnancy ChildbirthBMC Pregnancy ChildbirthBMC Pregnancy and Childbirth1471-2393BioMed Central London 104510.1186/s12884-016-1045-2Research ArticleAnalyzing video recorded support of postnatal transition in preterm infants following a c-section Konstantelos Dimitrios dimitrios.konstantelos@uniklinikum-dresden.de Dinger Jürgen juergen.dinger@uniklinikum-dresden.de Ifflaender Sascha sascha.ifflaender@uniklinikum-dresden.de Rüdiger Mario Mario.ruediger@uniklinikum-dresden.de Department of Neonatology and Pediatric Intensive Care, Medizinische Fakultät Carl Gustav Carus, TU Dresden, Fetscherstraße 74, Dresden, 01307 Germany 25 8 2016 25 8 2016 2016 16 1 24619 3 2016 26 7 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Over the past years, research on neonatal resuscitation has focused on single interventions. The present study was performed to analyze the process quality of delivery room management of preterm infants born by c-section in our institution. Methods We performed a cross-sectional study of videos of preterm infants born by c-section. Videos were analyzed according to time point, duration and number of performed medical interventions. The study period occurred between January 2012 and December 2013. Infants were caterogized in 3 groups according to their gestational age. Results One hundred eleven videos were analyzed. 100 (90 %) of the infants were transferred to NICU and 91 (83 %) received respiratory support after a median of 0.5 min. All infants were auscultated after 8 (5–16) seconds median (IQR) and an oxygen saturation sensor was placed after 37 (28–52) seconds. 23 infants were intubated after 9 (6–17) minutes and 17 received exogenous surfactant; 29 % according to INSURE (intubation-surfactant-extubation) technique. The duration of intubation attempts was 47 (25–60) seconds. 51 % of the newborns received a sustained inflation for 8 (6–9) seconds. A successful IV-line placement occurred after 15 (12–20) minutes. 4 % of the infants were transported to the NICU without an IV-line after 3 (difference range: 2–5) unsuccessful attempts. Conclusions Using video analysis as a tool to study process quality, we conclude that interventions differ not only between but also within similar age groups. This data can be used for benchmarking with current guidelines and practice in other centers. Electronic supplementary material The online version of this article (doi:10.1186/s12884-016-1045-2) contains supplementary material, which is available to authorized users. Keywords Delivery roomVideo recordingManagementPretermissue-copyright-statement© The Author(s) 2016 ==== Body Background Postnatal adaptation during the first “Golden Hour of Life” represents a crucial time. Especially on immature newborns on whom quality of delivery room (DR) management is of great importance. Thus, research has focused on optimizing postnatal adaptation [1–3]. New concepts of respiratory support during the DR-management such as continuous positive airway pressure (CPAP) [4–6], sustained inflations (SI) [7–11], and less invasive surfactant administration have been studied [12–14]. Furthermore, the importance of monitoring – not only of vital signs but also respiratory parameters or body temperature – has been proven in clinical studies [15, 16]. Finally, interventions that have been widely used in the past – such as suctioning of the oropharynx – have been now abandoned [17]. However, there is not a single “magic bullet” to improve DR-management of preterm infants [18, 19]. In order to reach an improvement outcome, the entire approach has to be changed from an aggressive resuscitation towards a gentle support of transition, which requires multimodal management [20, 21]. In an attempt to improve the outcome of preterm infants in our institution, we implemented a quality assurance program for DR-management [22–24]. The current study was performed to identify differences in current DR-management of preterms within and between different gestational age (GA) groups. Methods DR-management In our institution, postnatal transition of a preterm infant is supported by a caregiving team consisting of a neonatologist, a pediatric resident (or neonatal fellow) and a NICU nurse. After birth the midwife places the infant onto a scale to be weighed, thereafter the NICU-nurse places the infant under a radiant heater. Newborns weighing less than 1500 g are placed in a plastic wrap. The newborn management room has a temperature of 24–26° Celsius. DR-management should adhere to local guidelines, which are in accordance with international recommendations. In short, the neonatologist in charge is standing behind the head of the newborn and is responsible for the respiratory support. The second caregiver – responsible for initial auscultation – is standing on the left side of the newborn and shows the newborn’s heartbeat frequency by moving his finger until both a reliable heart rate and saturation readings are available. If the newborn is in need of a respiratory support, CPAP is administered via a facemask with a pressure of 6–8 cm H20. In case of an initial bradycardia (<100 bpm) a sustained inflation with a pressure of 20 cm H20 for 10 s is applied up to a maximum of 3 times. If bradycardia persists, positive pressure ventilation via a facemask is commenced. After respiratory stabilization of the newborn an intubation-surfactant-extubation (INSURE) procedure is performed if the predefined criteria are met. Prior to transferring the infant to the NICU, an intravenous line (IV-line) is placed, blood is collected, glucose infusion is administered and the newborn is shown to their mother. Video recording Video recording of DR-management was performed as previously described [22–24]. For the present study, all recordings of preterm infants (<37 GA) born after c-section in 2012 and 2013 during the morning shift on working days were analyzed. Videos of infants with known congenital malformation were excluded. Analysis started with the arrival of the newborn under the radiant heater (Time point 0), which takes place usually about 20 s after cord clamping [22]. For the current analysis we distinguished the infants according to their gestational age arbitrarily in the following three groups: low gestational age (LGA): ≥ 34 GA, very low gestational age (VLGA): 31+0 – 33+6, extremely low gestational age (ELGA): <31 GA. For infants that stayed in the delivery room, analysis ended either at transfer to the mother or after 10 min (whichever of the two preceded the other). For infants in need of a further treatment in NICU, analysis was extended until transfer to the NICU. Analysis of data The following medical interventions were analyzed with regard to start time, total duration and frequency of occurrence. Routine care All routine interventions were performed either to assess or stabilize the infant’s condition (auscultation, finger heart rate sign, new towel or plastic wrap placement to prevent heat loss, placement of saturation sensor, temperature management). Respiratory support CPAP administration via a facemask, endotracheal or nasopharyngeal tube (which is mainly used during NICU transfer), SI, ventilation with consecutive inflations via a facemask or through an endotracheal tube, intubation (defined as the time that laryngoscope stayed in infants mouth), surfactant administration, suctioning. IV-line placement Time point and duration (defined as the time between end of disinfection of the skin and stabilization of the IV-line with a band) of successful and unsuccessful IV-line placement attempts were determined. Other interventions The total time during which the infant was not manipulated was calculated, were manipulations are defined as any of the above-described interventions or any other contact with the infant’s skin. Statistical analysis Descriptive data are presented as median (interquartile range, IQR). Interventions were analyzed with respect to parameters such as time of first occurrence, duration and frequency. Two box-plot figures were used to display the variance with respect to the time point of first occurrence. STROBE statement We confirm that our research complies with the STROBE [Strengthening the Reporting of Observational Studies in Epidemiology (collaboration)] guidelines. Results One hundred eleven videos of DR-management in preterm infants born by c-section were analyzed. We excluded 5 videos due to congenital malformations. Videos had a total duration of 2619 min [median of 24 (IQR, 19–29) minutes per video]. 22 different caregivers performed DR-management on these infants with a median frequency of 2 (IQR 1–5) patients per caregiver. 50 videos of LGA, 46 of VLGA and 15 of ELGA infants were analyzed. Only 11 of the 111 (9.9 %) preterm infants could remain in the delivery room with their mother. These infants had a median gestational age of 35+6 weeks (IQR, 35+5 – 36+5). The remaining 100 infants [GA 33+1 (32 – 34+4)] were transferred for further treatment after a median of 24 (19–29) minutes (Table 1).Table 1 Median (IQR) duration of various events LGA VLGA ELGA n n n Timepoint of NICU transfer (minute) 39 21 (17–25) 46 24 (20–34) 15 28 (24–34) No manipulation (seconds) 38 44 (8–104) 13 11 (2–47) 6 3 (1–17)  - % of analyzed time 4 (1–13) 1 (0–4) 0 (0–1) Duration of single SI (seconds) 16 7.4 (6.1-8.2) 27 8.2 (5.8-9.4) 13 8.1 (5.8-9.9) Duration of single intubation attempt (seconds) 46 52 (27–64) 45 (24–60) For most of the time some interventions were performed on the infants. On 57 of the 111 infants there was a short time (25 s (6–79) in median) without any manipulation. As shown in Table 1, the more immature the infants were, the less the likelihood of manipulation. Prevention of heat loss All but 3 infants were initially placed under the radiant warmer in a towel and were dried; the remaining 3 were placed in a plastic wrap without drying. Fourteen of the dried infants were also placed in a plastic wrap (1 in LGA, 5 in VLGA and 11 in ELGA). Towels were changed on 41 infants twice and on 2 infants three times. Temperature was measured on 54 (16 in LGA, 25 in VLGA, 13 in ELGA) infants once (median 1–1 times) after a median of 13 (5–21) minutes. Assessment of infant’s condition Auscultation was performed on all newborns for a median of 3.9 (2–6) minutes, which represents between 0.4 and 70.1 % of the analyzed time. Pulse frequency was visualized on 95 infants (42 in LGA, 39 in VLGA and 14 in ELGA) and for 18.4 % (8.5–33.7) of the total auscultation time. A pulse-oximetry sensor was placed on all infants after a median of 37 (28–52) seconds (Fig. 1).Fig. 1 Time to first use of intervention I. Shown are how often the first 7 interventions were given (n in parenthesis - % of all of the groups). For each intervention the time point of first occurrence is given as median (horizontal line in the middle of each box), 25th and 75th percentiles (top and bottom of each box) and minimum/maximum (whiskers mark) of all analyzed recordings Respiratory support A total of 91 (83 %) of infants received respiratory support, starting in a median of 0.5 min after arrival under radiant heater (Fig. 1) and for a median duration of 22 min (Table 2). The more immature the infant, the earlier the respiratory support was started (Fig. 1). No newborn required chest compressions.Table 2 Detailed respiratory support durations LGA VLGA ELGA n Total Timea n Total Timea n Total Timea Total time spent for respiratory support 31 17.9 (11.6-23.2)b 45 22.8 (16.6-30.5) 15 28.4 (24–31.5)c CPAP  -CPAP through face mask or endotracheal tubed 31 9.6 (6.8-14.7) 45 10.9 (7.4-19.1) 15 9.7 (5.2-19.4)  -CPAP through nasopharyngeal tube 24 7.8 (4.1-12.2) 35 7.8 (4.3-13.7) 7 13.6 (3–15.8) Mechanical Ventilation  -Consecutive inflations 9 0.5 (0.2-1.5) 18 1.1 (0.5-2.8) 12 1.5 (0.6-2.7)  -Through endotracheal tube 1 11.6 7 12.1 (5.7-20.1) 9 14.9 (7.7-24.6) aduration is presented in minutes [median (IQR)] bp between LGA and VLGA < 0.05 cp between LGA and ELGA < 0.001 dNon-mechanical ventilation delivered in newborns intubated for surfactant application purposes only (intubation – surfactant application – extubation) A sustained inflation was performed on 56 newborns 3 times (2–3) in median. Duration of single sustained inflation was in median 8 s (5.9–9.4) (Table 1). Mechanical ventilation (through an endotracheal tube or as consecutive inflations via a face mask) was performed on 41 infants within a median of 2.4 min after arrival at the resuscitaire. Less than a quarter of the infants (1 from LGA, 11 from VLGA and 11 from ELGA) were intubated in the delivery room. The first intubation attempt was performed in a median after 9.3 min (Fig. 2). For successful intubation a median of 2 attempts (range 1–2) were required with a median duration of 47 (25–60) seconds (Fig. 3).Fig. 2 Time to first use of intervention II. Shown are how often the rest of the interventions were given (n in parenthesis - % of all of the groups). For each intervention the time point of first occurrence is given as median (horizontal line in the middle of each box), 25th and 75th percentiles (top and bottom of each box) and minimum/maximum (whiskers mark) of all analyzed recordings Fig. 3 Time taken to intubate according to GA. Shown are the intubation attempts according to gestational age (GA). Every dot represents the duration of an intubation attempt (including successful and unsuccessful attempts) A total of 17 infants received exogenous surfactant (1 from LGA, 6 from VLGA and 10 from ELGA) after a median of 15 min (Fig. 2). Only 5 (29 %) of them were extubated after surfactant administration (INSURE). Duration of surfactant administration was 392 s in LGA, 76 (35–115) in VLGA and 69 (44–113) in ELGA. A facemask was replaced by a naso-pharyngeal tube on 66 infants after 14 (10–18) minutes to administer CPAP for subsequent transport to the NICU (Fig. 2). Time spent for nasopharyngeal tube placement was between 6 and 65 s per attempt. 78 infants (71 %) were suctioned a median of 2 (1–4) times, lasting 7 (5–13) seconds per attempt. IV-line placement An attempt to place an IV-line was performed on 100 infants (39 on LGA, 46 on VLGA and 15 on ELGA) (Table 3). On 4 LGA infants after three attempts in median (difference range: 2–5), an IV-line could not be placed and these infants were transported without an IV-line to the NICU. No infant received an umbilical cord catheterization in the DR.Table 3 IV-Line placement data LGA VLGA ELGA Infants who received an IV-line (%) 35 (70) 46 (100) 15 (100) Attempts /infant 1 (1–5) 1 (1–7) 1 (1–3) % of success from first attempt 64 59 60 Total time (minutes) 4 (3–6) 5 (3–7) 5 (3–6) Timepoint at completion of the IV-Linea 13 (10–18) 15 (11–19) 19 (13–22) Time spent only for successful attempt (minutes) 3 (2–4) 3 (3–4) 3 (3–5) atimepoint is presented in minute on the resuscitaire [median (IQR)] Discussion The more immature the infant the more likely some form of medical support during postnatal transition will be required. The quality of DR-management does have a great impact on the outcome of these vulnerable infants. Whereas some interventions have been investigated in clinical studies, little is known about the procedure followed during the entire postnatal management. According to our knowledge, the present study is the first detailed analysis of DR-interventions on preterm infants. Using video analysis as a tool to study process quality, we show that interventions differ not only between but also within similar age groups. By detecting deviations from local guidelines we have provided a tool for improving the quality of DR-management and we have defined areas so as to achieve a quality improvement with goals that can be quantified and thus its achievement verified. Variation with respect to gestational age Immaturity often results in disturbances of respiratory adaptation and subsequent need of medical support. In the present study, most of the infants received respiratory support, however, time of initiation varied with gestational age. Whereas respiratory support started immediately after arrival at the resuscitaire on ELGA-infants, it started significantly later on VLGA. These findings raise an interesting question; should respiratory support be considered a prevention (with an immediate initiation without any symptoms) or a treatment (initiated if first symptoms are present) of respiratory distress? Prematurity and delivery by c-section are two major risk factors for subsequent development of respiratory distress. Early CPAP is more effective in preventing than in treating respiratory distress [25]. Since almost all VLGA infants required some respiratory support at some point in time, it could be argued, that initiation of CPAP should start as soon as possible on infants with a gestational age below 33 weeks. That aspect could be considered in future guidelines, however appropriate trials are needed. A similar variation between different GA-groups was found for sustained inflations. Whereas the majority (87 %) of ELGA infants received a SI, only 59 % of VLGA and 32 % of LGA received a SI. Furthermore, SI was administered within the first minute on ELGA and VLGA, but only after 7 min on LGA. During a SI the airway pressure is increased, pressing the lung fluid into the interstitial space. Studies on rabbits showed benefit when SI is administrated prior to the onset of breathing, e.g. in fluid filled lungs [26]. The observed practice of applying a SI after 7 min is questionable since there is no data concerning a benefit but rather possible side effects of administering this airway pressure for a long time in (partially) air-filled lungs. According to the present analysis, only infants with a gestational age below 31 weeks were considered to breathe insufficiently by the caregivers and thus received SI. However, studies have shown that breathing is present in smaller infants and the need for SI could be questioned. Adherence to guidelines SI – If administered appropriately – seems to be effective in preventing intubation in DR [9]. Thus, SI’s were incorporated into our local neonatal resuscitation guidelines in 2010. However, the present analysis shows that clinical practice differs from the current recommendation. Time of initiation and duration of SI varied in and between different GA-groups. On some ELGA infants a single inflation lasted up to 19 s and some infants received up to 15 SI. To date, there is a lack of studies on SI and little is known regarding the optimal duration, time point or pressure [9, 27]. Since the present study was not designed to analyze the outcome of the DR management it does not provide any data regarding the incidence of pneumothorax. In the light of limited evidence on the benefits or side effects it seems to be important to adhere strictly to local recommendations, which are based on existing clinical data. Local guidelines recommend, when surfactant administration is considered necessary, that infants should be intubated after a period of stabilization, surfactant should be given as a slow bolus during spontaneous breathing and infants should be extubated on CPAP [modified less invasive surfactant administration (LISA)-procedure] [28]. Our analysis showed that the management we followed deviates from these guidelines. About a quarter of infants that were intubated did not receive surfactant even though they met criteria for surfactant administration. Furthermore, only about 29 % of infants were extubated after surfactant administration. To improve quality of DR-management it will be necessary to detect reasons for not complying with the current guidelines [6]. As shown by Schilleman et al., guidelines could be too complicated [29]. According to our analysis, the duration of surfactant administration also varied significantly. Whereas slow surfactant infusion has been shown to be inefficient on ventilated infants, there is no sufficient data on the optimal duration of surfactant administration during spontaneous ventilation. For LISA-procedure with a feeding tube a time of 1–3 min has been suggested [28, 30]. However, it remains unclear whether it will be appropriate for intubated infants as well. Monitoring of vital parameters represents an important aspect of DR-management [16]. Since measurement of oxygen saturation is a prerequisite to guide oxygen therapy, the request for an immediate placement of saturation sensor after arrival of the infant was added to our local guidelines a few years ago. Our analysis showed that the sensor was placed in a median after 37 s – thus we achieved sufficient adherence to our guidelines with regard to saturation measurement. Our guidelines further recommend that during auscultation the actual heartbeat must be shown by a finger movement until a sufficient sensor signal is available. That happened in approximately 86 % of cases. Since heart rate is an important parameter to evaluate efficacy of postnatal adaptation, an improvement is needed. Defining fields for improvement International guidelines suggest 20 s as an appropriate time for intubation; however, this recommendation is based on limited data. Our current analysis showed great variation regarding duration of intubation, lasting up to 2 min. In approximately 11 % of all intubations attempts, time was as recommended in current guidelines. Interestingly, time needed for intubation did not differ significantly between ELGA and VLGA. It could be argued that the lack of medication for intubation and surfactant administration is a reason for the longer time needed for intubation. On the other hand, our data supports ongoing discussions regarding the optimal duration for intubation [31–33]. Further research is needed to find a correlation between a critical time and outcome. Even if the current analysis was not designed to study side effects of intubation, no major complications were observed. Nevertheless, improving the intubation skills of the caregiver has been defined as an enhancement target in our institution. The aim being to reduce median intubation time down to 30 s. Stress of the newborn is associated with increased energy demand and oxygen consumption, which will lead to severe complications on preterm infants [34, 35]. Therefore, it is the general consensus to minimize stress and to apply the principles of optimal handling on these vulnerable infants. Our analysis showed that the first minutes of extra-uterine life are rather stressful for preterm; almost all infants experienced some kind of handling during some point of postnatal adaptation. Considering the importance of undisturbed adaptation we decided to re-evaluate the necessity of all handling procedures during DR-management. In future studies we aim to investigate, whether prolonged periods of no handling will have a beneficial effect on postnatal adaptation. An important aspect to reduce energy consumption is appropriate temperature management. Our analysis shows good adherence to our internal guidelines, which recommends plastic wrap on infants with a birth weight below 1500 g. Whereas low admission temperatures are easily prevented by plastic wrap, there is a danger of overheating. Thus, according to our guidelines it is recommended to check the temperature. In the current analysis only about 49 % of the infant’s temperature was measured in a median after 13 min. Since we did not record the NICU admission temperature in the present study, we are not able to give any data concerning the effect. We have recently described great variations concerning the time needed for IV-placement in term newborns [22]. The present study shows similar results for preterm infants. Interestingly, the number of attempts and the duration was lowest on the most immature infants. This finding can be explained by the fact that IV-lines were placed by more experienced caregivers on these infants, whereas lines were placed by more junior staff in more mature infants. Nevertheless, improving skills of placing IV-lines have to be improved in the future. Since placement of IV-lines represents a common procedure in neonatology, it would be of interest to compare different centers or even the policy of nurses versus medical doctors. Whereas routine suctioning has been recommended in the past, it has been abandoned in both current international and local guidelines. Nevertheless, 70 % of the newborns were suctioned during the time period of current analysis. In a previous publication we have analyzed the effects of suctioning in term infants and did not find any beneficial effect (but also no side-effects) [24]. Thus, future training of our staff will focus on preventing routine suctioning. Conclusions Avoidance of stress during postnatal transition seems to be important, especially in the preterm population. Thus, manipulation should be restricted to a required minimum; delivery room management should be focused on supporting transition rather than resuscitation. Whereas that approach seems to be a simple mission, daily practice is rather different. Video monitoring of delivery room management, combined with a subsequent structured analysis and feedback represents an important tool to discuss the “appropriateness” of all administered interventions. The present study shows that analyzing of different processes of DR-management is feasible in great detail. Data shows variations in care of the infants which could have an impact on subsequent outcome. In the past, clinical studies were performed to examine the effect of a single intervention of DR-management. Despite the scientific advantage of this approach, it has a major limitation since it neglects the heterogeneity of DR-management in the daily routine. By analyzing all processes of current management, variations can be found which could be of clinical relevance. As a consequence habits can be changed. A subsequent re-evaluation of processes and outcome will show whether changes have improved outcomes. Even more interestingly, benchmarking of different centers can be initiated, based upon a valid process analysis. Additional files Additional file 1: Table S1. Exact duration of various events for every LGA, VLGA and ELGA that was analyzed. (XLSX 45 kb) Additional file 2: Figures S1, S2. The time point of first occurrence of an intervention for every LGA, VLGA and ELGA that the specific intervention was given. (XLSX 78 kb) Additional file 3: Table S2. Detailed respiratory support administration duration of every LGA, VLGA and ELGA. (XLSX 43 kb) Additional file 4: Figure S3. Duration of every intubation attempt (successful and unsuccessful) according to the GA. (XLSX 30 kb) Additional file 5: Table S3. IV-Line placement data according to GA. (XLSX 48 kb) Abbreviations CPAPContinuous positive airway pressure c-sectionCaesarean section ELGAExtremely low gestational age GAWeeks of gestation INSUREIntubation-surfactant-extubation IV-lineIntravenous line LGALow gestational age LISALess invasive surfactant administration SISustained inflation VLGAVery low gestational age Acknowledgements Special thanks are given to Kate Erin Shaw and Anna Dorgia for proofreading the manuscript. Funding The study was supported by Else Kröner-Fresenius-Stiftung. The sponsor had no involvement in the study design, in the collection, analysis and interpretation of data; in the writing of the manuscript; and in the decision to submit the manuscript for publication. We acknowledge support by the German Research Foundation and the Open Access Publication Funds of the TU Dresden. Availability of data and materials All data generated or analysed during this study are included in this published article as Additional files 1, 2, 3, 4 and 5. Authors’ contributions DK and MR conceptualized and designed the study, drafted the initial manuscript, carried out initial analyses, contributed to the acquisition of data. JD and SI contributed to the design of the study and for acquisition of data. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Ethics approval and consent to participate In our institution, video recording is part of routine patient care and approved by the Local Ethic Committee (Ethikkomission an der Technischen Universität Dresden), thus all data gained from these video recordings can be used for research purposes as long as no patient or medical care worker can be identified. Therefore, the need of a written informed consent for participation in the study from a parent/guardian of the infants was waived by the Local Ethic Committee. Videos were stored for later analysis in a way that identification of individual patients was not possible. Recording did not affect any aspect of care of the individual patient. ==== Refs References 1. O’Donnell CPF Turn and face the strange - ch.ch.ch.changes to neonatal resuscitation guidelines in the past decade J Paediatr Child Health 2012 48 735 739 10.1111/j.1440-1754.2012.02531.x 22970666 2. Raghuveer TS Cox AJ Neonatal resuscitation: an update Am Fam Physician 2011 83 911 918 21524031 3. Davis PG Dawson JA New concepts in neonatal resuscitation Curr Opin Pediatr 2012 24 147 153 10.1097/MOP.0b013e3283504e11 22327948 4. Gittermann MK Fusch C Gittermann AR Regazzoni BM Moessinger AC Early nasal continuous positive airway pressure treatment reduces the need for intubation in very low birth weight infants Eur J Pediatr 1997 156 384 388 10.1007/s004310050620 9177982 5. Finer NN Carlo WA Duara S Fanaroff AA Donovan EF Wright LL Kandefer S Poole WK Delivery room continuous positive airway pressure/positive end-expiratory pressure in extremely low birth weight infants: a feasibility trial Pediatrics 2004 114 651 657 10.1542/peds.2004-0394 15342835 6. Lindner W Voßbeck S Hummler H Pohlandt F Delivery room management of extremely Low birth weight infants : spontaneous breathing or intubation? Pediatrics 1999 103 961 967 10.1542/peds.103.5.961 10224173 7. Dani C Lista G Pratesi S Boni L Agosti M Biban P Del Vecchio A Gazzolo D Gizzi C Magaldi R Messner H Mosca F Sandri F Scopesi F Trevisanuto D Vento G Sustained lung inflation in the delivery room in preterm infants at high risk of respiratory distress syndrome (SLI STUDY): study protocol for a randomized controlled trial Trials 2013 14 67 10.1186/1745-6215-14-67 23497495 8. van Vonderen JJ Hooper SB Hummler HD Lopriore E te Pas AB Effects of a sustained inflation in preterm infants at birth J Pediatr 2014 165 5 903 908 10.1016/j.jpeds.2014.06.007 25039041 9. te Pas AB Walther FJ A randomized, controlled trial of delivery-room respiratory management in very preterm infants Pediatrics 2007 120 322 329 10.1542/peds.2007-0114 17671058 10. te Pas AB Siew M Wallace MJ Kitchen MJ Fouras A Lewis RA Yagi N Uesugi K Donath S Davis PG Morley CJ Hooper SB Establishing functional residual capacity at birth: the effect of sustained inflation and positive end-expiratory pressure in a preterm rabbit model Pediatr Res 2009 65 537 541 10.1203/PDR.0b013e31819da21b 19190537 11. Lindner W Högel J Pohlandt F Sustained pressure-controlled inflation or intermittent mandatory ventilation in preterm infants in the delivery room? a randomized, controlled trial on initial respiratory support via nasopharyngeal tube Acta Paediatr 2005 94 303 309 16028648 12. Kendig J Notter R Cox C Reubens L Davis J Maniscalco W Sinkin R Bartoletti A Dweck H Horgan M Risemberg H Phelps DS D A comparison of surfactant as immediate prophylaxis and as rescue therapy in newborns of less than 30 weeks’ gestation N Engl J Med 1991 324 866 871 10.1056/NEJM199103283241301 13. Doron MW Veness-Meehan KA Margolis LH Holoman EM Stiles AD Delivery room resuscitation decisions for extremely premature infants Pediatrics 1998 102 574 582 10.1542/peds.102.3.574 9738179 14. Horbar JD Carpenter JH Buzas J Soll RF Suresh G Bracken MB Leviton LC Plsek PE Sinclair JC Collaborative quality improvement to promote evidence based surfactant for preterm infants: a cluster randomised trial BMJ 2004 329 7473 1004 10.1136/bmj.329.7473.1004 15514344 15. Laptook AR Watkinson M Temperature management in the delivery room Semin Fetal Neonatal Med 2008 13 383 391 10.1016/j.siny.2008.04.003 18501693 16. Dawson JA Davis PG O’Donnell CPF Kamlin COF Morley CJ Pulse oximetry for monitoring infants in the delivery room: a review Arch Dis Child Fetal Neonatal Ed 2007 92 F4 F7 10.1136/adc.2006.102749 17185428 17. Velaphi S Vidyasagar D The pros and cons of suctioning at the perineum (intrapartum) and post-delivery with and without meconium Semin Fetal Neonatal Med 2008 13 375 382 10.1016/j.siny.2008.04.001 18474453 18. Mehler K Grimme J Abele J Huenseler C Roth B Kribs A Outcome of extremely low gestational age newborns after introduction of a revised protocol to assist preterm infants in their transition to extrauterine life Acta Paediatr Int J Paediatr 2012 101 1232 1239 10.1111/apa.12015 19. Klebermass-Schrehof K Wald M Schwindt J Grill A Prusa A-R Haiden N Hayde M Waldhoer T Fuiko R Berger A Less invasive surfactant administration in extremely preterm infants: impact on mortality and morbidity Neonatology 2013 103 252 258 10.1159/000346521 23446061 20. Jobe AH Transition/adaptation in the delivery room and less RDS: “Don’t just do something, stand there!” J Pediatr 2005 147 284 286 10.1016/j.jpeds.2005.05.003 16182661 21. Rüdiger M Braun N Gurth H Bergert R Dinger J Preterm resuscitation I: clinical approaches to improve management in delivery room Early Hum Dev 2011 87 749 753 10.1016/j.earlhumdev.2011.08.019 21920678 22. Konstantelos D Ifflaender S Dinger J Burkhardt W Rüdiger M Analyzing support of postnatal transition in term infants after c-section BMC Pregnancy Childbirth 2014 14 225 10.1186/1471-2393-14-225 25011378 23. Konstantelos D Gurth H Bergert R Ifflaender S Rüdiger M Positioning of term infants during delivery room routine handling - analysis of videos BMC Pediatr 2014 14 33 10.1186/1471-2431-14-33 24495525 24. Konstantelos D Ifflaender S Dinger J Rüdiger M Suctioning habits in the delivery room and the influence on postnatal adaptation - a video analysis J Perinat Med 2015 43 6 777 782 10.1515/jpm-2014-0188 25324437 25. Sweet DG Carnielli V Greisen G Hallman M Ozek E Plavka R Saugstad OD Simeoni U Speer CP Halliday HL European consensus guidelines on the management of neonatal respiratory distress syndrome in preterm infants - 2010 update Neonatology 2010 97 402 417 10.1159/000297773 20551710 26. Sobotka KS Hooper SB Allison BJ te Pas AB Davis PG Morley CJ Moss TJM An initial sustained inflation improves the respiratory and cardiovascular transition at birth in preterm lambs Pediatr Res 2011 70 56 60 10.1203/PDR.0b013e31821d06a1 21659961 27. te Pas AB Siew M Wallace MJ Kitchen MJ Fouras A Lewis RA Yagi N Uesugi K Donath S Davis PG Morley CJ Hooper SB Effect of sustained inflation length on establishing functional residual capacity at birth in ventilated premature rabbits Pediatr Res 2009 66 295 300 10.1203/PDR.0b013e3181b1bca4 19542905 28. Kribs A Roll C Göpel W Wieg C Groneck P Laux R Teig N Hoehn T Böhm W Welzing L Vochem M Hoppenz M Bührer C Mehler K Stützer H Franklin J Stöhr A Herting E Roth B Nonintubated surfactant application vs conventional therapy in extremely preterm infants: a randomized clinical trial JAMA Pediatr 2015 169 723 730 10.1001/jamapediatrics.2015.0504 26053341 29. Schilleman K Witlox RS van Vonderen JJ Roegholt E Walther FJ te Pas AB Auditing documentation on delivery room management using video and physiological recordings Arch Dis Child Fetal Neonatal Ed 2014 99 6 F485 F490 10.1136/archdischild-2014-306261 25125582 30. Göpel W Kribs A Ziegler A Laux R Hoehn T Wieg C Siegel J Avenarius S von der Wense A Vochem M Groneck P Weller U Möller J Härtel C Haller S Roth B Herting E Avoidance of mechanical ventilation by surfactant treatment of spontaneously breathing preterm infants (AMV): an open-label, randomised, controlled trial Lancet 2011 378 1627 1634 10.1016/S0140-6736(11)60986-0 21963186 31. Lane B Finer N Rich W Duration of intubation attempts during neonatal resuscitation J Pediatr 2004 145 1 67 70 10.1016/j.jpeds.2004.03.003 15238909 32. O’Donnell CPF Kamlin COF Davis PG Morley CJ Endotracheal intubation attempts during neonatal resuscitation: success rates, duration, and adverse effects Pediatrics 2006 117 e16 e21 10.1542/peds.2005-0901 16396845 33. Haubner LY Barry JS Johnston LC Soghier L Tatum PM Kessler D Downes K Auerbach M Neonatal intubation performance: room for improvement in tertiary neonatal intensive care units Resuscitation 2013 84 1359 1364 10.1016/j.resuscitation.2013.03.014 23562374 34. Speidel B Adverse effects of routine procedures on preterm infants Lancet 1978 1 8069 864 866 10.1016/S0140-6736(78)90204-0 76807 35. Sheila N Campbell Laura BS Nursing procedures and alterations in transcutaneous oxygen tension in premature infants Nurs Res 1982 31 330 336 6924215

PMC005xxxxxx/PMC5000428.txt

==== Front Stand Genomic SciStand Genomic SciStandards in Genomic Sciences1944-3277BioMed Central London 17610.1186/s40793-016-0176-4Extended Genome ReportFirst high quality draft genome sequence of a plant growth promoting and cold active enzyme producing psychrotrophic Arthrobacter agilis strain L77 Singh Ram N. 1Gaba Sonam 1Yadav Ajar N. 1Gaur Prakhar 1Gulati Sneha 1Kaushik Rajeev 1Saxena Anil K. +91 547 25300080saxena461@yahoo.com 121 Division of Microbiology, ICAR-Indian Agricultural Research Institute, New Delhi, 110012 India 2 Present Address: ICAR-National Bureau of Agriculturally Important Microorganisms, Kushmaur, Mau, 275103 Uttar Pradesh India 26 8 2016 26 8 2016 2016 11 1 5423 12 2015 15 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Arthrobacter agilis strain L77, is a plant growth promoting and cold active hydrolytic enzymes producing psychrotrophic bacterium, isolated from Pangong Lake, a subglacial lake in north western Himalayas, India. Genome analysis revealed metabolic versatility with genes involved in metabolism and cold shock adaptation, utilization and biosynthesis of diverse structural and storage polysaccharides such as plant based carbon polymers. The genome of Arthrobacter agilis strain L77 consists of 3,608,439 bp (3.60 Mb) of a circular chromosome. The genome comprises of 3316 protein coding genes and 74 RNA genes, 725 hypothetical proteins, 25 pseudo-genes and 1404 unique genes. Electronic supplementary material The online version of this article (doi:10.1186/s40793-016-0176-4) contains supplementary material, which is available to authorized users. Keywords ArthrobacterPsychrotrophicPGPBCold-active enzymesPangong LakeHimalayashttp://dx.doi.org/10.13039/501100001503Indian Council of Agricultural Research (IN)NAIP/Comp.4/C4/C.30033/2008-09Saxena Anil K. issue-copyright-statement© The Author(s) 2016 ==== Body Introduction The microorganisms from extreme environments are of particular importance in global ecology since the majority of terrestrial and aquatic ecosystems of our planet are permanently or seasonally submitted to cold temperatures. Microorganisms capable of coping with low temperatures are widespread in these natural environments where they often represent the dominant flora and they should therefore be regarded as the most successful colonizers of our planet. Members of the genus Arthrobacter [1, 2] are Gram-positive, show rods in exponential growth and cocci in their stationary phase, able to grow under aerobic as well as anaerobic conditions and belong to the phylum Actinobacteria [3]. Different species of Arthrobacter [1, 2] have been implicated in plant growth promotion [4], production of industrially important enzymes [5, 6] and as xeroprotectant [7, 8]. These reports suggest that species from Arthrobacter [1, 2] harbor genes for coding enzymes that can be useful in the industry, agriculture and biotechnology. Arthrobacter agilis [9] strain L77 was isolated from Pangong Lake, a subglacial lake in north western Himalayas, India and exhibit plant growth promoting attributes as well as production of hydrolytic enzymes. The culture was further characterized for production of EPS and anti-freeze compounds (AFCs). Here, we present the draft genome sequence of Arthrobacter agilis [9] strain L77 along with the description of genome properties and annotation. Organism information Classification and features Arthrobacter agilis [9] strain L77 was isolated from frozen sub-glacial Pangong Lake (33°82′55.59″N and 78°59′26.69″E) in north western Himalaya, India (Table 1). This psychrotrophic bacterium was isolated using standard serial dilution method on Trypticase soya agar [10] plate and has been reported to possess plant growth promoting attributes and could produce cold active enzymes and AFCs. It could solubilize phosphorus, zinc and could produce indole acetic acid and ammonia. It could produce cold active enzymes such as lipase, amylase, protease, chitinase and β-galactosidase.Table 1 Classification and general features of Arthrobacter agilis strain L77 MIGS ID Property Term Evidence codea Classification Domain Bacteria TAS [12] Phylum Actinobacteria TAS [3] Class Actinobacteria TAS [13] Order Actinomycetales TAS [2, 14] Family Micrococcaceae TAS [2, 15] Genus Arthrobacter TAS [1, 2] Species Arthrobacter agilis TAS [9] Strain L77 NAS Gram stain Positive IDA Cell shape Polymorphic: Coccus to rod shaped IDA Motility Non-motile TAS [9] Sporulation Non-sporulating TAS [9] Temperature range −10 °C −30 °C IDA Optimum temperature 15 °C IDA pH range; Optimum 6–9, 7 IDA Carbon source Yeast extract, glucose, lactose, mannose TAS [9] MIGS-6 Habitat Sub-glacial Lake IDA MIGS-6.3 Salinity Grown on 5 % > NaCl (w/v) IDA MIGS-22 Oxygen requirement Aerobic TAS [9] MIGS-15 Biotic relationship Free living TAS [9] MIGS-14 Pathogenicity Non-pathogeneic NAS MIGS-4 Geographic location India, Leh Ladakh, Jammu & Kashmir TAS [10] MIGS-5 Sample collection March 28, 2010 IDA MIGS-4.1 Latitude 33°82′55.59″N NAS MIGS-4.2 Longitude 78°59′26.69″E NAS MIGS-4.4 Altitude 3215 m NAS aEvidence codes - TAS Traceable Author Statement (i.e., a direct report exists in the literature), NAS Non-traceable Author Statement (i.e., not directly observed for the living, isolated sample, but based on a generally accepted property for the species, or anecdotal evidence). These evidence codes are from the Gene Ontology project [49] Strain L77 is a bright yellow colored (Fig. 1) Gram-positive, aerobic, non-motile bacterium exhibiting a rod-coccus cycle. The initial validation of bacterium was done by 16S rRNA gene sequencing using the universal eubacterial primers pA (5′-AGAGTTTGATCCTGGCTCAG-3′) and pH (5′-AAGGAGGTGATCCAGCCGCA-3′) [11]. The 16S rRNA gene sequence places Arthrobacter agilis strain L77 in the domain Bacteria [12] (Table 1), phylum Actinobacteria [3] and Class Actinobacteria [13], order Actinomycetales [2, 14] and family Micrococcaceae [2, 15] during homology search by BLAST [16]. Only few of the closely related species after reclassification [17] of genus Arthrobacter [1, 2,] with validly published names: A. agilisDSM 20550T [9], A. woluwensis 1551TDSM 10495 [18], A. methylotrophusDSM 14008T [19], A. tectiLMG 22282T [20], A. parietisLMG 22281T [20], A. subterraneus CH7TDSM 17585 [21], A. tumbaeLMG 19501T [20], Arthrobacter oryzae KV-651TDSM 25586 [22], Arthrobacter alkaliphilus LC6TDSM 23368 [23], Arthrobacter flavusJCM 11496T [24], A. cupressi D48TDSM 24664 [25], A. globiformisDSM 20124T [1, 2] were selected for drawing the phylogenetic position of strain L77.Fig. 1 Full grown yellow colored bacterial culture on Tripticase Soy Agar (TSA) medium A phylogenetic tree was constructed (Fig. 2) from the 16S rRNA gene sequence together with other Arthrobacter [1, 2] homologs using MEGA 6.0 software suite [26]. The evolutionary history was inferred by using the Maximum Likelihood method based on the Tamura-Nei model [27]. The tree with the highest log likelihood (0.14495825) is shown. The percentage of trees in which the associated taxa clustered together is shown next to the branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 13 nucleotide sequences. All positions containing gaps and missing data were eliminated. There were a total of 1553 positions in the final dataset. Evolutionary analyses were conducted in MEGA6.0 [26]. According to the 16S rRNA gene similarity, the nearest phylogenetic neighbors of Arthrobacter agilis strain L77 are Arthrobacter flavusJCM 11496T [24] (AB537168) with 97.8 %, A. tectiLMG 22282T [20] (AJ639829) with 97.13 %, A. parietisLMG 22281T [20] (AJ639830) with 97.41 %, A. subtrerraneus CH7TDSM 17585 [21] (DQ097525) with 97.66 % and A. tumbaeLMG 19501T [20] (AJ315069) with 97.68 % similarity. The 16S rRNA gene sequence also submitted to NCBI GenBank with the accession number KT804924.Fig. 2 Phylogenetic placements of Arthrobacter agilis strain L77 between known species of Arthrobacter genus Extended feature descriptions Arthrobacter agilis strain L77, a psychrotrophic bacterium, forms bright yellow color colonies (Fig. 1) on TSA medium and could grow in a pH range of 6–9 and tolerate 5 % NaCl. Growth studies showed that the isolate when incubated at 15 and 30 °C was in the exponential phase until 36 h, while at 4 °C, the exponential phase started after 24 h (Fig. 3). Freezing survival studies of Arthrobacter agilis strain L77 revealed that when the culture was initially grown at 4 °C prior to freezing at −10 and −20 °C, it showed significantly higher freezing survival rather than culture initially grown at 15 and 30 °C prior to freezing (Fig. 3).Fig. 3 Growth curves of Arthrobacter agilis strain L77 at three different temperatures 4, 15 and 30 °C Exopolysaccharide production was found to be higher at lower incubation temperatures (4 or 15 °C) in comparison to the optimal growth temperature (30 °C) for Arthrobacter agilis (L77) (Fig. 4). EPS production by psychrophilic bacteria is one of the adaptations at low temperatures. The high polyhydroxyl content of EPS lowers the freezing point and ice nucleation temperature of water. In addition, EPS can trap water, nutrients and metal-ions and facilitate surface adhesion, cellular aggregation and biofilm formation and may also play a role in protecting extracellular enzymes against cold denaturation and autolysis [28, 29].Fig. 4 The survival of Arthrobacter agilis strain L77 subjected to freezing temperature (−10 and −20 °C) shifted from three different temperatures 4, 15 and 30 °C Remarkable variations in terms of accumulation of various organic acids, sugars, polyols and amino acids were detected through HPLC at three different incubation temperatures (4, 15 and 30 °C) (Additional file 1: Table S1, Additional file 2: Table S2 and Fig. 5). Among the sugars, accumulation of mannitol and sorbitol was observed only at 4 °C. The amino acids expression pattern revealed that the most prominent increase was observed in the concentrations of glycine, cysteine and arginine at 4 °C (Additional file 2: Table S2). It has been reported that the cold active enzymes and efficient growth rates are used to facilitate and maintain the adequate metabolic fluxes at near freezing temperature for cold-adaptation [30]. The development of freezing tolerance by producing cryoprotectant compounds or adaptation of cytoplasmic enzymes to cold conditions for protecting cytoplasmic components is one of the strategy used by microbial cells to survive in freezing conditions as these molecules depress freezing point for the protection of cells [31].Fig. 5 EPS accumulation by Arthrobacter agilis strain L77 at three different temperatures 4, 15 and 30 °C Enhanced EPS production by the psychrophilic bacteria at low temperature suggests that EPS plays an important role in desiccation protection or prevention of drying of bacterial cells from freezing temperature. It can be assumed that the strain L77 follows a cold evading strategy to thrive in freezing conditions by synthesizing various cryoprotectants (sugars, polyols and amino acids). These cryoprotectants are known to depress freezing point to evade crystallization [32]. Genome sequencing information Genome project history This organism was selected for sequencing on the basis of its environmental and agricultural relevance to help in plant growth and ability to provide inorganic phosphate to crops at very low temperature. It also has biogeochemical importance of producing AFCs, so helpful for soil aeration. The genome project is deposited in the online genome database (NCBI-Genome). Sequencing, assembly and annotations were performed at Division of Microbiology, Indian Agricultural Research Institute (ICAR-IARI), New Delhi, India. A summary of the project information is shown in the Table 2.Table 2 Genome sequencing project information for Arthrobacter agilis strain L77 MIGS ID Property Term MIGS-31 Finishing quality Unfinished, improved high quality draft MIGS-28 Libraries used Paired End (insert size 250 bp) MIGS-29 Sequencing platforms Illumina MiSeq MIGS-31.2 Fold coverage 180× MIGS-30 Assemblers A5 pipeline v jan-2014 MIGS-32 Gene calling method Prodigal Locus Tag RY94 Genbank ID JWSU00000000.1-10.1 Genbank Date of Release 08-Jan-2015 GOLD ID Gp0117366 BIOPROJECT PRJNA270909 MIGS 13 Source Material Identifier L77 Project relevance Bioprospecting Growth conditions and genomic DNA preparation A culture of L77 was grown in Trypticase soya broth, until they reached an OD(600 nm) > 1.0. The cells were pelleted from 5 ml culture, washed thrice with TE buffer (10 mM Tris and 1 mM EDTA, pH 8.0) and the pellet was resuspended in 750 μl TE buffer. Genomic DNA was extracted from the suspended pellet using Zymo Research Fungal/Bacterial DNA MicroPrep™ following the standard protocol prescribed by the manufacturer. Genome sequencing and assembly The draft genome of Arthrobacter agilis strain L77 (PRJNA270909) was generated at the Division of Microbiology, ICAR-Indian Agricultural Research Institute (ICAR-IARI), New Delhi, India using Illumina [33] technology (Table 2). For this genome, we constructed and sequenced an Illumina MiSeq shotgun library which generated 1,568,654 reads totaling 321.8 Mb data. The raw fastq data was checked for quality using Fast QC [34]. Trimmomatic 0.32 [35] with Nextra adapter sequences was used to hard clip reads. Assembly of trimmed reads was carried out using a5 pipeline version 2014 [36] (Table 2). In terms of N50 and total number of scaffolds, the a5 pipeline [36] was found to be better than other genome assemblers. CONTIGuator [37] was used to improve the assembly draft. The final draft was identified as Arthrobacter agilis L77, using megablast with RDP 16S database, release 11–1 [38]. This whole-genome project (Bioproject ID: PRJNA270909) has been registered and assembled sequence data submitted at NCBI GenBank under the accession no. JWSU00000000.1-10.1. The version described in this paper is the first version. Genome annotation Genes were identified using Prokka 1.8 [39] based on Prodigal [40] (Table 2) as part of the Oak Ridge National Laboratory genome annotation pipeline. The predicted CDSs were further annotated on Pfam [41], and (COGs) [42]. These data sources were combined to assert a product description for each predicted protein. Non-coding genes and miscellaneous features were predicted using tRNAscan-SE [43], RNAMMer [44], Rfam [45], TMHMM [46], and signalP v4.1 [47] (Table 3).Table 3 Genome Statistics for Arthrobacter agilis strain L77 Attribute Value % of total Genome size (bp) 3,608,439 100.00 DNA coding (bp) 3,224,998 89.37 DNA G + C (bp) 2,518,329 69.79 DNA scaffolds 10 100.00 Total genes 3390 100.00 Protein coding genes 3316 97.81 RNA genes 84 2.18 Pseudo genes 25 0.73 Genes in internal clusters N/A N/A Genes with function prediction 2591 78.10 Genes assigned to COGs 2122 63.64 Genes assigned to Pfam domains 2855 85.11 Genes with signal peptides 126 5.51 Genes with transmembrane helices 852 25.6 CRISPR repeats N/A N/A Genome properties The genome is 3,608,439 bp in size, which has GC content of 69.79 mol % (Table 3). There are 47 tRNA, 1 tmRNA, 6 rRNA and 20 ncRNA genes. Of the 3390 predicted genes, 3316 are protein-coding genes (CDSs). Of the total CDSs, 63.64 % represent COG functional categories and 5.51 % consist of signal peptides (Table 3). The distribution of genes into COG functional categories are presented in Table 4. The genome map (Fig. 6) was visualized by CG view server [48].Table 4 Number of protein coding genes of Arthrobacter agilis strain L77 associated with general COG functional categories Code Value % agea COG category J 184 5.54 Translation, ribosomal structure and biogenesis A 1 0.03 RNA processing and modification K 208 6.27 Transcription L 109 3.28 Replication recombination and repair B 1 0.03 Chromatin structure and dynamics D 22 0.66 Cell cycle control, Cell division, chromosome partitioning V 49 1.47 Defense mechanisms T 113 3.40 Signal transduction mechanisms M 124 3.73 Cell wall/membrane biogenesis N 30 0.90 Cell motility U 19 0.57 Intracellular trafficking and secretion O 104 3.13 Posttranslational modification, protein turnover, chaperones C 110 3.31 Energy production and conversion G 213 6.42 Carbohydrate transport and metabolism E 200 6.03 Amino acid transport and metabolism F 71 2.14 Nucleotide transport and metabolism H 114 3.43 Coenzyme transport and metabolism I 88 2.65 Lipid transport and metabolism P 118 3.55 Inorganic ion transport and metabolism Q 38 1.14 Secondary metabolites biosynthesis, transport and catabolism R 204 6.15 General function prediction only S 166 5.00 Function unknown – 1030 31.06 Not in COGs aThe total is based on the number of protein coding genes in the annotated genome Fig. 6 Graphical map of genome of Arthrobacter agilis strain L77. From outside to centre: RNA genes (Brown, tRNA and light purple, rRNA) and other genes are colored according to COG categories. Inner circle shows the GC skew with positive (+) as dark green and negative (−) as dark purple. GC content is indicated in black Insights from the genome sequence The isolate was successfully screened for lipase, amylase, protease, chitinase and β-galactosidase. Genome analysis showed two important genes pstA and pstC which are required for the translocation of phosphate across the membranes. Another important gene, PstB (an ADP binding protein), of the phosphate transport system is responsible for giving energy to the phosphate transport system of the organism. PhoR and PhoP were also found which are important for regulation of phosphate operon. PhoH like protein has a probable ATPase which is induced when phosphate level decreases. Genome annotation also predicted a putative cold shock protein which is supposed to play an important role in low temperature conditions. There are other proteins which shares evolutionary relationship with bacterial cold shock proteins such as Rhodanase and S1 RNA binding protein suggesting their role in low temperature conditions. In-depth analysis of the genome could give us better insight into mechanism of tolerance of this strain to low temperature. Other temperature responsive proteins were found such as molecular chaperone Hsp31 and glyoxalase 3 that influence the exposure of hydrophobic domains of proteins and stabilize the early unfolding under high temperature stress conditions to provide stability to the isolate in temperature stress. Genes of heavy metal resistance were also found in the annotation. Mercuric resistance operon regulatory protein activates the mercury resistance operon in the presence of mercury thus protecting the bacteria from harmful side-effects of mercury. Mercuric reductase is also present which is responsible for conversion of Hg2+ to Hg0. copZ is a copper chaperone that replaces zinc with copper and releases copY from the DNA which is a negative regulator of copYZAB under excess copper. Gene of nitrogen regulation, nitrogen regulatory protein P-II was found that regulates the level of nitrogen by regulating glutamine. When the ratio of glutamine to 2-ketoglutarate decreases, uridine is added on a tyrosine of P-II to form P-II-UMP which in turn deadenylates glutamine synthase resulting in its activation. Putative genes coding for these activities were identified in the genome based on annotation (Table 5).Table 5 Candidate genes coding for putative lipase, amylase, chitinase, protease, β-galactosidase, phosphate transport regulation, cold shock proteins, chaperons and heavy metal resistance activities identified in Arthrobacter agilis strain L77 draft genome Putative Gene Annotation Size (aa) Lipase  ABAGL_00531 GDSL-like Lipase/Acylhydrolase 262  ABAGL_00732 Lipase 1 precursor 288  ABAGL_00875 GDSL-like Lipase/Acylhydrolase 267  ABAGL_01161 Lipase 1 precursor 350  ABAGL_03217 GDSL-like Lipase/Acylhydrolase 272 Amylase  ABAGL_00299 Glucose-resistance amylase regulator 338  ABAGL_01452 Glucose-resistance amylase regulator 336  ABAGL_01652 Trehalose synthase/amylase TreS 588  ABAGL_01737 Alpha-amylase precursor 905  ABAGL_01923 Alpha-amylase/pullulanase 257  ABAGL_01950 Glucose-resistance amylase regulator 327 Chitinase  ABAGL_01394 putative bifunctional chitinase/lysozyme precursor 520  ABAGL_01777 Chitinase 400 Protease  ABAGL_00100 Putative cysteine protease YraA 188  ABAGL_00190 Flp pilus assembly protein, protease CpaA 207  ABAGL_00447 Lon protease 364  ABAGL_00456 Putative serine protease HtrA 496  ABAGL_00667 Serine proteasec 401  ABAGL_00940 CAAX amino terminal protease self- immunity 268  ABAGL_00971 CAAX amino terminal protease self- immunity 247  ABAGL_01091 Serine protease Do-like HtrA 366  ABAGL_01213 Rhomboid protease GluP 291  ABAGL_01289 ATP-dependent zinc metalloprotease FtsH 689  ABAGL_01302 Putative ATP-dependent Clp protease ATP-binding subunit 835  ABAGL_01392 CAAX amino terminal protease self- immunity 266  ABAGL_01505 Minor extracellular protease vpr precursor 1059  ABAGL_01669 Flp pilus assembly protein, protease CpaA 168  ABAGL_01755 CAAX amino terminal protease self- immunity 326  ABAGL_02020 Putative serine protease HtrA 310  ABAGL_02206 Putative metalloprotease 303  ABAGL_02449 Putative zinc metalloproteasec/MT2700 388  ABAGL_02467 Modulator of FtsH protease HflK 310  ABAGL_02638 ATP-dependent Clp protease ATP-binding subunit ClpX 430  ABAGL_02639 ATP-dependent Clp protease proteolytic subunit 1 224  ABAGL_02640 ATP-dependent Clp protease proteolytic subunit 2 208  ABAGL_02862 ATP-dependent Clp protease adaptor protein ClpS 105  ABAGL_02923 ATP-dependent zinc metalloprotease FtsH 438  ABAGL_03163 Serine protease inhibitor-like protein 389  ABAGL_03211 CAAX amino terminal protease self- immunity 267  ABAGL_03271 Metalloprotease MmpA 447  ABAGL_00551 Protease PrtS precursor 355  ABAGL_00739 Protease 2 734  ABAGL_01958 Protease synthase and sporulation negative regulatory protein 215  ABAGL_02571 Protease PrsW 425  ABAGL_03295 Protease 3 precursor 455 β-galactosidase  ABAGL_00260 β-galactosidase bgaB 667  ABAGL_00292 β-galactosidase 687  ABAGL_01083 β-galactosidase precursor 708 Phosphate Transport Regulation  ABAGL_01317 Phosphate transport system permease protein PstA 310  ABAGL_01318 Phosphate import ATP-binding protein PstB 367  ABAGL_01316 Phosphate transport system permease protein PstC 259  ABAGL_00191 Alkaline phosphatase synthesis sensor protein PhoR 544  ABAGL_03137 Alkaline phosphatase synthesis sensor protein PhoR 555  ABAGL_01671 PhoH-like protein 443  ABAGL_02530 PhoH-like protein 344 Cold shock Proteins  ABAGL_01978 putative cold shock protein A 67 Chaperons  ABAGL_01554 Molecular chaperone Hsp31 and glyoxalase 3 255  ABAGL_01067 Copper chaperone CopZ 74 Heavy Metal Resistance  ABAGL_02628 Mercuric resistance operon regulatory protein 134 Conclusions The 3.6 Mb draft genome of Arthrobacter agilis strain L77 was assembled and annotated. The isolate was successfully screened for production of EPS and AFCs with potential application in biotechnology. The candidate genes coding for hydrolytic enzymes and cold shock proteins were identified in the genome. Arthrobacter agilis strain L77 will serve as a source for antifreeze proteins, functional enzymes and other bioactive molecules in future bioprospecting projects. Additional files Additional file 1: Table S1. Quantitative analysis of organic acid and sugars/polyols from Arthrobacter agilis strain L77 by HPLC. (DOCX 13 kb) Additional file 2: Table S2. Quantitative analysis of amino acids content of Arthrobacter agilis strain L77 by HPLC. (DOCX 15 kb) Abbreviations AFCsAnti-freeze compounds EPSExopolysaccharides The authors are grateful to the National Agricultural Innovation Project (NAIP), Indian Council of Agricultural Research, Govt. of India, New Delhi and Division of Microbiology, ICAR-Indian Agricultural Research Institute (IARI), Pusa, New Delhi and for providing the financial support facilities, to undertake the investigations. Author’s contributions RNS and SGa equally contributed to the work. RNS carried out the sample collection, participated in the strain identification, sequence alignment, assembly and annotation analysis and drafted the manuscript. SGa participated in the sequence assembly and annotation analysis. ANY and SGu carried out the bacterial isolation and performed the physiological assays. PG did the initial sequence assembly of the raw data. RK participated in sample collection and sequencing of 16S rRNA gene. AKS conceived of the study, and participated in its design, coordination and helped to finalize the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. ==== Refs References 1. Conn H Dimmick I Soil bacteria similar in morphology to Mycobacterium and Corynebacterium J Bacteriol 1947 54 3 291 16561362 2. Skerman V McGowan V Sneath P Moore W Moore LV Approved lists of bacterial names Int J Syst Bacteriol. 1980 30 225 420 10.1099/00207713-30-1-225 3. Garrity GM, Holt JG. The road map to the manual. Bergey’s Manual® of Systematic Bacteriology. Springer New York; 2001. p. 119–66. 4. Manzanera M Narváez-Reinaldo JJ García-Fontana C Vílchez JI González-López J Genome sequence of Arthrobacter koreensis 5J12A, a plant growth-promoting and desiccation-tolerant strain Genome Announc 2015 3 3 e00648 15 26067978 5. Kallimanis A LaButti KM Lapidus A Clum A Lykidis A Mavromatis K Complete genome sequence of Arthrobacter phenanthrenivorans type strain (Sphe3) Stand Genomic Sci 2011 4 2 123 10.4056/sigs.1393494 21677849 6. Kiran S Swarnkar MK Pal M Thakur R Tewari R Singh AK Complete genome sequencing of protease-producing novel Arthrobacter sp. strain IHBB 11108 using PacBio single-molecule real-time sequencing technology Genome Announc 2015 3 2 e00346 15 10.1128/genomeA.00346-15 25908143 7. Manzanera M Santa-Cruz-Calvo L Vílchez J García-Fontana C Silva-Castro G Calvo C Genome sequence of Arthrobacter siccitolerans 4 J27, a xeroprotectant-producing desiccation-tolerant microorganism Genome Announc 2014 2 3 e00526 14 10.1128/genomeA.00526-14 24948757 8. SantaCruz-Calvo L González-López J Manzanera M Arthrobacter siccitolerans sp. nov., a highly desiccation-tolerant, xeroprotectant-producing strain isolated from dry soil Int J Syst Evol Microbiol 2013 63 Pt 11 4174 80 10.1099/ijs.0.052902-0 23771623 9. Koch C Schumann P Stackebrandt E Reclassification of Micrococcus agilis (Ali-Cohen 1889) to the genus Arthrobacter as Arthrobacter agilis comb. nov. and emendation of the genus Arthrobacter Int J Syst Bacteriol 1995 45 4 837 9 10.1099/00207713-45-4-837 7547308 10. Yadav AN Sachan SG Verma P Tyagi SP Kaushik R Saxena AK Culturable diversity and functional annotation of psychrotrophic bacteria from cold desert of Leh Ladakh (India) World J Microbiol Biotechnol 2015 31 1 95 108 10.1007/s11274-014-1768-z 25371316 11. Edwards U Rogall T Blöcker H Emde M Böttger EC Isolation and direct complete nucleotide determination of entire genes. Characterization of a gene coding for 16S ribosomal RNA Nucleic Acids Res 1989 17 19 7843 53 10.1093/nar/17.19.7843 2798131 12. Woese CR Kandler O Wheelis ML Towards a natural system of organisms: proposal for the domains Archaea, Bacteria, and Eucarya Proc Natl Acad Sci U S A 1990 87 12 4576 9 10.1073/pnas.87.12.4576 2112744 13. Stackebrandt E Rainey FA Ward-Rainey NL Proposal for a new hierarchic classification system, Actinobacteria classis nov Int J Syst Bacteriol 1997 47 2 479 91 10.1099/00207713-47-2-479 14. Buchanan R Studies in the nomenclature and classification of the bacteria: II. The primary subdivisions of the schizomycetes J Bacteriol 1917 2 2 155 16558735 15. Pribram E A contributionto the classification of microörganisms J Bacteriol 1929 18 6 361 16559403 16. Altschul SF Gish W Miller W Myers EW Lipman DJ Basic local alignment search tool J Mol Biol 1990 215 3 403 10 10.1016/S0022-2836(05)80360-2 2231712 17. Busee HJ Review of the taxonomy of the genus Arthrobacter , emendation of the genus Arthrobacter sensu lato, proposal to reclassify selected species of the genus Arthrobacter in the novel genera Glutamicibacter gen. nov., Paeniglutamicibacter gen. nov., Pseudoglutamicibacter gen. nov., Paenarthrobacter gen. nov. and Pseudoarthrobacter gen. nov, and emended description of Arthrobacter roseus Int J Syst Evol Microbiol 2016 66 9 37 10.1099/ijsem.0.000702 26486726 18. Funke G Hutson RA Bernard KA Pfyffer GE Wauters G Collins MD Isolation of Arthrobacter spp. from clinical specimens and description of Arthrobacter cumminsii sp. nov. and Arthrobacter woluwensis sp. nov J Clin Microbiol 1996 34 10 2356 63 8880479 19. Borodina E Kelly DP Schumann P Rainey FA Ward-Rainey NL Wood AP Enzymes of dimethylsulfone metabolism and the phylogenetic characterization of the facultative methylotrophs Arthrobacter sulfonivorans sp. nov., Arthrobacter methylotrophus sp. nov., and Hyphomicrobium sulfonivorans sp. nov Arch Microbiol 2002 177 2 173 83 10.1007/s00203-001-0373-3 11807567 20. Heyrman J Verbeeren J Schumann P Swings J De Vos P Six novel Arthrobacter species isolated from deteriorated mural paintings Int J Syst Evol Microbiol 2005 55 4 1457 64 10.1099/ijs.0.63358-0 16014466 21. Chang H Bae J Nam Y Kwon H Park JR Shin K Arthrobacter subterraneus sp. nov., isolated from deep subsurface water of the South Coast of Korea J Microbiol Biotechnol 2007 17 11 1875 18092474 22. Kageyama A Morisaki K Ōmura S Takahashi Y Arthrobacter oryzae sp. nov. and Arthrobacter humicola sp. nov Int J Syst Evol Microbiol 2008 58 1 53 6 10.1099/ijs.0.64875-0 18175682 23. Ding L Hirose T Yokota A Four novel Arthrobacter species isolated from filtration substrate Int J Syst Evol Microbiol 2009 59 4 856 62 10.1099/ijs.0.65301-0 19329620 24. Reddy G Aggarwal R Matsumoto G Shivaji S Arthrobacter flavus sp. nov., a psychrophilic bacterium isolated from a pond in McMurdo Dry Valley, Antarctica Int J Syst Evol Microbiol 2000 50 4 1553 61 10.1099/00207713-50-4-1553 10939663 25. Zhang J, Ma Y, Yu H. Arthrobacter cupressi sp. nov., a novel actinomycete isolated from the rhizosphere soil of Cupressus sempervirens. Int J Syst Evol Microbiol. 2012. doi:10.1099/ijs.0.036889-0. 26. Tamura K Stecher G Peterson D Filipski A Kumar S MEGA6: molecular evolutionary genetics analysis version 6.0. Mol Biol Evol 2013 30 12 2725 9 10.1093/molbev/mst197 24132122 27. Tamura K Nei M Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees Mol Biol Evol 1993 10 3 512 26 8336541 28. Nichols CM Guezennec J Bowman J Bacterial exopolysaccharides from extreme marine environments with special consideration of the southern ocean, sea ice, and deep-sea hydrothermal vents: a review Mar Biotechnol 2005 7 4 253 71 10.1007/s10126-004-5118-2 16075348 29. Poli A Anzelmo G Nicolaus B Bacterial exopolysaccharides from extreme marine habitats: production, characterization and biological activities Marine Drugs 2010 8 6 1779 802 10.3390/md8061779 20631870 30. Shivaji S Prakash JS How do bacteria sense and respond to low temperature? Arch Microbiol 2010 192 2 85 95 10.1007/s00203-009-0539-y 20049417 31. Yamashita Y Nakamura N Omiya K Nishikawa J Kawahara H Obata H Identification of an antifreeze lipoprotein from Moraxella sp. of Antarctic origin Biosci Biotech Biochem 2002 66 2 239 47 10.1271/bbb.66.239 32. Chattopadhyay M Cold-adaptation of Antarctic microorganisms–possible involvement of viable but nonculturable state Polar Biol 2000 23 3 223 4 10.1007/s003000050030 33. Bennett EA Massilani D Lizzo G Daligault J Geigl E-M Grange T Library construction for ancient genomics: Single strand or double strand? BioTechn 2014 56 6 289 10.2144/000114176 34. Andrews S. FastQC: a quality control tool for high throughput data. Reference Source. 2010. 35. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014. doi:10.1093/bioinformatics/btu170. 36. Coil D, Jospin G, Darling AE. A5-miseq: an updated pipeline to assemble microbial genomes from Illumina MiSeq data. Bioinformatics. 2014. doi:10.1093/bioinformatics/btu661. 37. Galardini M Biondi EG Bazzicalupo M Mengoni A CONTIGuator: a bacterial genomes finishing tool for structural insights on draft genomes Source Code Biol Med 2011 6 11 10.1186/1751-0473-6-11 21693004 38. Cole JR Wang Q Cardenas E Fish J Chai B Farris RJ The Ribosomal Database Project: improved alignments and new tools for rRNA analysis Nucleic Acids Res 2009 37 suppl 1 D141 5 10.1093/nar/gkn879 19004872 39. Seemann T. Prokka: rapid prokaryotic genome annotation. Bioinformatics. 2014. doi:10.1093/bioinformatics/btu153. 40. Hyatt D Chen G-L LoCascio PF Land ML Larimer FW Hauser LJ Prodigal: prokaryotic gene recognition and translation initiation site identification BMC Bioinformatics 2010 11 1 119 10.1186/1471-2105-11-119 20211023 41. Bateman A Coin L Durbin R Finn RD Hollich V Griffiths‐Jones S The Pfam protein families database Nucleic Acids Res 2004 32 suppl 1 D138 41 10.1093/nar/gkh121 14681378 42. Tatusov RL Koonin EV Lipman DJ A genomic perspective on protein families Science 1997 278 5338 631 7 10.1126/science.278.5338.631 9381173 43. Lowe TM Eddy SR tRNAscan-SE: a program for improved detection of transfer RNA genes in genomic sequence Nucleic Acids Res 1997 25 5 0955 64 10.1093/nar/25.5.0955 44. Lagesen K Hallin P Rødland EA Stærfeldt H-H Rognes T Ussery DW RNAmmer: consistent and rapid annotation of ribosomal RNA genes Nucleic Acids Res 2007 35 9 3100 8 10.1093/nar/gkm160 17452365 45. Griffiths-Jones S Bateman A Marshall M Khanna A Eddy SR Rfam: an RNA family database Nucleic Acids Res 2003 31 1 439 41 10.1093/nar/gkg006 12520045 46. Krogh A Larsson B Von Heijne G Sonnhammer EL Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes J Mol Biol 2001 305 3 567 80 10.1006/jmbi.2000.4315 11152613 47. Bendtsen JD Nielsen H von Heijne G Brunak S Improved prediction of signal peptides: SignalP 3.0. J Mol Biol 2004 340 4 783 95 10.1016/j.jmb.2004.05.028 15223320 48. Grant JR Stothard P The CGView server: a comparative genomics tool for circular genomes Nucleic Acids Res 2008 36 suppl 2 W181 4 10.1093/nar/gkn179 18411202 49 Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, Harris MA, Hill DP, Traver LIKasarskis A, Lewis S, Matese JC, Richardson JE, Ringwald M, Rubin GM, Sherlock G. Gene Ontology: tool for the unification of biology. Nat Genet. 2000; 25(1):25–29.

PMC005xxxxxx/PMC5000429.txt

==== Front BMC Sports Sci Med RehabilBMC Sports Sci Med RehabilBMC Sports Science, Medicine and Rehabilitation2052-1847BioMed Central London 5110.1186/s13102-016-0051-zStudy ProtocolIs an in-home telerehabilitation program for people with proximal humerus fracture as effective as a conventional face-to face rehabilitation program? A study protocol for a noninferiority randomized clinical trial Cabana François francois.cabana@usherbrooke.ca 1Pagé Catherine catherine.page@usherbrooke.ca 2Svotelis Amy amy.svotelis@usherbrooke.ca 1Langlois-Michaud Samuel samuel.langlois-michaud@usherbrooke.ca 1Tousignant Michel michel.tousignant@usherbrooke.ca 21 Department of surgery, Faculty of Medicine and Health Sciences, Université de Sherbrooke, 3001 12e Avenue Nord, Sherbrooke, J1H 5N4 QC Canada 2 Research Centre on Aging, Centre intégré universitaire de santé et de services sociaux de l’Estrie - Centre hospitalier universitaire de Sherbrooke (CIUSSS de l’Estrie CHUS), 1036 Belvédère Sud, Sherbrooke, J1H 4C4 QC Canada 26 8 2016 26 8 2016 2016 8 1 2711 12 2015 17 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Proximal humerus fractures can be treated surgically (eg: pinning, plate and screws) or conservatively by wearing a splint or a cast. Following both of these approaches, rehabilitation has proven effective to prevent functional limitations and to re-establish normal shoulder function. However, access to these rehabilitation services and compliance tends to be limited in elderly patients due to travelling difficulties caused by their precarious health status and, in some cases, social and marital status. Since the majority of patients with a proximal humerus fracture are elderly, it becomes relevant to find a new way to offer quick, simple and suitable rehabilitation service. Thus, the use of promising alternative approaches, as in-home telerehabilitation, can enhance access to rehabilitation services for such population. The main objective of the study is to compare the clinical effects of the innovative telerehabilitation approach (TELE group) compared to face-to-face visits to a clinic (CLINIC group) for patients treated for a proximal humerus fracture. Methods/Design In this randomized controlled trial, individuals who have had a proximal humerus fracture treated conservatively at the Centre intégré universitaire de santé et de services sociaux de l’Estrie - Centre hospitalier universitaire de Sherbrooke (CIUSSS de l’Estrie CHUS), and who are returning home will be included. Participants will be recruited during their visit to the emergency ward or outpatient clinic by the medical or research team and will then sign the informed consent form if they are interested to participate in the study. We expect to recruit 52 participants (26 per group). Randomization will be done by a random number generator with sealed envelopes. Each patient will be evaluated before the beginning of the rehabilitation (T1), and immediately after the 2-month intervention (T2). The following outcomes will be measured: 1) upper extremity function (Constant Shoulder Score and Disability of the Arm, Shoulder and Hand questionnaire [DASH]); 2) range of motion (conventional goniometer); 3) user satisfaction (Health Care Satisfaction questionnaire); and 4) cost of services to the public healthcare system. The difference between the two groups will be compared using a t-test or a chi-squared test, and through a cost-effectiveness economic analysis. Discussion We hypothesize that in-home telerehabilitation will provide a good alternative to conventional rehabilitation, in terms of its efficacy, simplicity, patient satisfaction, and low associated costs. Trial registration ClinicalTrials.gov: NCT02425267. April 22nd, 2015. Keywords RehabilitationProximal humerus fractureTelerehabilitationEffectivenessissue-copyright-statement© The Author(s) 2016 ==== Body Background The aging population has placed in interesting pressure on our health system. In addition to different age-related conditions, the prevention and treatment of fragility fractures has become an important issue. Although osteoporosis treatments have improved in recent years, in addition to falls, the associated fractures are highly related to morbidity and mortality. Each year, one person out of three falls, and from them, 10 % causes a fracture [1]. The proximal humerus is one of the three most common fractured sites, along with the hip and wrist, making up 4 to 5 % of fractures, all sites combined [2]. Humerus fractures can be treated surgically (e.g. pinning, plate and screws, etc), or conservatively, typically through immobilisation in a splint or an orthotic device [3]. Conservative medical treatment tends to be advocated, mostly because of the complications that may occur following surgical intervention. However, the type of fracture, the surgeon, the patient and the fracture classification will influence the treatment decision. Among the existing proximal humerus fracture classifications (i.e. Neer [4], AO [5], Codman-Hertel [6], and Resch [7]), the Neer classification is the most widely used among orthopaedists. Neer group I and II fractures are often treated with the conservative approach, while group III and IV fractures most likely are surgical cases [8, 9]. Following both of these medical approaches, the individual needs rehabilitation to prevent functional limitations. Indeed, intervention programs in rehabilitation for humerus fractures have demonstrated their efficacy to improve shoulder joint mobility, diminish pain and improve functional state [10, 11]. Immobilisation time between the fracture and the beginning of rehabilitation is a parameter that has a significant influence on conservative treatment outcome [10–12], and, in reality, varies greatly between cases. A recent randomised clinical trial (RCT) [12] addressed the question of the efficacy of an immediate versus delayed mobilisation, showing that patients with a shorter immobilisation time (1 week) recovered faster functionally than patients with a longer immobilisation time. A similar study demonstrated that, in addition to a faster functional recovery, an immediate rehabilitation following the fracture leads to experiencing less pain for the patient [11]. Another study even claimed that a 3-day immobilisation, followed by passive mobilisation is sufficient and safe to restore physical capacities and post-fracture performance [10]. Finally, a RCT has compared the difference between the efficacy of a conventional physical therapy treatment and an individual in-home training program with instructions [13]. The results demonstrated no significant difference between treatment types. Interestingly, advantages emerged in the individual approach with instructions, such as: 1) allowing the patient to stay at home while receiving their rehabilitation, eliminating the need to travel to a clinic, and 2) patient satisfaction derived from having the responsibility of their rehabilitation at home. Presently, in Canada, rehabilitation for proximal humerus fractures occurs at the hospital (37.2 %) [14], in external clinics, and in various in-home services. However, not all patients have access to rehabilitation services, mainly due the difficulty of the elderly to travel to the clinic because of their precarious health condition [15]. Therefore, new service delivery strategies are essential to enhance accessibility to rehabilitation services, with a focus on in-home services. Telerehabilitation, defined as a telehealth application that uses telecommunication technologies to provide physical therapy services [16–19], is an interesting solution to the lack of available services. This approach allows the patient to receive rehabilitation at home, without the need for a health care professional to travel to the patient’s home or to travel to the clinic for the patient. As such, telerehabilitation could permit the elderly with a proximal humerus fracture to have access to rehabilitation directly in their home with adequate and safe supervision. Services provided by telerehabiliation can be very diverse, including telecheckups (phone calls to ensure the wellness of the patient), telemonitoring (record physiological data), teleconsultation (between two service centers) and teletreatment [19, 20]. This long-distance therapeutic intervention also includes the notion of frequency (a set number of sessions per week) and duration (over a specific time period) [21, 22]. A systematic review of telerehabilitation interventions [18] showed a tendency for these services to generate clinical improvements that are generally equal to those resulting from conventional rehabilitation programs. For example, efficacy of teletreatment has been demonstrated: in older populations with loss of autonomy [23], or living at home will mild to moderate dementia [24]; in individuals with mobility impairments [25]; and following total knee [26–29] or shoulder [30] arthroplasty. Interestingly, in the latest study [30], rehabilitation treatment was provided to 10 elderly patients through a videoconferencing system installed directly in their home. The intervention period length was 2 months and included supervised sessions, where a physical therapist was able to adapt exercises according to each patient’s evolution. Results of this study demonstrated that patients are able to adhere to teletreatment, and that they are able to maintain a good relationship with their physical therapists even though there was no direct physical interaction. Moreover, this remote method of receiving their rehabilitation facilitated their daily living since they were not required to travel to receive treatment. By staying home, patient motivation increased and they felt more independent in their exercises. However, an important limitation to this study was the lack of evaluation of clinical results. Therefore, although a promising avenue for rehabilitation, especially in older populations, the efficacy and functional results of telerehabilitation remain to be studied with high quality evaluations. To our knowledge, only one study exists on telerehabilitation among patients with a proximal humerus fracture [31]. This study, realised by our research team, demonstrated the feasibility of such a service delivery to this specific population. However, our previous study was more focussed on proof-of-principle, and as such, the lack of a control group (treatment at clinic) did not permit us to affirm that clinical improvement was due to the patient’s natural recovery or to the telerehabilitation received. Thus, the main objective of the present study is to evaluate the efficacy of in-home telerehabilitation (TELE group) compared to the conventional rehabilitation in a clinic (CLINIC group) in a population with a proximal humerus fracture. Our hypothesis is that in-home telerehabilitation will prove to be a good alternative to conventional rehabilitation. Methods/Design Study design In our study, teletreatment will be used to provide physical therapy sessions from a service center (clinic) directly into the patients’ home. The study design is a RCT. As described in Fig. 1, there will be an evaluation at baseline (T1), 1 to 2 weeks post-fracture. Then, the 8-week intervention phase (telerehabilitation) or conventional rehabilitation in a clinic will begin at the moment prescribed by the orthopaedist, approximately 2 to 3 weeks post-fracture. Finally, a second evaluation (T2) will be held at the end of the intervention period.Fig. 1 Study timeline. The patient is recruited post-fracture and evaluated at baseline (T1). Then, the participant is randomized into either Telerehabilitation group or Conventional rehabilitation group. Following 8 weeks of treatments, the patient is evaluated again (T2) Participants This study will be conducted in the population with a proximal humerus fracture treated conservatively at the Centre intégré universitaire de santé et de services sociaux de l’Estrie - Centre hospitalier universitaire de Sherbrooke (CIUSSS de l’Estrie CHUS). To be included, participants will have to: 1) return home after discharge from hospital or emergency; 2) be apt to do exercises; 3) have a sufficient verbal and written comprehension to participate to the treatment and evaluations; 4) have access to high speed internet connection at home. Participants with the following characteristics will be excluded from the study: 1) intra-articular proximal humerus fracture types (often susceptible to longer rehabilitation periods and a higher risk of complications); 2) presence of any other upper-limb fracture that can interfere with rehabilitation; and 3) surgical treatment following the fracture. A sample of 52 participants with a proximal humerus fracture and responding to the eligibility criteria will be included. Participants will be screened by the orthopaedic research team through the orthopaedic resident or nurse who proceeds to the installation of the removable splint. At that time, the research team will briefly inform the patient on the study and will obtain his/her authorisation to be contacted by phone by a member of the teletreatment research for a more detailed evaluation of eligibility and explanation of the study. Following this telephone interview, if the participant is still interested and meets all the eligibility criteria, an appointment will be scheduled to obtain the informed consent (approved by local ethic committee; CIUSSS de l’Estrie - CHUS, and perform the first evaluation (T1). Immediatly following visit T1, each patient will be randomised to either the TELE or the CLINIC group. The randomised list has been generated electronically using block randomization of size 4. The evaluator is the only one who will be blind to the randomisation. Independent variable: rehabilitation program The rehabilitation program is identical in both randomised groups, and will be dispensed by physical therapists of the Clinique universitaire de réadaptation de l’Estrie (CURE). Only the delivery mode will differ; TELE group (in-home telerehabilitation) or CLINIC group (conventional face-to-face rehabilitation in a clinic). The intervention consists of a rehabilitation program with constant qualified physical therapist supervision. The exercise program, based on a post-prosthesis and post-fracture rehabilitation program developed by the orthopaedic surgery division of the CIUSSS de l’Estrie - CHUS, includes stretching, pain management, range of motion and muscular strengthening, in addition to a question period. An example of a rehabilitation session is described in Table 1. The attending physical therapist will also adjust the exercises according to the progression of each patient’s condition (see Table 2 for the rehabilitation session progression).Table 1 Example of a rehabilitation session Length (minutes) Exercise types ≈5 à 10 Warm-up and stretching ≈15 à 30 According to progression • Weeks 1–2 : Pain management and range of motions • Weeks 3–4 : Range of motion renewal (active assisted) • Weeks 5 à 8 : Range of motion maintaining (active) and muscle strengthening ≈5 à 10 Question period Total : ≈ 30–45 End of the intervention Table 2 Rehabilitation session progression Exercices Weeksa 1 2 3 4 5 6 7 8 Circulatory movements x x Pendulum movements x x Wrist and elbow movements x x Thermal method (if necessary) x x x x x x x x Range of motion exercices x x x x x x Muscle strengthing x x x x aWeek 1 of rehabilitation matches to approximately week 3 post-fracture The training program consisting of 30 to 45-min sessions, which will be realised for 8 weeks at a frequency of twice daily, either supervised (TELE or CLINIC) or unsupervised at home. During weeks 1, 3 and 5, patients will have to perform their exercises twice with direct supervision of the physiotherapist and others without this supervision. For the other weeks (2, 4, 6, 7, 8), patients will only have one supervised sessions, and the others without supervision. Supervised sessions will allow both the therapist and the patient to adjust the program if a problem occurs and assure the proper execution of the exercises. Telerehabilitation technological platform The originality of the in-home telerehabilitation intervention stems from the superior level and type of interaction between users (health care professional and patient), which exceeds the usual use of a videoconferencing system. As such, the technological support must be flexible in order to respond to the needs and constraints of such a system. The telerehabilitation platform used in this study was developed in collaboration with Vigilent Telesystems to address these issues. Figure 2 describes the telerehabilitation platform used in the study.Fig. 2 Telerehabilitation technological platform. The patient and clinician systems include a 22″ touch monitor, a mini-PC (Intel NUC), a pan-tilt-zoom (PTZ) camera with embedded h264 video codec, a microphone array and a speaker. The telerehabilitation software, Vigil2, runs on both systems. The software includes functionalities for management (users, systems and sessions), patient status (online, offline, previous sessions, planned sessions), secure video, audio and data transfer over the Internet, and intuitive camera control (point-and-click control scheme). It also includes an easy way for the patient to turn on and off the system using the touch screen. Audio, video and sensor data coming from the patient’s home are transferred to the clinician using an application and database server over a secure link, allowing real-time sessions to occur Dependent variables Participants will attend a total of two evaluations of approximately 1.5 h each with a trained research assistant at the Research Centre on Aging before (T1) and after (T2) the intervention period. Each assessment will be executed in the same order to optimise the validity of the collected information: 1) range of motion (flexion, extension, abduction, external and internal rotation); 2) upper limb function; 3) global shoulder function and pain; and 4) participant satisfaction. Every item will be evaluated at both assessments, except for the satisfaction, which will be assessed only at T2. Range of motion Shoulder range of motion (flexion, extension, abduction, and internal and external rotations) will be assessed by a universal goniometer [32]. This standard, simple and reliable instrument used to measure angles has a degree-graduated scale. Active and passive measurements will be taken with standardised procedures. Upper limb function Shoulder function will be evaluated with the the Disability of the Arm, Shoulder and Hand questionnaire (DASH) [33]. This self-administered questionnaire includes 30 questions evaluated on a 5-point Likert scale, most of which relate to the individual’s capacity to realise a task. This tool was chosen for its scientific validity, ease of use, and ability to accurately reflect activity levels in daily living. The result is on a total of 100, where a high score indicates a greater disability. The original English version questionnaire demonstrated a good test-retest reliability (intraclass correlation coefficient = 0.95), a good internal consistency (Cronbach alpha = 0.96) and a moderate construct validity according to Spearman correlations varying from -0.58 to -0.76 [34]. However, factorial analyses demonstrated that five factors explained 67 % of total variance [35]. Shoulder functional measures Shoulder global function will be measured with the Constant score [36], which is the primary outcome measure. This questionnaire allows the assessment of four outcomes related to shoulder function: 1) pain; 2) activities of daily living (sleeping, work, leisure); 3) range of motion; and 4) muscle strength. The total is on 100 and a higher score indicates a higher shoulder function. According to a systematic review of the psychometric properties of the Constant score, the reliability of this questionnaire is excellent (ICC, 0.89; 95 % confidence interval, 0.79–0.94) [37], and the internal consistency, evaluated with the Cronbach alpha, ranges from 0.60 to 0.75 suggesting that it measures different aspects of function [38, 39]. Patients’ satisfaction Every patient will complete the French version of the validated Healthcare Satisfaction Questionnaire at T2 to evaluate their general satisfaction toward the health care service received [40]. This questionnaire also showed a good internal consistency (Cronbach’s alpha coefficient of the overall scale = 0.92) [40]. Satisfaction construct is determined by three distinct factors, whether satisfaction with the: 1) relationship with the healthcare professional; 2) services delivered; and 3) general healthcare organization. A score on 100 is computed for the general satisfaction, as well as for satisfaction of each domain. In addition, patients’ satisfaction toward teletreatment (only for the TELE group) will be measured after removal of the technology by the technician using the Telemedicine Perception Questionnaire [41]. This validated questionnaire includes 14 items placed on a five-point Likert scale (where 1 = totally disagree and 5 = totally agree) for a total score of 56. The first six items explore the communication quality between the professional and the patient, and the other items concentrate on the patient’s perception of the quality of the received service, including coherence, accessibility and needs met. Costs Economic analyses are based on the health system perspective [42]. A grid previously developed and already used to collected costs associated to teletreatment (TELE group) and conventional rehabilitation (CLINIC group) for our post-knee arthroplasty [43] will also be part of this study. Statistical methods The principal analyses intend to test the noninferiority of the TELE intervention versus the CLINIC intervention. We aim to recruit a total of 52 participants, or 26 participants per group. Sample size was determined with data from a previous study on the Constant score [44], which is considered in this study as the primary outcome measure. With 26 participants per group, within the context of a unilateral t-test with an alpha level set at 0.05, the power of the study to reject non-equivalence hypothesis would be at 69 % for a difference of 10.3 or less (standard deviation = 17). This noninferiority point is under the minimal clinically important difference of 10.4 points on the Constant score [44]. Participant characteristics in each group will be described pre-intervention (T1) using mean and standard deviation (continuous variables) or proportion (categorical variables). Groups at baseline will then be compared using t-test or chi-squared test. First, the data will be analysed according to the received intervention (per protocol), and then, according to the assigned group (intention to treat analysis). A sensitivity analysis will be used to explore the effect of compliance or screen failures: missing data will be replaced by extreme data (no change following intervention or most favorable change noted in the study). Any non-robustness of the revealed results by comparison of strategies will be noted. Economic analysis is of cost-efficacy type. Cost by change-unit of the principal dependant variable (Constant score) will be determined for both groups. Differential cost will then be established. Discussion This trial is the next logical step to the feasibility study conducted by our research team on 12 post-proximal humerus facture patients [31]. The previous study demonstrated that telerehabilitation seems to be a promising avenue to provide rehabilitation services to this population without adverse events. The results obtained following the present protocol will affirm the cost-efficacy of telereahabilitation treatment in an orthopaedic context, more precisely, after a proximal humerus fracture treated conservatively. In order to control a potential selection bias, randomisation will be blind to the evaluator and participants’ characteristics will be compared pre-intervention between each group. If a group differs on some characteristics despite randomisation, corrective measures will be made in subsequent analyses. Furthermore, information bias will be controlled by using standardized measures and by calibrating all the assessors for each assessment. This study will verify the noninferiority of in-home telerehabilitation compared to face-to-face intervention at a clinic for patients with proximal humerus fracture. In accordance with our hypothesis, we think that telerehabilitation will improve access to a rapid, less expensive, satisfactory and effective rehabilitation services. Trial status Recruitment has begun since June 2015. Abbreviations CIUSSS de l’Estrie CHUSCentre intégré universitaire de santé et de services sociaux de l’Estrie - Centre hospitalier universitaire de Sherbrooke CLINICConventional face-to-face rehabilitation in a clinic CUREClinique universitaire de réadaptation de l’Estrie DASHDisability of the Arm, Shoulder and Hand questionnaire RCTRandomised clinical trial T1Baseline evaluation (pre-intervention) T2Post-intervention evaluation TELEIn-home telerehabilitation Acknowledgements Not applicable. Funding This research project is supported by a grant received from the Department of surgery of the Université de Sherbrooke, as well as from the Chair in Telerehabilitation. Availability of data and materials Not applicable. Authors’ contributions All authors participated in the conception and design of the study. MT has the expertise in telerehabilitation and FC and SML in orthopeadic. CP and AS are responsible of the recruitment and coordination of the study. All authors read and approved the final version of the manuscript to be published. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate The ethical research committee of the CIUSSS de l’Estrie – CHUS (reference 2014-725) approved all procedures. Written informed consent will be obtained from all participants. ==== Refs References 1. Arcand M Hébert R Précis pratique de gériatrie 2007 3 Montreal EDISEM 2. Brorson S Olsen BS Frich LH Jensen SL Johannsen HV Sørensen AK Effect of osteosynthesis, primary hemiarthroplasty, and non-surgical management for displaced four-part fractures of the proximal humerus in elderly: a multi-centre, randomised clinical trial Trials 2009 10 51 10.1186/1745-6215-10-51 19586546 3. Handoll HH Ollivere BJ Interventions for treating proximal humeral fractures in adults Cochrane Database Syst Rev 2010 12 CD000434 21154345 4. Neer CS 2nd Displaced proximal humeral fractures: part I. Classification and evaluation Clin Orthop Relat Res 2006 442 77 82 10.1097/01.blo.0000198718.91223.ca 16394743 5. AO Foundation Müller AO Classification of fractures - Long bones 2010 6. Hertel R Hempfing A Stiehler M Leunig M Predictors of humeral head ischemia after intracapsular fracture of the proximal humerus J Shoulder Elbow Surg 2004 13 427 433 10.1016/j.jse.2004.01.034 15220884 7. Majed A Macleod I Bull AM Zyto K Resch H Hertel R Proximal humeral fracture classification systems revisited J Shoulder Elbow Surg 2011 20 1125 1132 10.1016/j.jse.2011.01.020 21481611 8. Murray IR Amin AK White TO Robinson CM Proximal humeral fractures: current concepts in classification, treatment and outcomes J Bone Joint Surg (Br) 2011 93 1 11 10.1302/0301-620X.93B1.25702 21196536 9. Neer CS 2nd Displaced proximal humeral fractures. I. Classification and evaluation J Bone Joint Surg Am 1970 52 1077 1089 5455339 10. Lefevre-Colau MM Babinet A Fayad F Fermanian J Anract P Roren A Immediate mobilization compared with conventional immobilization for the impacted nonoperatively treated proximal humeral fracture. A randomized controlled trial J Bone Joint Surg Am 2007 89 2582 2590 10.2106/JBJS.F.01419 18056488 11. Hodgson SA Mawson SJ Stanley D Rehabilitation after two-part fractures of the neck of the humerus J Bone Joint Surg (Br) 2003 85 419 422 10.1302/0301-620X.85B3.13458 12729121 12. Hodgson SA Mawson SJ Saxton JM Stanley D Rehabilitation of two-part fractures of the neck of the humerus (two-year follow-up) J Shoulder Elbow Surg 2007 16 143 145 10.1016/j.jse.2006.06.003 17399621 13. Bertoft ES Lundh I Ringqvist I Physiotherapy after fracture of the proximal end of the humerus. Comparison between two methods Scand J Rehabil Med 1984 16 11 6 6710092 14. Lübbeke A Stern R Grab B Herrmann F Michel JP Hoffmeyer P Upper extremity fractures in the elderly: consequences on utilization of rehabilitation care Aging Clin Exp Res 2005 17 276 280 10.1007/BF03324610 16285192 15. Burdea G Popescu V Hentz V Colbert K Virtual reality-based orthopedic telerehabilitation IEEE Trans Rehabil Eng 2000 8 430 432 10.1109/86.867886 11001524 16. Romanow R Building on Values: The Future of Health Care in Canada 2002 Ottawa Santé Canada 17. Georgeadis AC Brennan DM Barker LM Baron CR Telerehabilitation and its effect on story retelling by adults with neurogenic communication disorders Aphasiology 2004 18 639 652 10.1080/02687030444000075 18. Kairy D Lehoux P Vincent C Visintin M A systematic review of clinical outcomes, clinical process, healthcare utilization and costs associated with telerehabilitation Disabil Rehabil 2009 31 427 447 10.1080/09638280802062553 18720118 19. Pineau G, Moqadem K, St-Hilaire C, Levac E, Hamel B. Télésanté : Lignes directrices cliniques et normes technologiques en téléréadpatation, in Agences des technologies et des modes d’intervention en santé. 2006, Gouvernement du Québec: Québec, PQ. Available at: https://www.inesss.qc.ca/fileadmin/doc/AETMIS/Rapports/Telesante/ETMIS2006_Vol2_No3.pdf 20. Joseph AM Care coordination and telehealth technology in promoting self-management among chronically ill patients Telemed J E Health 2006 12 156 159 10.1089/tmj.2006.12.156 16620170 21. Winters JM Telerehabilitation research: emerging opportunities Annu Rev Biomed Eng 2002 4 287 320 10.1146/annurev.bioeng.4.112801.121923 12117760 22. Bashshur R Shannon G Krupinski E Grigsby J The taxonomy of telemedicine Telemed J E Health 2011 17 484 494 10.1089/tmj.2011.0103 21718114 23. Hoenig H Nusbaum N Brummel-Smith K Geriatric rehabilitation: state of the art J Am Geriatr Soc 1997 45 1371 1381 10.1111/j.1532-5415.1997.tb02939.x 9361665 24. Poon P Hui E Dai D Kwok T Woo J Cognitive intervention for community-dwelling older persons with memory problems: telemedicine versus face-to-face treatment Int J Geriatr Psychiatry 2005 20 285 286 10.1002/gps.1282 15717335 25. Hoenig H Sanford JA Butterfield T Griffiths PC Richardson P Hargraves K Development of a teletechnology protocol for in-home rehabilitation J Rehabil Res Dev 2006 43 287 298 10.1682/JRRD.2004.07.0089 16847794 26. Russell TG Buttrum P Wootton R Jull GA Low-bandwidth telerehabilitation for patients who have undergone total knee replacement: preliminary results J Telemed Telecare 2003 9 Suppl 2 S44 S47 10.1258/135763303322596246 14728759 27. Russell T Buttrum P Wootton R Jull G Rehabilitation after total knee replacement via low-bandwidth telemedicine: the patient and therapist experience J Telemed Telecare 2004 10 Suppl 1 85 87 10.1258/1357633042614384 15603622 28. Tousignant M Moffet H Boissy P Corriveau H Cabana F Marquis F A randomized controlled trial of home telerehabilitation for post-knee arthroplasty J Telemed Telecare 2011 17 195 198 10.1258/jtt.2010.100602 21398389 29. Moffet H Tousignant M Nadeau S Merette C Boissy P Corriveau H In-Home Telerehabilitation Compared with Face-to-Face Rehabilitation After Total Knee Arthroplasty: A Noninferiority Randomized Controlled Trial J Bone Joint Surg Am 2015 97 1129 1141 10.2106/JBJS.N.01066 26178888 30. Eriksson L Lindström B Ekenberg L Patients’ experiences of telerehabilitation at home after shoulder joint replacement J Telemed Telecare 2011 17 25 30 10.1258/jtt.2010.100317 21075802 31. Tousignant M Giguere AM Morin M Pelletier J Sheehy A Cabana F In-home telerehabilitation for proximal humerus fractures: a pilot study Int J telerehabil 2014 6 31 37 10.5195/ijt.2014.6158 25945227 32. Norkin CC White DJ Measurement of joint motion: a guide to goniometry 1995 Philadelphia F.A. Davis Company 33. Beaton D Davis A Hudak P McConnell S The DASH (Disabilities of the Arm, Shoulder and Hand) Outcome Measure: What do we know about it now? British J Hand Therapy 2001 64 109 118 34. Slobogean GP Noonan VK O’Brien PJ The reliability and validity of the Disabilities of Arm, Shoulder, and Hand, EuroQol-5D, Health Utilities Index, and Short Form-6D outcome instruments in patients with proximal humeral fractures J Shoulder Elbow Surg 2010 19 342 348 10.1016/j.jse.2009.10.021 20189839 35. Fayad F Lefevre-Colau MM Macé Y Fermanian J Mayoux-Benhamou A Roren A Validation of the French version of the Disability of the Arm, Shoulder and Hand questionnaire (F-DASH) Joint Bone Spine 2008 75 195 200 10.1016/j.jbspin.2007.04.023 17983829 36. Constant CR Murley AH A clinical method of functional assessment of the shoulder Clin Orthop Relat Res 1987 214 160 164 3791738 37. Roy JS MacDermid JC Woodhouse LJ A systematic review of the psychometric properties of the Constant-Murley score J Shoulder Elbow Surg 2010 19 157 164 10.1016/j.jse.2009.04.008 19559630 38. Razmjou H Bean A Macdermid JC van Osnabrugge V Travers N Holtby R Convergent validity of the constant-murley outcome measure in patients with rotator cuff disease Physiother Can 2008 60 72 79 10.3138/physio/60/1/72 20145743 39. Livain T Pichon H Vermeulen J Vaillant J Saragaglia D Poisson MF Intra- and interobserver reproducibility of the French version of the Constant-Murley shoulder assessment during rehabilitation after rotator cuff surgery Rev Chir Orthop Reparatrice Appar Mot 2007 93 142 149 10.1016/S0035-1040(07)90217-2 17401287 40. Gagnon M Hebert R Dube M Dubois MF Development and validation of the Health Care Satisfaction Questionnaire (HCSQ) in elders J Nurs Meas 2006 14 190 204 10.1891/jnm-v14i3a004 17278339 41. Yip MP Chang AM Chan J MacKenzie AE Development of the Telemedicine Satisfaction Questionnaire to evaluate patient satisfaction with telemedicine: a preliminary study J Telemed Telecare 2003 9 46 50 10.1258/135763303321159693 12641893 42. Drummond MF O’Brien B Stoddart GL Torrance GW Methods for the economic evaluation of health care programmes 1997 2 Oxford Oxford University Press 43. Tousignant M Moffet H Nadeau S Merette C Boissy P Corriveau H Cost analysis of in-home telerehabilitation for post-knee arthroplasty J Med Internet Res 2015 17 e83 10.2196/jmir.3844 25840501 44. Kukkonen J Kauko T Vahlberg T Joukainen A Aarimaa V Investigating minimal clinically important difference for Constant score in patients undergoing rotator cuff surgery J Shoulder Elbow Surg 2013 22 1650 1655 10.1016/j.jse.2013.05.002 23850308

PMC005xxxxxx/PMC5000430.txt

==== Front BMC Public HealthBMC Public HealthBMC Public Health1471-2458BioMed Central London 354410.1186/s12889-016-3544-5Research ArticleWorkforce capacity to address obesity: a Western Australian cross-sectional study identifies the gap between health priority and human resources needed http://orcid.org/0000-0002-5448-8932Begley Andrea +61 8 9266 2773a.begley@curtin.edu.au 1Pollard Christina Mary c.pollard@curtin.edu.au 121 School of Public Health, Curtin University, Kent Street, GPO Box U1987, Perth, 6845 WA Australia 2 Department of Health in Western Australia, 189 Royal Street, East Perth, 6004 WA Australia 25 8 2016 25 8 2016 2016 16 1 88131 8 2015 18 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background The disease burden due to poor nutrition, physical inactivity and obesity is high and increasing. An adequately sized and skilled workforce is required to respond to this issue. This study describes the public health nutrition and physical activity (NAPA) practice priorities and explores health managers and practitioner’s beliefs regarding workforce capacity to deliver on these priorities. Methods A workforce audit was conducted including a telephone survey of all managers and a postal survey of practitioners working in the area of NAPA promotion in Western Australia in 2004. Managers gave their perspective on workforce priorities, current competencies and future needs, with a 70 % response rate. Practitioners reported on public health workforce priorities, qualifications and needs, with a 56 % response rate. Results The top practice priorities for managers were diabetes (35 %), alcohol and other drugs (33 %), and cardiovascular disease (27 %). Obesity (19 %), poor nutrition (15 %) and inadequate physical activity (10 %) were of lower priority. For nutrition, managers identified lack of staff (60.4 %), organisational and management factors (39.5 %) and insufficient financial resources (30.2 %) as the major barriers to adequate service delivery. For physical activity services, insufficient financial resources (41.7 %) and staffing (35.4 %) and a lack of specific physical activity service specifications (25.0 %) were the main barriers. Practitioners identified inadequate staffing as the main barrier to service delivery for nutrition (42.3 %) and physical activity (23.3 %). Ideally, managers said they required 152 % more specialist nutritionists in the workforce and 131 % specialists for physical activity services to meet health outcomes in addition to other generalist staff. Conclusion Human and financial resources and organisational factors were the main barriers to meeting obesity, and public health nutrition and physical activity outcomes. Services were being delivered by generalists rather than specialists, which may reduce service effectiveness. Although conclusions from this research need to take into account the fact that the audit was conducted in 2004, the findings suggest that there was a need to equip health services with an adequately skilled workforce of sufficient capacity to deliver an effective public health response to the obesity epidemic, particularly addressing poor nutrition and physical inactivity. Keywords WorkforceNutritionPhysical activityCapacityhttp://dx.doi.org/10.13039/501100000960Healthway (AU)19986Pollard Christina Mary http://dx.doi.org/10.13039/501100006065Department of Health, Government of Western Australia (AU)issue-copyright-statement© The Author(s) 2016 ==== Body Background The increasing prevalence of obesity and non-communicable chronic disease in Australia requires a range of actions and interventions to enable effective prevention policy and programs [1]. The health and economic costs of poor nutrition and physical inactivity contributing to obesity are greater than that of smoking and harmful and hazardous alcohol consumption [2]. Healthy eating and regular physical activity at any age can substantially protect against weight gain, obesity and diet-related chronic illness, and therefore reduce preventable chronic disease and associated healthcare costs [3]. It is acknowledged that public health services designed to improve NAPA are essential to reduce the increasing prevalence of chronic disease [4]. Effective interventions require sufficiently sized and skilled workforce to achieve prevention targets [5]. An appropriately trained workforce to implement healthy eating and physical activity disease prevention strategies is a priority public health infrastructure needed to impact on rising obesity rates [6]. It is not easy to quantify the size of the workforce required but there is no doubt that an appropriate workforce will have a profound impact on the ability to achieve effective outcomes [7]. A critical mass in workforce is required for effective service delivery [8]. To foster workforce adequacy there is a need to firstly consider workforce development through appropriate training and curriculum and secondly to consider the existing workforce capacity to design and deliver effective obesity prevention programs including planning considerations to address future challenges. Australian public health policy asserted that a range of professionals in public and primary health are required to support population and community based activities and indicated that public health nutritionists and health promotion officers specializing in physical activity are important health professionals to deliver these services [1]. Research suggests that the prevention workforce in other countries is lacking practitioners with specific skills and responsibility for effective public health NAPA action [9, 10]. Little is known about Australia’s obesity prevention workforce or the public health workforce more broadly. However, there has been concern since 2009 that the level of capacity in the specialist obesity prevention workforce is lacking in most jurisdictions including local government, state government and non-government organisations across Australia [1]. It is likely that the promotion of healthy eating and physical activity is relegated to general staff with lack of additional resources and variable levels of training and/or there is a lack of service delivery. The lack of workforce capacity has been identified as the result of several factors including a lack of specific workforce development efforts and workforce effectiveness associated with population health outcomes [11]. Public health nutrition is a discipline defined as the promotion and maintenance of nutrition-related health and wellbeing of populations through organised efforts and informed choices of society [12, 13]. Workforce development is a key strategic domain for building capacity for public health nutrition practice therefore it has been necessary to define the role and scope of the workforce and the competencies required [9, 14–16]. There is international agreement that whilst public health work is multi-sectorial and multidisciplinary, the most effective programs to achieve public health nutrition goals are those facilitated by a specialist workforce identified by specific competencies [17]. Australia’s 10 year national agenda for action for public health nutrition, Eat Well Australia 2001–10, provided the mandate for capacity building priorities to consider workforce development as a central strategy [18]. Global efforts for public health action in physical activity have also recognised the opportunistic nature of past workforce development and the recurrent need for systematic workforce development [19]. The International Society for Physical Activity and Health was formed in 2009 with a view to moving physical activity to mainstream public health services [20]. The physical activity workforce broadly includes practitioners from health, education, sport and recreation, planning, transport and other disciplines such as medicine [21]. Whilst the broad range of sectors involved can be mobilised to engage in physical activity promotion, the variability in knowledge, skills and training may hinder population based program development efforts. The range of programs provided include examples such as the medicalisation of physical activity risk to exercise physiology where athletic performance is the target, or physiotherapists with rehabilitation as their target [21]. The public health physical activity workforce is emerging with specific positions created, however, there is an imperative to develop a physical activity promotion workforce across a range of disciplines [22]. Little is known about the priority placed on NAPA public health programs or the workforce size needed to support effective efforts to build workforce capacity [23]. As the policy environment continues to focus on reducing obesity in Australia there is an urgent need to profile the obesity prevention workforce. The composition, practice methods, resource allocation and organisation issues are all likely to impact on workforce capacity to address obesity. An audit of the NAPA workforce was carried out in Western Australia in 2004 to explore the policy environment and future workforce needs. This audit was commissioned by the Nutrition and Physical Activity Branch (NPAB) of the Department of Health in Western Australia. The Western Australian Health Promotion Foundation (Healthway) funded Curtin University’s Food Law, Policy and Communications to Improve Public Health Research Translation Project to enable results to be published. The specific objectives of the audit were to describe the current priorities for NAPA and workforce structure of the WA NAPA workforce, as determined by health managers and practitioners. This paper reports on the 2004 workforce audit to determine the appropriateness of priorities and size of the workforce to meet the challenges of addressing obesity prevention as an important function of workforce capacity. These results are significant as they are the only workforce data for both workforce areas to be published for Australia and the findings enable the retrospective exploration of factors impacting on workforce capacity and development in relation to policy directives so as to inform future strategies. Methods NAPA services were defined for the purposes of the audit as any service offered in the form of education, program delivery, community or policy development that seeks to improve the food intake and physical activity levels of specific target groups or the population in general. The audit consisted of two surveys; the first was a telephone survey of managers of the obesity prevention workforce in NAPA services and the second a postal survey of the existing workforce (practitioners). Workforce definitions To describe the current workforce it was necessary to elucidate the types of workforce, with a variety of qualifications currently employed in obesity prevention. Workforce definitions describing different paradigms in the nutrition workforce were used as the basis to describe key workforce areas for consideration. Workforce positions in public health nutrition, community nutrition or dietetics and clinical dietetics formed the specialist nutrition workforce. In Australia, people working in these positions would have Bachelor and/or postgraduate university nutrition and/or dietetics qualifications. It is expected that other professionals working in health, such as health promotion officers and Aboriginal health workers, would also have some role in the delivery of nutrition services. This workforce may have little or no training in nutrition but be experts in other areas, for example health promotion program delivery. For the purpose of this audit this section of the workforce is described as the generalist nutrition workforce. Detailed descriptions of the specialist and generalist nutrition workforce to represent a spectrum of workforce, as adapted from Hughes and Somerset (1997) [24]. These descriptions were then applied to definitions of the physical activity workforce as no previous literature had identified a taxonomy for defining that workforce at the time of the survey. Definitions of delineated service delivery describing the different features of methods and processes were also adapted and defined for the purpose of the study [8]. Community and Public Health delivery are usually differentiated by intended reach, prevention level, and the wellness/or illness paradigm for operation in Australia [24]. One way to consider workforce was to differentiate between the multiple workforce tiers by the determinant driving the service delivery. Determinants such as community development, needs assessment and policy directives indicate that different workforce competencies are required for community and public health NAPA approaches. Questionnaire development Separate manager and practitioner surveys were developed to measure the research objectives based on previous survey’s including an unpublished state government Review of Allied Health Professionals Recruitment and Retention Taskforce Survey (1999), the Dietitians Association of Australia’s professional competencies [25], general health promotion competencies [26] and a public health nutrition workforce development study [27]. Questions selected aimed to measure priority placed on NAPA services based on required service reporting areas and were mostly closed ended. The Department of Health required service reporting areas and potential national health priority and target areas were listed and managers could select those applicable. Other questions required the enumeration of specialist and generalist workforce and perceptions of current workforce in relation to adequacy, competency and training needs and perceived ability to meet NAPA service goals. A workforce profiling was conducted using the position title, fractional appointment and location of specialist and generalist workforce and description of services provided. Current workforce and future workforce requirements were then calculated for each region and totalled for the state. The practitioner postage survey included the questions described above with additional details on years working in their current position, methods of training and continuing professional development and perceived barriers to service delivery. Both questionnaires were developed in conjunction with NPAB staff for content validity and were piloted on university staff for comprehensibility and face validity. Ethics approval was granted from Curtin University’s Human Research Ethics Committee. All participants signed consent to participate and data was anonymised and aggregated for regions and then the state. Confidentially was maintained at all times and all participants consented to the publication of the results in various formats for the Department of Health’s purposes. Recruitment Western Australia is a geographically large state (2,532,400 square kilometres) with a population in 2004 estimated to be just under 2 million [28]. There were four metropolitan government health regions including public and community health, seven regional public and community health units and several non-government organisations and welfare organisations involved in prevention service delivery at the time of the survey. Fifteen medical general practices were also organised in geographic areas across the state with a mandate for promoting NAPA [29]. Managers were defined as a person who directly or indirectly line managed practitioner/s that have a functional responsibility to deliver nutrition and/or physical activity services for an area/region or organisation in community, public and population health. The term ‘services’ was used to broadly cover interventions and strategies designed to improve the risk factors of interest (public health NAPA). During the audit several revisions were made to the recruitment list due to restructuring and people on leave or acting in positions, with 69 managers identified by the end of the survey period. An email was sent to managers with an introductory letter explaining the aim of the audit including research consent and a copy of the questionnaire and workforce descriptions. Managers were asked to respond with details of their current NAPA workforce and a telephone interview was arranged to complete the questionnaire and elicit other comments regarding the workforce. A $20 gift voucher was sent at the completion of the interviews as an incentive to encourage a high response rate. Manager’s interviews were carried out over 3 months and lasted between 30 and 60 min. All interviews were carried out by the primary author. Practitioners were defined as a person who delivers nutrition and/or physical activity services as part of their employment. Practitioners were identified from contact lists of the NPAB and professional organisation mailing lists. As well as the original lists, a snowball approach was used to identify additional practitioners by asking survey participants to nominate other specialist or generalist practitioners. All 185 practitioners identified at the start of the survey were mailed an introductory letter, research consent form and questionnaire with replied paid envelope. They were also asked to send a copy of their job description form outlining organisational structure, key duties and competencies required for the position. Analysis Responses to closed-ended questions were coded directly onto the questionnaire and responses to open-ended questions were summarised and then coded according to a pre-established coding protocol developed after the interviews. Both sets of questionnaires were analysed using SPSS version 11 (SPSS Inc., Chicago, IL, USA), using descriptive statistics and chi-square test of association to assess relationships between data. . Results Forty eight managers were interviewed (a 70 % response rate) and 101 of the 185 practitioners identified participated (a 56 % responses rate). The representative spread across all WA health regions and organisations enabled enumeration of the current NAPA workforce. Demographic characteristics Over half of the managers (55.8 %) had been in their current position for 2 years or less. Their main service delivery was in population services (37.5 %), community and clinical (29 %), solely community (25 %), public health (6.3 %) and clinical only (2.1 %). The majority of managers (62.5 %) were located in country areas with 87.5 % having regional service delivery. There was variability in the highest qualification held, with only 10.4 % having attained a Master of Public Health qualification. NAPA practitioners were mostly female (97 %) with a mean age of 36.5 years. Most (90 %) delivered nutrition services and 54.5 % delivered physical activity services. The nutrition workforce was more experienced, 41 % had over 10 years’ experience compared to 8.9 % of the physical activity practitioners. Most practitioners (76.2 %) had nutrition and/or dietetic qualifications, 4.9 % had health promotion qualifications and 13.8 % had diabetes educator qualifications. The main employers were the Department of Health (70.3 %) and nongovernment organisations (9.9 %) and the remainder from private business. Two thirds (65.4 %) of practitioners were employed in the metropolitan area reflecting the population distribution. Services and health priority All managers had some responsibility for nutrition and/or physical activity service delivery. Table 1 shows that managers rated NAPA services as priority service delivery areas along with many other competing priorities, particularly in regional areas where alcohol and other drugs and injury prevention (including assault & suicide) were rated higher. Diabetes (35.4 %), reducing harm from alcohol and other drugs (33.3 %), cardiovascular disease (27.1 %) and injury prevention (25 %) were the priority health risks. As key risk factors for chronic disease, poor nutrition was ranked 11th and inadequate physical activity 13th in priorities.Table 1 Managers self-reported major health issues for their regions/organisations (n = 48) Major Health Issue % (n = 48) Diabetes 35.4 Drugs & Alcohol 33.3 Cardiovascular Disease 27.1 Injury, Assault, Suicide 25.0 National Health Priority Areasa 22.9 Mental Health 20.1 Maternal and Child Health 20.1 Social Impacts/Socioeconomic Status 18.8 Obesity 18.8 Indigenous Health 16.6 Poor Nutrition 14.6 Lifestyle Risk Factors 12.5 Inadequate Physical Activity 10.4 Smoking 10.4 Cancer 8.3 Asthma 8.3 Renal 6.2 aAustralia’s seven national health priority areas recognised by government in 2004 as Cardiovascular Health, Cancer Control, Diabetes Mellitus, Injury Prevention and Control, Mental Health, Arthritis and Musculoskeletal Conditions; Asthma (http://www.aihw.gov.au/national-health-priority-areas/) The health issues reflected in the ranking of the top five intervention strategies used by managers for their region or organisation. Eight key interventions were predetermined based on expected Department of Health service reporting and the top five listed by managers were improving physical activity (75 %); improving nutrition (70.8 %); capacity building (68.7 %); reducing drugs and alcohol (68.7 %) and addressing obesity (62.5 %). Indigenous people were key target areas identified by three quarters of managers for NAPA services. The second key target areas for managers were women and children however the focus for practitioners were adults in general for both areas of service delivery. Size and type of NAPA workforce One quarter of managers had no direct management of positions that were involved in physical activity service delivery and 10 % had no direct responsibility for staff delivering nutrition services. Table 2 shows the 18 different job titles identified as delivering nutrition services. The total specialist nutrition workforce was estimated to be 53.1 full time equivalents (FTE) state-wide or 9 % of the total workforce with the majority having a dietetic qualification as reflected by job descriptions. Practitioners who identified with community delivery roles also had position descriptions that required delivering clinical services (35 %). The majority of managers’ capacity to deliver nutrition services fell to a generalist workforce of Aboriginal health workers and community nurses without explicit public health or community nutrition skills in their job descriptions (528.8FTE in total).Table 2 Types of Positions Responsible for Delivering NAPA Services under direct supervision by Managers (FTE)) Type Job Description % (n = 43) FTE % FTE of Total Nutrition Workforce Specialist Workforce Community/Clinical Dietitians 32.6 18.4 8 % Community Dietitians 20.9 13.9 Public Health Nutritionist 16.3 6.2 Clinical Dietitians 7.0 2.5 Nutrition Co-ordinators 9.3 4.0 Population Health Nutritionist 2.3 1.0 Community Nutritionist 0 0 TOTAL Specialist FTE 46.0 FTE Generalist Workforce Aboriginal Health Workers 62.7 88.0 92 % Nurses 48.8 371.0 Health Promotion Officers/Project Officers 44.0 25.0 Diabetes Educators 46.5 21.2 Project Officers 13.9 12.0 CVD Coordinators 4.6 2.0 Chronic Disease Co-ordinators 4.6 2.0 Health Advancement Officers 2.3 0.6 Research Officers 2.3 1.0 Secondary Prevention manager 2.3 1.0 Early Intervention Staff 2.3 1.0 Liaison Officer 2.3 1.0 TOTAL Generalist FTE 525.8 FTE Department of Health Head Office Project Officers 7.1 TOTAL FTE 578.9 FTEa Job Description % (n = 36) FTE % FTE Specialist Physical Health Promotion Officer 58.3 43.5 Activity Workforce Physical Activity Co-ordinators 8.3 2.5 14 % TOTAL Specialist 46.0 FTE Community Physiotherapists 50.0 38.0 Nurses 44.4 152.0 Aboriginal Health Workers 36.1 54.0 Project Officer 22.2 14.0 86 % Chronic Disease Co-ordinators 12.5 8.0 Generalist Physical Public Health Nutritionists 11.1 3.6 Activity Workforce Community Dietitians 11.1 6.0 Diabetes Co-ordinator 11.1 4.0 Therapy Assistant 8.3 3.0 Clinical Dietitians 5.5 2.0 Researcher 5.5 2.0 Occupational Therapist 2.8 1.0 Population Health Nutritionist - - Community Nutritionist - - TOTAL Generalist 287.6 Department of Health –Head Office Physical activity project officer 1.5FTE TOTAL FTE 335.1 FTEb a 2 managers unable to estimate FTE b 2 managers unable to estimate FTE The majority of physical activity services were delivered by health promotion officers, community physiotherapists, nurses and/or Aboriginal health workers in a preventive role (see Table 2). The specialist workforce was estimated at 47.5 FTE or 14 % of the total physical activity workforce, and the general physical activity workforce was estimated to be 335.1 FTE. NAPA service delivery Managers and practitioners were in agreement about the achievement of service delivery against policy goals or strategic plan objectives. Few managers (4.3 %) and practitioners (5.9 %) thought that physical activity goals were being met (%) while 10.4 % of managers and 9.0 % of practitioners indicated nutrition goals were being met. Implications of not meeting goals including the recognition that services were stretched, and the limited ability to use capacity building or community development approaches to respond to the issues and lack of ability to service disadvantaged groups. The major barriers to full nutrition service delivery identified by managers was a lack of staff (60.4 %), organisational and management factors (39.5 %) and financial resources (30.2 %). The major barriers for full physical activity service delivery were financial (41.7 %), lack of staff (35.4 %) and physical activity not being clearly identified in service specifications (25.0 %). Recruitment and retention of staff to deliver nutrition services were barriers to service delivery reported by managers, particularly in relation to attracting staff to regional areas (20.8 %) and staff burn out (10.4 %). Lack of funding (14.5 %) and the limited number of dietetics trained professionals applying for public health nutrition (PHN) positions (14.5 %) were also considered barriers to delivering nutrition services. There were similar issues to the recruitment and retention of staff to deliver physical activity services, however physical activity was viewed by some managers (10.4 %) as being a newer or untested area for service delivery. Future workforce requirements Three quarters of managers said more staff were needed to fully deliver on nutrition service goals, particularly from specialist workforce. An additional 81FTE of specialist workforce (152 % more) and 62FTE (12 % more) of generalist workforce such as health promotion officers was identified as necessary which included filling currently vacant positions. Ideally, the additional specialist workforce would be dietitians (45 %), health promotion officers (17 %), and public health nutritionists (13 %). Figure 1 illustrates the comparison between the current workforce and the estimated additional specialist workforce required by managers to fully deliver on nutrition service goals.Fig. 1 Comparison between current and additional specialist and generalist NAPA workforce required to fully meet goals In relation to full physical activity service delivery, the majority of managers said that an additional 56.6 FTE (131 % more) of specialist physical activity workforce and 52FTE (16 % more) from generalist workforce was required including filling currently vacant positions. Discussion The 2004 WA nutrition and physical activity (NAPA) workforce audit described and quantified the priority and capacity for service delivery from a public health perspective. Even though NAPA are key risk factors for preventable chronic disease and obesity they were considered a low service delivery priority in 2004. Broad policy priorities did not always reflect practice priorities, particularly in regional areas. Increasing decision makers’ awareness of the health, economic and social benefits of improving NAPA appears to be warranted. Human and financial resources were identified as major weaknesses in health service delivery only 9 % of positions responsible for delivering nutrition services occupied by suitably qualified personnel. Most managers and practitioners believed they were ‘not or only partially’ meeting NAPA service delivery requirements, suggesting a reduced or stretched service delivery primarily due to a lack of specialist workforce. Organisational and managerial workforce support Organisational and managerial support directed the services provided as well as mandated requirements by the state based Department of Health and/or other organisations. Manager’s focus was on the seven chronic disease outcomes reflected in Government policy priorities at the time. The program delivery focus in WA at the time was promoting increased fruit and vegetable consumption with the Go for 2&5® social marketing campaign [30, 31]. In some instances other immediate local issues, for example, reducing alcohol and other drug usage were higher priorities than poor nutrition, physical inactivity or obesity. The policy priority of preventing obesity continues to increase [32], as does the need for an appropriately sized and skilled public health and primary health care workforce to deliver programs [18]. In 2004, addressing obesity was approached by encouraging employers to ensure a healthy workforce rather than building the workforce to implement actions to improve diet and physical activity [18]. Australia’s public health nutrition strategic plan of action, Eat Well Australia, expressed uncertainty about whether the current workforce was large enough to undertake the tasks required to address obesity and highlighted the lack of a specific workforce development strategy [18]. The first action “Investigating workforce requirements, including training needs and the systems necessary to deliver activities in light of current funding arrangements, workforce capacity and composition” was never undertaken ([18]:26). The policy priority assigned to specific health issues has the potential to limit service delivery. Unsupported low priority issues result in an undersized and unqualified workforce or alternatively, an undersized and underqualified workforce can influence the priority managers placed on the health issue and subsequent service delivery because they have limited capacity to act. Addressing poor nutrition is complex, there are multiple stakeholders and numerous dietary targets (e.g., increasing fruit and vegetable consumption) and approaches needed [8]. Managers’ and practitioners’ opinions differed in regard to meeting NAPA expectations with potential misalignment between practice and the work needed. The Indigenous population was an important target for managers yet practitioners focussed on adults in general; suggesting that disadvantaged groups, with great health need could be left out of service delivery. Workforce profiling A specialist workforce is critical to obesity prevention program success [18]. Findings showed an urgent need to increase the size of the specialist NAPA workforce in WA to develop the critical mass of human resources required. Managers estimated 152 % more specialist nutrition and 131 % more specialist physical activity workforce was required to achieve policy/program goals. The findings are consistent with research in California which found 70 % of local public health department managers rated their staff capacity for obesity prevention in NAPA environments as less than effective [33]. Benchmarking the recommendations for staffing public health in NAPA areas is limited. The type of workforce is dependent on the size, training, experience and work to be achieved in the target population or the socio-ecological interventions needed. Just prior to the audit, Australian advanced level public health nutritionists were estimated as a specialist workforce capacity at 20 % of that required [14], estimating that WA needed to increase to 265FTE. The only other published figures from the United States (US) planning models for workforce enumeration for government funded programs set the US ratio of 1FTE public health nutritionist to 133 000 head of population in the 90s [34, 35]. Updated in 2000 by the US Association of State and Territorial Public Health Nutrition Directors to 1FTE for every 50 000 head of population in consideration of the complexity of addressing obesity and nutrition of vulnerable population groups [36, 37]. Australian nutrition workforce enumeration demonstrates variability amongst states. Figures from South Australia suggested that the ratio for dedicated community nutrition positions was between 1.04 and 1.69 per 100 000 people in 2003, and the Queensland specialist workforce rose to 4.8FTE per 100 000 in 2003–4 and to 137.3FTE in total in 2009 [38]. WA’s 2004 population was 1,982,204 with 53.1FTE specialist nutritionists [28]. Matching Queensland’s investment, an additional 95.2 FTE would be required, similar to the 134.1 FTE (current and required) indicated by WA managers to fully deliver on nutrition service goals. The exemplar Queensland workforce was disbanded in 2012 following a newly elected State Government restructure which resulted in the devolution of public health with a 90 % reduction to 14 FTE in total [38]. Physical activity workforce human resource requirements are more challenging to estimate as there are no clear professional recommendations. The mixture of health promotion, physiotherapy, and nursing-trained practitioners highlights the need to develop a specialist workforce by defining both the competencies and numeration requirements to contribute to effective physical activity program delivery [11]. Consistent with the 2008 National Preventive Health Taskforce recommendation to expand the supply and support training of relevant primary health workers, health promotion workers, nutritionists, and dietitians, the findings suggests an obvious way to increase workforce capacity is to invest in workforce growth [1]. In Victoria, developing workforce capacity including the FTE, benefited obesity prevention strategies [39]. The variety of position titles and selection criteria used to recruit workers may lead to variability in the WA workforce. Whilst there has been growth in dietetics as a profession this has predominantly been in clinical services [40]. The WA NAPA workforce has not grown substantially since 2004, a worrying implication for achieving obesity targets. The importance of a diverse generalist workforce for service delivery was demonstrated but there were skill deficits in the respective areas. Reliance on the generalist workforce with limited or no training in NAPA to deliver interventions is likely to be problematic. Existing WA programs required dietetic input, e.g., FoodCents® [41] and future interventions needed to address the obesogenic environment require a coordinated and skilled workforce. Whilst it is important to work in a multidisciplinary and intersectorial way to reach the whole population, a lack of training and specialist workforce to deliver targeted workforce training is also a problem. These challenges are not confined to the Australian workforce. The US identified a lack of understanding of the complexity of the dietary change process by other practitioners and managers, lack of resources, training and mentoring to do the work, job insecurity and expectations that nutritionists would assume a variety of other roles [34, 42, 43]. Several European countries identify major constraining factors to public health nutrition workforce development [44]. Variable expectations about work roles and differences in priority placed on NAPA by managers may be due to their own preferences and/or past work experience. Many managers were clinically trained in disciplines such as nursing, suggesting that practitioners were reporting to managers without NAPA qualifications or delivery of community public health interventions. Other workforce development issues were the impending shortage of experienced workers as many are approaching retirement age, the overall staff and the workforce instability due to high turnover or unfilled positions. Short term funding, the high proportion of female staff and dissatisfaction with career pathways were reasons identified. Interruptions to service delivery, loss of partnerships, and loss of experience when staff leave without positions being filled are priority workforce issues [8]. Professional isolation is a challenge in rural areas [43] and the with ability to work effectively with peers due to competing pressures or risk factors were identified in this study. Policy implications for building workforce capacity Obesity prevention requires a strategic approach to workforce planning within governments and organisations. An appropriately trained and skilled workforce can help improve diet quality and physical activity to reduce obesity and improve population health [45]. Policy level support, organisational level workforce management, and continued competency and capacity building in the existing workforce are required. Workforce development is often not part of the range of policy options for public health nutrition [46]. Although human resource capacity and training were identified in strategic Australian policy as essential to build capacity to achieve Australian obesity outcomes the policies have since been rescinded and not replaced. The chronic disease or obesity prevention emphasis rather than the direct focus on addressing poor nutrition and physical inactivity may contribute to this. Governments are focussing on educating the individual rather than environmental, organisational, policy and legislative and economic approaches [47]. Efforts to reduce budget expenditure such as moving to contract, part-time or generalist practitioners or less experienced practitioners also have a negative impact on overall service delivery [34]. Although there is now a mandate for implementing a workforce development strategy [48], amid growing concern about the lack and potential loss of NAPA workforce capacity, there have been no subsequent workforce audits. More research also is required on how best to train and maintain a NAPA workforce to meet current challenges and future needs. Limitations of the research The survey was conducted over 10 years ago and a follow-up survey is timely and urgently needed. Caution should be taken when interpreting the results of this workforce audit as the interventions delivered by the Department of Health in Western Australia at the time of the audit were largely directed by national health priority areas. This study findings show that manager’s recognition of nutrition and physical activity as major health issues was a lower priority than other factors such as obesity, social impacts and mental health, see Table 1, which may have changed since the audit. Obesity remains a public health priority and research into effective public health policy options interventions has progressed [49, 50] and emphasise the need for inter-sectoral action and approaches. For example, there is increasing recognition of mental health issues and stigma related to body weight [51, 52]. Further government workforce audits are recommended and would need to consider the current policy and intervention context and the broader workforce involved in prevention. However, the findings maybe valuable for future workforce development given the lack of evidence on NAPA workforce in Australia and may contribute to evidence on the lack of progress in addressing issues such as obesity presently. The sampling was designed to target all NAPA service providers however individual practitioners in other settings who may have been involved with promotion in their clinical roles may not have been captured. The relatively low response rate among practitioners compared to managers is a limitation, however, other workforce audits have reported rates as less than 50 % percent [38]. The use of snowball sampling and the uniqueness of the WA context may limit the generalisability of findings. In addition, it should be noted that the practitioner survey relied on self-report data. Also several managers were unable to estimate some of their generalist workforce’s time dedicated to nutrition and or physical activity service delivery. Variable size of organisations meant some had more managers and practitioners included, although this was taken into account when enumerating the workforce so that positions were only counted once. The Department of Health NPAB manager (secondary author) who commissioned the audit was not included in the survey but the workforce at the NPAB has been included in enumeration estimates. Conclusion Workforce development needs to be a key strategic determinant for obesity prevention. The 2004 WA NAPA workforce audit highlighted a lack of responsibility for workforce development, an unclear and fragmented strategy, and a lack of fit for purpose workforce to deliver interventions. There is no doubt the programs required to effectively influence NAPA are challenging and complex yet there is little evidence of workforce considerations. Abbreviations FTEFull-time equivalents NAPANutrition and physical activity NHPANational Health Priority Areas NPABNutrition and physical activity Branch PAPhysical Activity PHNPublic Health Nutrition WAWestern Australia Funding The Department of Health in Western Australia funded the 2003–2004 Western Australian Audit of NAPA and Healthway, the Health Promotion Foundation, funded Curtin University to assist the translation of research into practice through the “Food Law, Policy and Communications to Improve Public Health Project”. http://foodpolicy.org.au/. Availability of data and materials The data that support the findings of this study are available from Department of Health, Western Australia but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Department of Health, Western Australia. Authors’ contributions AB designed the study instruments and conducted the data collection and carried out the collation of information and contributed extensively to drafting and reviewing the manuscript. CP conceptualized and designed the study and assisted with the development of the instruments, and conceptualizing and drafting the manuscript, and edited and approved the final manuscript as submitted. Both authors read and approved the final manuscript. Authors’ information AB is currently a Senior Lecturer and the Course Coordinator for the Master of Dietetics at Curtin University. She has been at Curtin as a permanent staff member since 1996 is a highly experienced lecturer winning teaching awards. She is responsible for teaching the public health nutrition content for a variety of courses. Her research interests are in food literacy capabilities and effective programs, nutrition during the lifecycle and food and nutrition policy. In 2012 she was awarded Fellow of the Public Health Association of Australia in recognition of significant contribution to PHAA and the field of public health. In 2014 the Dietitians Association of Australia recognised AB as an Advanced Accredited Practicing Dietitian in recognition of leadership, education, supervision, teaching and health professional training CP works part-time for Curtin University and the Western Australian Department of Health to try to build the capacity for nutrition epidemiology in Western Australia to inform policy and practice. CP is best known for managing the Go for 2&5® fruit and vegetable social marketing campaign. She has been awarded the International Fellow of the World Cancer Research Fund, bestowed September 2012, and has achieved Fellowship of the Public Health Association of Australia, appointed September 2012. CP has a particular interest in improving nutrition for population groups who are vulnerable to poor nutrition due to their social, environmental or economic circumstances. CP was manager of the former state-wide Nutrition and Physical Activity Branch of the Department of Health at the time of the survey. Competing interests The authors declare that they have no competing interests. Consent for publication Aggregated results from individuals are presented and no one individual can be identified. Research Information Sheets indicated the intended use of the results for reports and publications. Ethics approval and consent to participate Human Research Ethics approval was obtained from Curtin University’s Human Research Ethics Committee Protocol Approval HR 180/2003 Part 1 and Part 2. All participants were required to provide written consent for their participation. ==== Refs References 1. National Health Preventative Taskforce National Preventative Health Strategy, Obesity in Australia- a need for urgent action Technical Paper 2009 Canberra Commonwealth Department of Health and Aging 2. Institute for Health Metrics and Evaluation. Global Burden of Disease: Australia. In: Global Burden of Diseases, Injuries and Risk Factors Study 2010. volume 2014. Seattle: Institute for Health Metrics and Evaluation; 2014. http://www.healthdata.org/australia. 3. National Health and Medical Research Council. Australian Dietary Guidelines. In: National Health and Medical Research Council, vol. Commonwealth of Australia, Canberra; 2013. 4. Bodenheimer T Chen E Bennett H Confronting the growing burden of chronic disease: Can the U.S. Health Care Workforce do the job? Health Affairs 2009 28 64 74 10.1377/hlthaff.28.1.64 19124856 5. Gortmaker S Swinburn BA Levy D Carter R Mabry P Finegood D Huang T Marsh T Moodie M Changing the future of obesity: science, policy and action The Lancet 2011 378 838 47 10.1016/S0140-6736(11)60815-5 6. Kugelberg S. Public health nutrition in Europe: Workforce development and policy change. Stockholm: Karolinska Institutet; 2013. https://openarchive.ki.se/xmlui/handle/10616/41338?locale-attribute=sv. 7. Swinburn B Gill T Kumanyika S Obesity prevention: a proposed framework for translating evidence into action Obesity Rev 2005 6 23 33 10.1111/j.1467-789X.2005.00184.x 8. Hughes R A socio-ecological analysis of the determinants of national public health nutrition workforce capacity: Australia as a case study Fam. Community Health 2006 29 55 67 10.1097/00003727-200601000-00007 16340678 9. Jonsdottir S Thorsdottir I Kugelberg S Yngve A Kennedy N Hughes R Core functions for the public health nutrition workforce in Europe: a consensus study Public Health Nutr 2012 15 1999 2004 10.1017/S1368980012003357 22929414 10. Yancey A Fielding J Flores G Sallis J McCarthy W Breslow L Creating a robust public helth infrastructure for physical activity promotion Am J Prev Med 2007 32 68 78 10.1016/j.amepre.2006.08.029 17218192 11. Beck AJ Boulton ML Building an effective workforce: a systematic review of public health workforce literature Am J Prev Med 2012 42 S6 16 10.1016/j.amepre.2012.01.020 22502927 12. World Public Health Nutrition Association. Public Health Nutrition Definition. 2007. 13. Serra-Majem L Moving forward in public health nutrition- the 1st World Congress of Public Health Nutrition-Introduction Nutr Rev 2009 67 s2 6 10.1111/j.1753-4887.2009.00150.x 19453672 14. Hughes R Public health nutrition workforce composition, core functions, competencies and capacity: perspectives of advanced-level practitioners in Australia Public Health Nutr. 2003 6 607 13 14690042 15. Hughes R, Begley A, Yeatman H. Aspirational competency expectations for public health nutritionists in Australia: a consensus study. Nutr Diet. 2015;72:122–131. 16. Jonsdottir S Hughes R Thorsdottir I Yngve A Consensus on the competencies required for public health nutrition workforce development in Europe - the JobNut project Public Health Nutr 2011 14 1439 49 10.1017/S1368980010000625 20353620 17. Hughes R Competencies for effective public health nutrition practice: a developing consensus Public Health Nutr 2003 7 683 93 15251059 18. Strategic Intergovernmental Nutrition Alliance Eat Well Australia: An agenda for action for public health nutrition 2000–2010 2001 Department Health and Aged Care Canberra 19. Blanchard C Shilton T Bull F Global Advocacy for Physical Activity (GAPA): global leadership towards a raised profile Glob Health Promot 2013 20 113 21 10.1177/1757975913500681 24722750 20. Kohl H Craig C Lambert E Inoue S Alkandari JR Leetongin G Kahlmeier S Group. TLPASW The pandemic of physical inactivity: a global action for public health The Lancet 2012 380 294 305 10.1016/S0140-6736(12)60898-8 21. Shilton T Advocacy for physical activity- from evidence to influence IUPHE- Promot Educ 2006 X111 118 26 10.1177/10253823060130020106 22. Giles-Corti B People or places: What should be the target? J Sci Med Sport 2006 9 357 66 10.1016/j.jsams.2006.06.021 16931155 23. Lawrence MA Galal O Margetts BM Yngve A Building global alliances for public health nutrition training Nutr Rev 2009 67 Suppl 1 S66 8 10.1111/j.1753-4887.2009.00162.x 19453681 24. Hughes R Somerset S Definitions and conceptual frameworks for public health nutrition and community nutrition: a discussion paper Aust J Nutr Diet 1997 54 40 5 25. Dietitians Association of Australia. National competency standards for entry level dietetics. Canberra; 1993. 26. Shilton T Howat P James R Lower T Final Report: Review of competencies for Australian Health Promotion 2002 Perth Western Australian Centre for Health Promotion Research, Curtin University and National Heart Foundation of Australia (WA Division) Perth Western Australia 27. Hughes R Descriptive analysis of the designated public health nutrition workforce profile, practices and workforce development needs: Western Australia Australian Public Health Nutrition Workforce Project 2001 Brisbane Nutrition Unit Griffith University 28. Australian Bureau of Statistics. Population by Age and Sex, Western Australia, Jun 2004. In 3235555001; 2004. 29. Harris M Hobbs C Davies G Simpson S Bernard D Stubbs A Implementation of a SNAP intervention in two divisions of general practice: a feasibilty study Med J Aust 2005 183 54 60 15992345 30. Pollard CM Miller MR Daly AM Crouchley KE O'Donoghue KJ Lang AJ Binns CW Increasing fruit and vegetable consumption: success of the Western Australian Go for 2&5 campaign Public Health Nutr 2008 11 314 20 10.1017/S1368980007000523 17612423 31. Pollard C, Lewis J, Binns C. Selecting interventions to promote fruit and vegetable consumption from policy to action, a planning framework case study in Western Australia. Aust N Z Health Policy. 2009;2009. 32. Department of Health: National chronic disease strategy. Australian Health Ministers Conference, Editor. Canberra; 2005. 33. Schwarte L Samuels S Boyle M Clark S Flores G Prentice B Local public health departments in California: Changing nutrition and physical activity environments for obesity prevention J Public Health Management Prac 2010 16 E17 28 10.1097/PHH.0b013e3181af63bb 34. Haughton B George A The Public Health Nutrition workforce and its future challenges: the US experience Public Health Nutr 2008 11 782 91 10.1017/S1368980008003455 18413011 35. Haughton B Story M Keir B Profile of public health nutrition personnel: challenges for population/system-focused roles and state-level monitoring J Am Diet Assoc 1998 98 664 70 10.1016/S0002-8223(98)00151-5 9627624 36. Kaufman M Heimendinger J Foerster S Carroll MA Survey of nutritionists in state and local public health agencies J Am Diet Assoc 1986 86 1566 70 3771969 37. Association of State and Territorial Public Health Nutrition Directors. Survey of the Public Health Nutrition Workforce :1999–2000. In: Nutrition Assistance Program Report Series The Office of Analysis, Nutrition and Evaluation. Virginia: United States Department of Agriculture- Food and Nutrition Service; 2003. http://www.fns.usda.gov/sites/default/files/Publichealthsurvey.pdf. 38. Vidgen H, Adam M, Gallegos D. Who does nutrition prevention work in Queensland? An investigation of structural and political workforce reforms. Nutr Diet. 2015. early online view. doi:10.1111/1747-0080.12239. 39. Haby M Doherty R Welch N Mason V Community-based interventions for obesity prevention: lessons learned by Australian policy-makers BMC Res Notes 2012 5 20 10.1186/1756-0500-5-20 22233586 40. Health Workforce Australia Australia’s Health Workforce Series Dietitians in Focus 2014 Canberra Commonwealth of Australia 41. Foley R Pollard C Food Cent$--implementing and evaluating a nutrition education project focusing on value for money Aust N Z J Public Health 1998 22 494 501 10.1111/j.1467-842X.1998.tb01420.x 9659779 42. Chapman-Novakofski K Reicks M Dietetics practitioners in extension: what is the current climate and future demand? J Acad Nutr Diet 2013 113 1299 301 10.1016/j.jand.2013.05.020 23899609 43. Devine CM Jastran M Bisogni CA On the front line: practice satisfactions and challenges experienced by dietetics and nutrition professionals working in community settings in New York State J Am Diet Assoc 2004 104 787 92 10.1016/j.jada.2004.02.023 15127065 44. Kugelberg S Jonsdottir S Faxelid E Jonsson K Fox A Thorsdottir I Yngve A Public health nutrition workforce development in seven European countries: constraining and enabling factors Public Health Nutr 2012 15 1989 98 10.1017/S1368980012003874 22894796 45. Yngve A Sjostrom M Warm D Margetts B Perez Rodrigo C Nissinen A Effective promotion of healthy nutrition and physical activity in Europe requires skilled and competent people; European Master’s Programme in Public Health Nutrition Public Health Nutr 1999 2 449 52 10.1017/S1368980099000622 10610087 46. Yngve A Tseng M Haapala I Hodge A A robust and knowledgeable workforce is essential for public health nutrition policy implementation Public Health Nutr 2012 15 1979 80 10.1017/S1368980012004570 23075510 47. Caraher M Coveney J Public health nutrition and food policy Public Health Nutr 2004 7 591 8 10.1079/PHN2003575 15251049 48. Fox A, Chenhall C, Traynor M, Scythes C, Bellman J. Public health nutrition practice in Canada. Public Health Nutr. 2008;11:773–781. 49. Swinburn B Sacks G Vandevijvere S Kumanyika S Lobstein T Neal B Barquera S Friel S Hawkes C Kelly B INFORMAS (International Network for Food and Obesity/non-communicable diseases Research, Monitoring and Action Support): overview and key principles Obes Rev 2013 14 Suppl 1 1 12 10.1111/obr.12087 50. Hawkes C Jewell J Allen K A food policy package for healthy diets and the prevention of obesity and diet-related non-communicable diseases: the NOURISHING framework Obes Rev 2013 14 Suppl 2 159 68 10.1111/obr.12098 24103073 51. Puhl RM Heuer CA Obesity stigma: important considerations for public health Am J Public Health 2010 100 1019 28 10.2105/AJPH.2009.159491 20075322 52. Macpherson-Sanchez AE Integrating fundamental concepts of obesity and eating disorders: implications for the obesity epidemic Am J Public Health 2015 105 e71 85 10.2105/AJPH.2014.302507 25713933

PMC005xxxxxx/PMC5000431.txt

==== Front BMC Pregnancy ChildbirthBMC Pregnancy ChildbirthBMC Pregnancy and Childbirth1471-2393BioMed Central London 104110.1186/s12884-016-1041-6Research ArticleRelationship between hyperemesis gravidarum and small-for-gestational-age in the Japanese population: the Japan Environment and Children’s Study (JECS) Morokuma Seiichi +81-92-642-5105morokuma@med.kyushu-u.ac.jp 124Shimokawa Mototsugu shimokawa.m@nk-cc.go.jp 3Kato Kiyoko kkato@med.kyushu-u.ac.jp 14Sanefuji Masafumi sane26@pediatr.med.kyushu-u.ac.jp 15Shibata Eiji age-s@med.uoeh-u.ac.jp 67Tsuji Mayumi tsuji@med.uoeh-u.ac.jp 8Senju Ayako senju-a@med.uoeh-u.ac.jp 69Kawamoto Toshihiro kawamott@med.uoeh-u.ac.jp 68Kusuhara Koichi kkusuhar@med.uoeh-u.ac.jp 69Japan Environment & Children’s Study GroupKawamoto Toshihiro Saito Hirohisa Kishi Reiko Yaegashi Nobuo Hashimoto Koichi Mori Chisato Hirahara Fumiki Yamagata Zentaro Inadera Hidekuni Kamijima Michihiro Konishi Ikuo Iso Hiroyasu Shima Masayuki Ogawa Toshihide Suganuma Narufumi Kusuhara Koichi Katoh Takahiko jecscore@nies.go.jp 1 Research Center for Environmental and Developmental Medical Sciences, Kyushu University, Fukuoka, Japan 2 Department of Obstetrics and Gynecology, Kyushu University Hospital, Kyushu University, Fukuoka, Japan 3 Department of Cancer Information Research, Clinical Research Institute, National Kyushu Cancer Center, Fukuoka, Japan 4 Department of Obstetrics and Gynecology, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, 812-8582 Japan 5 Department of Pediatrics, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan 6 Japan Environment and Children’s Study, UOEH Subunit Center, University of Occupational and Environmental Health, Kitakyushu, Fukuoka Japan 7 Department of Obstetrics and Gynecology, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Fukuoka Japan 8 Department of Environmental Health, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Fukuoka Japan 9 Department of Pediatrics, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan 26 8 2016 26 8 2016 2016 16 1 24728 1 2016 20 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Small-for-gestational-age in infancy is a known risk factor not only for short-term prognosis but also for several long-term outcomes, such as neurological and metabolic disorders in adulthood. Previous research has shown that severe nausea and vomiting in early pregnancy (NVP) and hyperemesis gravidarum, which is an extreme form of NVP, represent risk factors for small-for-gestational-age birth. However, there is no clear consensus on this association. Thus, in the present study, we investigated the correlation between hyperemesis gravidarum and NVP on the one hand, and infant birth weight on the other, using data from the Japan Environment and Children’s Study (JECS). Methods The data utilized in the present study were obtained from the JECS, an ongoing cohort study that began in January 2011. Our sample size was 8635 parent–child pairs. The presence or absence of severe NVP, hyperemesis gravidarum, and potential confounding factors were noted. A multivariable regression analysis was used to estimate risks for small-for-gestational-age birth, and the results were expressed as risk ratios and 95 % confidence intervals. Results The risk ratios of small-for-gestational-age birth (95 % confidence interval) for mothers with severe NVP and those with hyperemesis gravidarum were 0.86 (0.62–1.19) and 0.81 (0.39–1.66), respectively, which represents a non-significant result. Conclusions In our analysis of JECS data, neither severe NVP nor hyperemesis gravidarum was associated with increased risk for small-for-gestational-age birth. Keywords Hyperemesis gravidarumSmall-for-gestational-ageBirth cohortMinistry of the Environment (JP)issue-copyright-statement© The Author(s) 2016 ==== Body Background There is a high incidence of nausea and vomiting in early pregnancy (NVP), reported at 35–91 % [1–4]. NVP can become severe in 0.3–3.6 % of cases, with hyperemesis gravidarum (HG) as an extreme form of NVP that is associated with weight loss [1–4]. The incidence of HG varies by country, and was reported at nearly 3.6 % in Japan [4]. The condition known as small-for-gestational-age (SGA) is a concern in infants, as it carries with it a multitude of risks, including a poorer life prognosis, neurological disorders, and metabolic diseases during adulthood [5, 6]. SGA is defined using the 10th percentile for birth weight as the cutoff value [7, 8]. There are many risk factors for SGA, but most of these are not well understood. Extreme NVP may result in poor health during pregnancy, which can influence the prognosis of fetuses [9, 10], possibly leading to an increase in the risk of SGA birth [9, 11–13]. Recent systematic reviews suggest that HG increases the risk of low birth weight and SGA by 42 and 28 %, respectively [12]. Furthermore, severe maternal weight loss in early pregnancy, typically linked with extreme NVP, has been linked with growth restriction [9]. However, other reports have suggested that HG does not influence growth restriction [14, 15], birth weight [11, 16, 17], or risk for SGA [18]. Thus, there is as yet no clear consensus on this issue [11, 16, 17]. In the present study, we investigated the effect of severe NVP and HG (extreme NVP), with respect to the risk for SGA birth in the Japanese population. Methods The data used in this study were obtained from The Japan Environment and Children’s Study (JECS), which is an ongoing cohort study that began in January 2011. The objective of the JECS is to determine the effect of environmental factors on children’s health. More than 100,000 pregnant women were recruited over a period of approximately 3 years. The recruitment period ended in March 2014. The pregnant women lived in one of the 15 study regions included in the JECS. The 15 regions were selected to cover wide geographical areas in Japan. We made contact with as many of these expecting mothers as possible. Either or both of the following two recruitment protocols were applied: 1) recruitment at the time of the first prenatal examination at cooperating health care providers, i.e., obstetric facilities (provider-mediated community-based recruitment), and/or 2) recruitment at local government offices issuing pregnancy journals, namely the Mother-Child Health Handbook, which is an official complimentary booklet that all expecting mothers in Japan are given when they become pregnant in order to receive municipal services for pregnancy, delivery, and childcare. The JECS protocol was approved by the Review Board on epidemiological studies of the Ministry of the Environment, and by the Ethics Committees of all participating institutions. The JECS is conducted in accordance with the Helsinki Declaration and other nationally valid regulations, and with written informed consent from all participants. However, those who had difficulty filling out the questionnaire in Japanese or had other unavoidable circumstances preventing them from participating in the survey, such as being in their hometown at the time of childbirth, were excluded from the analysis [19, 20]. As of the end of 2011, a total of 9646 participants had successful childbirths. After excluding cases with missing data and preterm births, we analyzed the records of the remaining 8631 women who had single, full-term (37–42 weeks) pregnancies (Fig. 1). The present study is based on the data set “jecs-ag-ai-20131008”, which was released in October 2013.Fig. 1 Participant inclusion flowchart Follow-up was conducted using a self-administered questionnaire. The questionnaires were completed during the first and second trimesters, as well as at 1 month postpartum. We obtained medical information from medical records transferred for examinations during the same time periods. The questionnaires were designed to collect information on pregnancy and medical history as well as on confounding and modifying factors, such as social and lifestyle factors. We collected information on birth, such as the birth weight, from the transferred medical records. The following question was included in the questionnaire for the second trimester to determine the status of HG: “Did you have morning sickness from conception until about week 12 of the pregnancy?” (1 = no, 2 = just nausea, 3 = vomiting, but was able to eat, 4 = vomiting, and was unable to eat). We thus defined the following groups for analysis: the “food intake group”, which included the women who answered 1, 2, or 3; the “no food” or severe NVP group, which included the women who answered 4; and the HG group, which was a subset of participants from the NVP group that included women with severe NVP and weight loss of >5 % from pre-pregnant weight in the first trimester. The participants underwent ultrasound examinations during the first trimester, and these results were used to determine the expected date of delivery if there was more than a 7-day difference between this date and the date calculated from the last menstrual period. Birth weight was transferred from medical records, and SGA was concluded if the weight was below the 10th percentile according to primiparous and multiparous birth size standards for both genders by gestational age in Japanese neonates [21]. The following covariates were included in the questionnaire for the first trimester: maternal age, pre-pregnancy body mass index (BMI), parity, smoking status, and alcohol consumption; the covariates of education and income were included in the questionnaire for the second trimester; the covariates of weight gain during pregnancy were calculated based on information from medical records. Statistical analysis Based on the records of mothers of singletons delivered at full term, we evaluated the relationship between SGA and NVP, HG, factors related to the patient’s background, and social factors. Continuous variables were expressed as mean ± standard deviation (SD). We calculated crude relative risk ratios (RRs) and 95 % confidence intervals (CIs) using the chi-squared test. The interrelationship between patient background, social factors, and birth weight was evaluated by univariate analysis. Covariates of maternal age, pre-pregnancy BMI, weight gain during pregnancy, gestational age at birth, smoking, alcohol consumption, education, and income were included in the calculation of adjusted risk ratios. The adjusted relative RR was calculated using a log-binomial regression model. All statistical analyses were performed using SAS version 9.3 (SAS Institute Inc., Cary, NC, USA). Results There were 880 patients (10.2 %) who experienced severe NVP, and 136 patients (1.6 %) who experienced HG. The mean age of participants, weeks of pregnancy at birth, and birth weight were 30.6 ± 5.02 years, 39.0 ± 1.14 weeks, and 3050.0 ± 371.32 g, respectively. The results of the univariate analysis are shown in Table 1. The adjusted risk ratios for mothers with a pre-pregnancy BMI of <18.5 kg/m2, mothers with a weight gain of <7 kg during pregnancy, and those who smoked were 1.58 (95 % CI, 1.32–1.90), 1.28 (95 % CI, 1.05–1.55), and 1.48 (95 % CI, 1.11–1.97), respectively, indicating a slightly higher risk of SGA birth. Moreover, the risk ratio was 0.60 (95 % CI, 0.43–0.85) for mothers with a pre-pregnancy BMI of >25 kg/m2, and 0.52 (95 % CI, 0.41–0.66) for mothers with a weight gain of >12 kg during pregnancy, indicating a lower risk of SGA birth.Table 1 Characteristics of all parent-child pairs included in this study (N = 8631) No. (%) Missing data No. of non-SGA births No. of SGA births % SGA RR for SGA birth 95 % CI Mother’s age (years)  19 or less 103 1.3 13 85 5 5.6 0.72 0.30 1.69  20–34a 5893 75.6 372 5093 428 7.8 (1.0)  35 or more 1801 23.1 83 1610 108 6.3 0.81 0.66 0.99  missing 834 15 774 45 Mother’s education   > 12 years 5227 62.8 320 4558 349 7.1 0.97 0.82 1.14   ≤ 12 yearsa 3098 37.2 135 2745 218 7.4 (1.0)  missing 306 28 259 19 Parity  0a 3150 38.4 18 2893 239 7.6 (1.0)   ≥ 1 5043 61.6 27 4669 347 6.9 0.91 0.77 1.06  missing 438 438 - - Pre-pregnancy body mass index   < 18.5 1363 16.2 69 1153 141 10.9 1.58 1.32 1.90  18.5–24.9a 6212 73.7 346 5462 404 6.9 (1.0)   ≥ 25 856 10.2 34 788 34 4.1 0.60 0.43 0.85  missing 200 34 159 7 Weight gain during pregnancy   < 7 kg 1354 17.7 64 1162 128 9.9 1.28 1.05 1.55  7–12 kga 4104 53.6 244 3560 300 7.8 (1.0)   > 12 kg 2198 28.7 102 2011 85 4.1 0.52 0.41 0.66  missing 975 73 829 73 Income   < 4 million yen 3266 41.1 183 2854 229 7.4 1.07 0.90 1.27  4–8 million yena 3836 48.3 208 3375 253 7.0 (1.0)   > 8 million yen 843 10.6 53 732 58 7.3 1.05 0.80 1.39  missing 686 39 601 46 Smoked during pregnancy  Noa 7991 94.5 455 7013 523 6.9 (1.0)  Yes 462 5.5 15 401 46 10.3 1.48 1.11 1.97  missing 178 13 148 17 Alcohol intake during pregnancy  Noa 7654 90.2 435 6710 509 7.1 (1.0)  Yes 835 9.8 38 732 65 8.2 1.16 0.90 1.48  missing 142 10 120 12 Data extracted from the Japan Environment and Children’s Study No. number, SGA small-for-gestational-age, RR risk ratio, CI confidence interval aUsed as reference in the calculation of risk ratios Tables 2 and 3 show the crude and adjusted risk ratios calculated using covariates such as the mother’s age, pre-pregnancy BMI, weight gain during pregnancy, parity, smoking and drinking, education, and income, to determine the effect of severe NVP or HG on the risk of SGA. The risk ratios for mothers with severe NVP and those with HG were 0.86 (95 % CI, 0.62–1.19) and 0.81 (95 % CI, 0.39–1.66), respectively, indicating a non-significant effect of NVP or HG on the risk for SGA birth.Table 2 Risk for small-for-gestational-age (SGA) birth associated with severe nausea and vomiting in early pregnancy (NVP) Total number No data on birth n (%) Non-SGA birth n (%) SGA birth n (%) Crude Confounder-adjusted RR 95 % CI RR 95 % CI Severe NVP 880 48 (5.5) 773 (87.8) 59 (6.7) 0.98 0.75–1.27 0.86 0.62–1.19 No severe NVPa 7563 420 (5.6) 6625 (87.6) 518 (6.8) (1.0) (1.0) No data on NVP state 188 15 (8.0) 164 (87.2) 9 (4.8) The crude and adjusted risk ratios calculated using covariates such as the mother’s age, pre-pregnancy body mass index, weight gain during pregnancy, parity, smoking and alcohol consumption status, education, and income, to determine the effect of severe NVP on the risk of SGA birth RR risk ratio, CI confidence interval aUsed as reference in the calculation of risk ratios Table 3 Risk for small-for-gestational-age (SGA) birth associated with hyperemesis gravidarum (HG) Total number No data on birth n (%) Non-SGA birth n (%) SGA birth n (%) Crude Confounder-adjusted RR 95 % CI RR 95 % CI HG 136 8 (5.9) 119 (87.5) 9 (6.6) 0.97 0.51–1.83 0.81 0.39–1.66 No HGa 6393 331 (5.2) 5622 (87.9) 440 (6.9) (1.0) (1.0) No data on HG state 2102 144 (6.9) 1821 (86.6) 137 (6.5) The crude and adjusted risk ratios calculated using covariates such as the mother’s age, pre-pregnancy BMI, weight gain during pregnancy, parity, smoking and alcohol consumption status, education, and income, to determine the effect of HG on the risk of SGA birth RR risk ratio, CI confidence interval aUsed as reference in the calculation of risk ratios Discussion In our analysis of JECS data, neither NVP nor HG was associated with the risk for SGA birth. The incidence of HG was 1.6 %, which is lower than the 3.6 % incidence reported by the latest study in the general Japanese population [4], but within the range of 0.3–2.0 % reported by other studies [1–3]. In addition, the participants in our study reported an incidence of NVP of 10.2 %, which is lower than the 33 % incidence reported by Chortatos et al. [22]; the difference is likely related to the fact that we defined NVP based on self-reported accounts of reduced food intake. Our study has a methodological limitation, because data regarding the severity of NVP were collected via a self-response questionnaire, while data regarding maternal weight loss were collected from the Mother-Child Health Handbooks and hospital records, and it is unknown whether participants required hospitalization for severe HG, how long severe NVP or HG persisted, and whether the condition reflected in the biochemical parameters. Another limitation is the fact that the questionnaire was applied in the second trimester, but the questions themselves referred to early pregnancy; thus, there might be the risk of recall bias, resulting in an overestimation of the severity of NVP. However, we do not believe that this effect was significant, because the questionnaire was applied during the pregnancy period; moreover, the definition of HG was based on independent records of maternal weight loss. A further limitation is related to the fact that our results were obtained based on the data regarding 136 cases of HG, which may be considered a small number in the context of an epidemiologic study. Nonetheless, given that the incidence of HG is expected to be under 2 %, and there is yet no consensus regarding the influence of HG on the risk for SGA birth, we believe that a sample size of 136 cases can ensure sufficient power to detect relevant trends, as some reports indicate that HG may increase the risk for SGA birth by up to 40 %; moreover, even if the power is low, the potential tendencies should be recognizable, because the confidence interval for our results is narrow. Finally, another limitation of the study is related to the fact that the incidence of SGA birth in the group of mothers for whom weight gain information was missing was relatively high. Unfortunately, the reason for this higher incidence of SGA births cannot be assessed based on the data available to us. While it is possible that the characteristics of the mothers excluded from the study because of missing information on weight gain may have an influence on the results, we do not expect this influence to extend to the conclusions of our study. Previous research demonstrating HG as a risk factor for SGA includes a study by Bailit et al., which showed that neonates born from mothers requiring hospitalization for HG were 125 g smaller compared to those born from mothers without such symptoms [11]. However, that study employed hospital admission rates for defining HG, which is a more subjective measure than is maternal weight loss. On the other hand, in other studies, which reported that HG leads to SGA birth [9, 10], the HG definition was based on maternal weight loss throughout pregnancy period; however, it was unclear whether the weight change was due to HG. In our study, the HG group included mothers with severe NVP (vomiting and not able to eat) and with weight loss of >5 % from pre-pregnant weight in the first trimester. Based on such a strict definition, our results showed that neither severe nor extreme NVP (i.e., HG) represented a risk factor for SGA birth. The recent Norwegian Mother and Child Cohort Study reported that HG-exposed babies had slightly reduced birthweight, but there were no association between HG and SGA birth [18, 23], although it should be noted that no adjustment for weight gain was made, while adjusting for smoking status slightly increased the effect of HG. Further reports have suggested that HG does not influence birth weight [11, 16, 17]. Our results are in agreement with the findings of the studies that reported no relationship between HG and SGA birth; nevertheless, the relevance of adjusting for weight gain when evaluating the influence of HG should be noted, implying that the risk for SGA birth is reduced when sufficient weight gain is ensured during pregnancy. It is important to note that both sets of studies (i.e., those concluding an effect and those concluding a lack of an effect) studied patients who required hospitalization. Even under these conditions, there is no conclusive evidence regarding the effect of HG on birth weight. Therefore, precise diagnostic criteria for HG should be developed for use in future investigations. Conclusions Our results suggest that neither NVP nor HG affect birth weight. Despite the methodological limitations of the study, we believe that these results indicate that pregnant women need not be concerned about potential risk for SGA birth due to NVP or HG. Abbreviations 95 % CI95 % confidence interval BMIBody mass index HGHyperemesis gravidarum JECSJapan Environment and Children’s Study NVPNausea and vomiting in early pregnancy RRRisk ratio SDStandard deviation SGASmall-for-gestational-age Acknowledgements We would like to express our gratitude to all participants of this study, and all individuals involved in data collection. Members of JECS as of 2015 (principal investigator, Toshihiro Kawamoto): Hirohisa Saito (National Center for Child Health and Development, Tokyo, Japan), Reiko Kishi (Hokkaido University, Sapporo, Japan), Nobuo Yaegashi (Tohoku University, Sendai, Japan), Koichi Hashimoto (Fukushima Medical University, Fukushima, Japan), Chisato Mori (Chiba University, Chiba, Japan), Fumiki Hirahara (Yokohama City University, Yokohama, Japan), Zentaro Yamagata (University of Yamanashi, Chuo, Japan), Hidekuni Inadera (University of Toyama, Toyama, Japan), Michihiro Kamijima (Nagoya City University, Nagoya, Japan), Ikuo Konishi (Kyoto University, Kyoto, Japan), Hiroyasu Iso (Osaka University, Suita, Japan), Masayuki Shima (Hyogo College of Medicine, Nishinomiya, Japan), Toshihide Ogawa (Tottori University, Yonago, Japan), Narufumi Suganuma (Kochi University, Nankoku, Japan), Koichi Kusuhara (University of Occupational and Environmental Health, Kitakyushu, Japan), Takahiko Katoh (Kumamoto University, Kumamoto, Japan). Funding JECS was funded by the Japanese Ministry of the Environment. The findings and conclusions of this article are solely the responsibility of the authors and do not represent the official views of the above government. This article was supported in part by MEXT KAKENHI (24119004) at the time of the design and composition. The funding bodies had no role in the design of the study, collection and analysis of data, interpretation of the results, writing the manuscript, or decision to publish. Availability of data and materials The data used to derive our conclusions are unsuitable for public deposition due to ethical restrictions and specific legal framework in Japan. It is prohibited by the Act on the Protection of Personal Information (Act No. 57 of 30 May 2003, amended on 9 September 2015) to publicly deposit data containing personal information. The Ethical Guidelines for Epidemiological Research enforced by the Japan Ministry of Education, Culture, Sports, Science and Technology and the Ministry of Health, Labor and Welfare also restricts the open sharing of the epidemiologic data. All inquiries about access to data should be sent to jecs-en@nies.go.jp. The person responsible for handling inquiries sent to this e-mail address is Dr Shoji F. Nakayama, JECS Programme Office, National Institute for Environmental Studies. Authors’ contributions K Kusuhara, K Kato, TK, and SM designed the study. MS, MT, and SM analyzed and interpreted the data. SM, MS, ES, and AS wrote the manuscript. All authors contributed critical revisions to the manuscript, and read and approved the final draft of the manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate The JECS protocol was approved by the Review Board on epidemiological studies of the Ministry of the Environment, and by the Ethics Committees of all participating institutions. The JECS is conducted in accordance with the Helsinki Declaration and other nationally valid regulations, and with written informed consent from all participants. ==== Refs References 1. Einarson TR Piwko C Koren G Quantifying the global rates of nausea and vomiting of pregnancy: a meta analysis J Popul Ther Clin Pharmacol 2013 20 e171 83 23863575 2. Källén B Hyperemesis during pregnancy and delivery outcome: a registry study Eur J Obstet Gynecol Reprod Biol 1987 26 291 302 10.1016/0028-2243(87)90127-4 3691940 3. Einarson TR Piwko C Koren G Prevalence of nausea and vomiting of pregnancy in the USA: a meta analysis J Popul Ther Clin Pharmacol 2013 20 e163 70 23863545 4. Matsuo K Ushioda N Nagamatsu M Kimura T Hyperemesis gravidarum in Eastern Asian population Gynecol Obstet Invest 2007 64 213 6 10.1159/000106493 17664884 5. Kok JH den Ouden AL Verloove-Vanhorick SP Brand R Outcome of very preterm small for gestational age infants: The first nine years of life Br J Obstet Gynaecol 1998 105 162 8 10.1111/j.1471-0528.1998.tb10046.x 9501780 6. Patterson RM Prihoda TJ Gibbs CE Wood RC Analysis of birth weight percentile as a predictor of perinatal outcome Obstetrics Gynecol 1986 68 459 63 7. Committee on Practice Bulletins-Gynecology, American College of Obstetricians and Gynecologists, Washington, DC 20090-6920, USA Intrauterine growth restriction. Clinical management guidelines for obstetrician-gynecologists. American College of Obstetricians and Gynecologists Int J Gynaecol Obstet 2001 72 85 96 10.1016/S0020-7292(00)90000-6 11203689 8. Goldenberg RL Cutter GR Hoffman HJ Foster JM Nelson KG Hauth JC Intrauterine growth retardation: Standards for diagnosis Am J Obstet Gynecol 1989 161 271 7 10.1016/0002-9378(89)90497-3 2669484 9. Dodds L Fell DB Joseph KS Allen VM Butler B Outcomes of pregnancies complicated by hyperemesis gravidarum Obstet Gynecol 2006 107 2 Pt 1 285 92 10.1097/01.AOG.0000195060.22832.cd 16449113 10. Gross S Librach C Cecutti A Maternal weight loss associated with hyperemesis gravidarum:a predictor of fetal outcome Am J Obstet Gynecol 1989 160 906 9 10.1016/0002-9378(89)90307-4 2712122 11. Bailit JL Hyperemesis gravidarium: epidemiologic findings from a large cohort Am J Obstet Gynecol 2005 193 3 Pt 1 811 4 10.1016/j.ajog.2005.02.132 16150279 12. Veenendaal MV van Abeelen AF Painter RC van der Post JA Roseboom TJ Consequences of hyperemesis gravidarum for offspring: a systematic review and meta-analysis BJOG 2011 118 1302 13 10.1111/j.1471-0528.2011.03023.x 21749625 13. Zhou Q O’Brien B Relyea J Severity of nausea and vomiting during pregnancy: what does it predict? Birth 1999 26 108 14 10.1046/j.1523-536x.1999.00108.x 10687575 14. Bashiri A Neumann L Maymon E Katz M Hyperemesis gravidarum: epidemiologic features, complications and outcome Eur J Obstet Gynecol Reprod Biol 1995 63 135 8 10.1016/0301-2115(95)02238-4 8903768 15. Hallak M Tsalamandris K Dombrowski MP Isada NB Pryde PG Evans MI Hyperemesis gravidarum. Effects on fetal outcome J Reprod Med 1996 41 871 4 8951141 16. Tan PC Jacob R Quek KF Omar SZ Pregnancy outcome in hyperemesis gravidarum and the effect of laboratory clinical indicators of hyperemesis severity J Obstet Gynaecol Res 2007 33 457 64 10.1111/j.1447-0756.2007.00552.x 17688612 17. Tsang IS Katz VL Wells SD Maternal and fetal outcomes in hyperemesis gravidarum Int J Gynaecol Obstet 1996 55 231 5 10.1016/S0020-7292(96)02778-6 9003948 18. Vandraas K Vikanes A Vangen S Magnus P Stoer N Grjibovski A Hyperemesis gravidarum and birth outcomes-a population-based cohort study of 2.2 million births in the Norwegian Birth Registry BJOG 2013 13 1654 60 10.1111/1471-0528.12429 24021026 19. Kawamoto T Nitta H Murata K Toda E Tsukamoto N Hasegawa M Rationale and study design of the Japan environment and children’s study (JECS) BMC Public Health 2014 14 25 10.1186/1471-2458-14-25 24410977 20. Ministry of the Environment. Japan Environment and Children Study. [homepage on the Internet] http://www.env.go.jp/en/chemi/hs/jecs/ [updated 9 Sep 2013; Accessed 22 Sept 2013]. 21. Itabashi K Miura F Uehara R Nakamura Y New Japanese neonatal anthropometric charts for gestational age at birth Pediatr Int 2014 56 702 8 10.1111/ped.12331 24617834 22. Chortatos A Haugen M Iversen PO Vikanes Å Eberhard-Gran M Bjelland EK Magnus P Veierød MB Pregnancy complications and birth outcomes among women experiencing nausea only or nausea and vomiting during pregnancy in the Norwegian Mother and Child Cohort Study BMC Pregnancy Childbirth 2015 15 138 10.1186/s12884-015-0580-6 26100060 23. Vikanes ÅV Støer NC Magnus P Grjibovski AM Hyperemesis gravidarum and pregnancy outcomes in the Norwegian Mother and Child Cohort - a cohort study BMC Pregnancy Childbirth 2013 13 169 10.1186/1471-2393-13-169 24004605

PMC005xxxxxx/PMC5000432.txt

==== Front Tob Induc DisTob Induc DisTobacco Induced Diseases2070-72661617-9625BioMed Central London 9610.1186/s12971-016-0096-5ResearchModelling intentions to provide smoking cessation support among mental health professionals in the Netherlands Blankers Matthijs mblankers@trimbos.nl 123Buisman Renate r.s.m.buisman@fsw.leidenuniv.nl 14Hopman Petra phopman@trimbos.nl 1van Gool Ronald r.vangool@ggzingeest.nl 56van Laar Margriet mlaar@trimbos.nl 11 Netherlands Expertise Centre on Tobacco Control (NET), Trimbos Institute, Utrecht, The Netherlands 2 Department of Research, Arkin Mental Health Care, Amsterdam, The Netherlands 3 Department of Psychiatry, Academic Medical Centre, University of Amsterdam, Amsterdam, The Netherlands 4 Centre for Child and Family Studies, Leiden University, Leiden, The Netherlands 5 GGz inGeest Mental Health Institute, Amsterdam, The Netherlands 6 Cluster of Nursing, Leiden University of Applied Sciences, Leiden, The Netherlands 26 8 2016 26 8 2016 2016 14 1 3211 3 2016 18 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Tobacco use prevalence is elevated among people with mental illnesses, leading to elevated rates of premature smoking-related mortality. Opportunities to encourage smoking cessation among them are currently underused by mental health professionals. In this paper, we aim to explore mechanisms to invigorate professionals’ intentions to help patients stop smoking. Methods Data stem from a recent staff survey on the provision of smoking cessation support to patients with mental illnesses in the Netherlands. Items and underlying constructs were based on the theory of planned behaviour and literature on habitual behaviour. Data were weighted and only data from staff members with regular patient contact (n = 506) were included. Descriptive statistics of the survey items are presented and in a second step using structural equation modelling (SEM), we regressed the latent variables attitudes, subjective norms (SN), perceived behavioural control (PBC), past cessation support behaviour (PB) and current smoking behaviour on intentions to provide support. In optimisation steps, models comprising a subset of this initial model were evaluated. Results A sample of 506 mental health workers who had direct contact with patients completed the survey. The majority of them were females (70.0 %), respondents had an average age of 42.5 years (SD = 12.0). Seventy-five percent had at least a BSc educational background. Of the respondents, 76 % indicated that patients should be encouraged more to quit smoking. Respondents were supportive to train their direct colleagues to provide cessation support more often (71 %) and also supported the involvement of mental health care facilities in providing cessation support to patients (69 %). The majority of the respondents feels capable to provide cessation support (66 %). Two thirds of the respondents wants to provide support, however only a minority (35 %) intends to actually do so during the coming year. Next, using SEM an acceptable fit was found of the constructs derived from the theory of planned behaviour and literature on habitual behaviour to the weighted data (χ2 (322) = 1188, p < .001; RMSEA = 0.067; CFI = 0.983), after removal of insignificant latent variables (SN and current smoking) and inclusion of covariates. Attitudes, PBC and PB of staff are the strongest identified correlates of intention toward providing cessation support to patients. SN and staff smoking behaviour were found to be weaker, non-significant correlates. Conclusions To nudge staff towards providing cessation support to people with mental illnesses one should aim at influencing attitudes and perceived behavioural control. Electronic supplementary material The online version of this article (doi:10.1186/s12971-016-0096-5) contains supplementary material, which is available to authorized users. Keywords PsychiatrySurvey researchTreatment and interventionStructural equation modellingNetherlands Ministry of Health, Welfare and Sportissue-copyright-statement© The Author(s) 2016 ==== Body Background Tobacco use prevalence is elevated among people with mental illnesses, compared to the general population. Estimates indicate that on average, smoking prevalence is two to four times higher among people with mental illnesses than in the general population [1]. People with schizophrenia have notably high smoking prevalence rates: a review by de Leon and Diaz [2] estimates the global daily smoking rate for people with schizophrenia to be 62 %. Common mental disorders such as depression (37 %), bipolar disorder (69 %) and substance use disorders (77 %–93 %) also have been found to be associated with high smoking rates [1, 3, 4] compared to smoking prevalence estimates of the World Health Organisation for the general population worldwide (21 % smoked tobacco in 2013 [5]). As a consequence, premature smoking-related mortality is common among people with mental illnesses. For example, based on data from individuals hospitalized with a primary psychiatric diagnosis in California from 1990 to 2005, mortality was associated with tobacco smoking in 23,620 of the 44,469 patients with schizophrenia (53 %) [1]. In comparison, an estimated 480,000 [6] of the 2,596,993 deaths in the general population of the United States in 2013 (18 %) died as a consequence of cigarette smoking and exposure to tobacco smoke. Therefore, a moderate proportion of all premature mortality among people with mental illnesses may be prevented if successful measures would be taken to reduce smoking rates among them. To involve this population in smoking cessation treatment, mental health care facilities are a promising setting [4, 7]. However, based on recent evaluations, many opportunities to encourage and support smoking cessation in mental health care institutes are currently not being used [8–10]. This also applies to the Netherlands [11], where the current study was performed. In a recent study conducted in the Netherlands, it was found that little more than half (54.7 %) of the staff of inpatient facilities had ever helped a patient stop smoking; 24 % had done so in the last year, whereas 35 % intended to provide cessation support to a patient next year [11]. Smoking cessation support in mental health institutes might not be provided more often for a number of reasons. These include tolerant smoking policies and informal norms regarding the acceptability of smoking among staff [4], staff members’ and patients’ opinions that smoking is often a lesser concern for people with mental illnesses [4], or even that smoking is helpful in reducing symptoms of disorders (eg self-medication hypothesis) [12]. Staff members’ own smoking status is also hypothesised to affect the likelihood they will provide cessation support [13]. Other possible reasons may be a lack of training, skills and support for staff to help patients stop smoking, or the limited availability of (effective) interventions to aid smoking cessation [4]. In the current paper, we aim to explore these and other possible mechanisms underlying staff members’ intentions to help patients stop smoking. In this exploration, constructs from the Theory of Planned Behaviour (TPB) [14] are used. The TPB is an established theory to model (health) intentions and behaviour [15–17]. According to the TPB, intentions are the most proximal determinants of behaviour. Intentions in turn are a product of behavioural attitudes (beliefs, feelings and tendencies towards a behaviour), the subjective norm (SN) regarding a behaviour and perceived behavioural control (PBC), which reflects the perception of being able to perform or control a behaviour. In addition, PBC is also hypothesized to have a direct influence on behaviour [14]. The TPB has frequently served as a basis for designing successful smoking cessation and other addiction treatment programs [18]. The number of studies in which the TPB is applied to modelling and changing clinicians’ behaviour is however smaller. A systematic review published in 2007 [19] identified 20 studies in which the TPB (or its predecessor, the theory of reasoned action) has been used in relation to clinicians’ behaviour. The authors conclude that the small number of studies is striking and unfortunate, as the discrepancy between clinicians’ prescribed (based on evidence based treatment guidelines) and actual behaviour implies a need for more research on possible approaches to narrow this gap [19]. Of the 20 studies that were included, two focussed on the provision of smoking cessation interventions by clinicians, neither of the two focussed on mental health care providers. The study by McCarty and colleagues among 397 staff nurses at four hospitals in the United States found that providing cessation advice was related to attitudes toward offering advice and perceived ability to offer advice [20]. The other study, by Puffer and colleagues found that attitudes and PBC were the most important predictors of intention to offer smoking cessation advice in accordance with coronary heart disease guidelines among community practise nurses in England [21]. Our study aims to contribute to this knowledge base and is (to our best knowledge) the first to evaluate the applicability of the TPB in modelling the intention of mental health care treatment staff to provide cessation support to their patients. Providing cessation support can range from single session brief advice to an extensive psychosocial or pharmacological intervention. We will test whether attitudes, subjective norms, PBC, past cessation support behaviour and current smoking behaviour together are significantly associated with intentions to provide future support. We will also test whether a subset of this model, consisting of only the significant paths between these constructs and intention adequately fits the data. This will identify key constructs to address in order to increase the rate at which mental health care staff will provide cessation support. Methods Data source Data were obtained from a survey (August – November 2014) on attitudes, norms, smoking policy, perceived behavioural control, intentions and behaviour towards smoking cessation support in mental health institutes in the Netherlands. Survey items were developed by the authors of the study, with input taken from interviews with the target audience (which were part of the general report [11]), from previous studies on mental health care staff opinions on smoking (cessation) and from the TPB literature [14–19]. The survey frame consists of the 57,310 employees [22] of three types of institutes: (a) integrated mental health care institutes, which usually offer both in- and outpatient mental health care and substance abuse treatment (35 institutes), (b) substance abuse treatment centres (9 institutes), and (c) regional institutes for sheltered housing (20 institutes). Together, these institutes comprise the voluntary inpatient mental health facilities for adults in the Netherlands. At times of the study, 64 institutes were represented by the overarching sector organisation of specialist mental health and addiction care providers. Employees working for these 64 institutes were invited to participate in this internet survey. Recruitment of participants Participants were recruited through invitations circulated among staff by the treatment institutes’ newsletters and via the Trimbos Institute (Netherlands institute of mental health and addiction) website. In order to motivate the target audience to participate, three iPads were raffled off. Ethics, consent and permissions All participants provided informed consent before participating in the survey, in line with the Dutch Medical Research Involving Human Subjects Act. Based on previous consultation with the Netherlands’ Central Committee on Research Involving Human Subjects, survey research as performed for this study is exempted from medical ethics approval. Measures Attitudes towards their role in providing cessation support to patients were measured with 12 items, answered on a 5-point Likert scale (range: completely disagree-completely agree). An example of an item is: “Patients should be encouraged more often to quit smoking”. Subjective norms regarding smoking and cessation support in the institutes participants worked for were measured with four items, answered on a 5-point Likert scale. Subjective norms are operationalized as perceived smoking policy, which is an injunctive norm. An example of an item is: “The institute I work for enforces a strict smoking policy”. Perceived behavioural control towards providing cessation support to patients is measured with four items, answered on a 5-point Likert scale. An example of an item is: “If I want to, I am able to help a patient quit smoking”. Intention to provide cessation support to patients in the near future is measured with four items. An example of an item is: “Next year, I intend to help at least one patient quit smoking”. Past behaviour regarding providing cessation support to patients was measured with three items. An example of an item is: “In the past year, I have helped at least one patient quit smoking”. Respondent’s smoking behaviour, comprising of smoking status, time until first cigarette after waking up in the morning (if respondent is a daily smoker, otherwise set to 0) and quit intentions (if applicable) was measured with three items. Survey weighting In order to improve the representativeness of the sample, survey weighting was applied. Weights were calculated in order to optimize the representativeness of our sample regarding type of organization, number of inhabitants of the province, gender, age, part time factor and type of function. Survey weights were estimated using raking calibration in R 3.2.1. As a reference value, information regarding the labour market for mental health workers in the three types of organizations was used [22]. Weight bounds were set at 1/6 (lowest possible weight) and 6 (highest possible weight). All analyses in this paper were performed using unweighted data, and corroborated using weights. In the results section, it is indicated whether weighted or unweighted data are reported. Analysis plan As a first step in the analysis procedure, missing data were analysed and addressed. Overall, the missing rate was low, with an average of 3 % missing or invalid responses on all items in the survey (per item range: 0-18 %). However, a principled approach to data missingness is important even under relatively low missingness rates, especially if multivariate analyses including structural equation modelling (SEM) are planned. Therefore, missing observations were imputed under the Missingness At Random assumption using Amelia-2 [23] for R version 3.2.1 [24]. Next, the reliability of the scales was tested using maximum-likelihood factor analysis and Cronbach’s α coefficient for internal consistency. Scoring of contra-indicative items was reversed. Variables that were poor factor indicators (loadings <0.4) on a one-factor solution were excluded from the scales. After Cronbach’s α reliability coefficients were calculated, a SEM was constructed with the TPB constructs (attitudes, subjective norms, PBC and intention), past behaviour and current smoking behaviour as latent variables. The a priori hypothesis was that the full model comprising the five latent variables associated with intention would optimally fit the data. In optimisation steps, alternative models, consisting of a subset of the five initial latent variables were created and tested for their association with intention. Therefore, the SEM approach can be described as model-generating, starting with theory-based constructs. The outcome variable (intention) is categorical. Therefore, diagonally weighted least squares (DWLS) with robust standard errors and mean and variance adjusted test statistics were used for the estimation of the SEM. SEM analyses were performed based on the covariance matrices using the R package lavaan version 0.5-19 [25]. The residual variances and the variances of exogenous latent variables are included in the model and set free. The metric of each latent variable is determined by fixing their variances to 1.0 (which gave the same results as fixing the first indicator to 1.0). The means of the observed variables are entered in the model. To estimate SEMs with categorical outcomes and a DWLS estimator while taking survey weights in account is not possible in lavaan version 0.5-19. Therefore, a parametric bootstrapping procedure was performed in which the process of SEM estimation was repeatedly (1000 iterations) performed on a parametrically bootstrapped dataset, in which the probability for a given case to be sampled in the bootstrapped dataset was proportional to its survey weight. Through this approach, the application of survey weights in the SEM estimation process was computationally approached. The SEMs were evaluated based on common SEM fit indices: (1) chi-square test of model fit (χ2), (2) comparative fit index (CFI), (3) and root mean square error of approximation (RMSEA). For the presentation of the SEM analyses, we adhered to Hoyle and Isherwood’s recommended reporting standards [26] and take in account the reporting recommendations by Jackson and colleagues [27]. Results Participants In total, 770 staff members submitted the survey via the submit button on the last page. Of those, 170 were excluded. The majority (n = 132) of these 170 were excluded because they worked for other mental health care organizations than the three types we intended to include in this study. Others were excluded because filling out the survey took them an unreasonably short (<7 min) or long (>6 h) time (n = 7), because of inconsistencies in their demographic data which indicated invalid input (n = 19) or because they were multivariate outliers based on Mahalanobis distances over all variables (n = 12). Of the remaining 600 respondents, 94 were excluded as they did not regularly have treatment contact with patients. The result was a net sample of 506 participants who had direct contact with patients. This can be considered a sufficient sample size based on a common rule-of-thumb for sample sizes in SEM (minimum of 10 cases per parameter [28]), and on a recent simulation study (n > 460 for relatively complex models [29]). Based on weighted data, the majority of the respondents were females (70.0 %). Respondents had an average age of 42.5 years (SD = 12.0). Seventy-five percent had at least a BSc educational background. The most common vocational background of the respondents was in nursing (38.2 %), followed by social work (15.6 %), psychology (8.0 %), medicine (6.1 %), or other vocational backgrounds (e.g. drama therapists, music therapists; 2.4 %). The other 29.7 % did not have a vocational background in mental health. Analysis of the unweighted data indicated a slight underrepresentation of females (62.7 % vs. 70.0 % after applying survey weights) and an overrepresentation of respondents with a social work background (23.2 % vs. 15.6 % after applying survey weights) in the sample–compared to the weighted data. Descriptive statistics of items and latent variables The measurement items which comprise the latent variables attitudes (ATT), subjective norms (SN), perceived behavioural control (PBC) and intention (INT) are summarised in Fig. 1. The full list of items and response options for each latent variable is available as Additional file 1. The items used in the SEM analysis for the latent variables “past cessation support behaviour” (PB) and “smoking behaviour” (SMO) are listed in Table 1, with a summary of the responses.Fig. 1 Measurement items for attitudes, subjective norms, perceived behavioural control and intention Table 1 Responses to the items comprising the latent variables “Past cessation support behaviour” and “Smoking behaviour” Latent variable Item Response Past cessation support behaviour (PB) Have you ever helped a patient quit? No: 45.3 % (n = 229) Yes, not during the last year: 30.6 % (n = 155) Yes, during the last year: 24.1 % (n = 122) How many patients have you helped quit in the last year? 0 patients: 75.4 % (n = 381) 1-3 patients: 16.5 % (n = 83) >3 patients: 8.3 % (n = 42) What percentage of your patients has received cessation support? <10 % of patients: 73.7 % (n = 373) 10 % - 30 % of patients: 15.7 % (n = 79) >30 % of patients: 10.6 % (n = 54) Smoking behaviour (SMO) Do you smoke cigarettes yourself? I have never smoked: 29.3 % (n = 148) I used to smoke but quit: 42.1 % (n = 213) I am a regular smoker: 28.1 % (n = 142) How long after waking up do you smoke your first cigarette? I don’t smoke daily: 80.2 % (n = 406) Within 5 min: 1.3 % (n = 7) Between 6 and 30 min: 8.7 % (n = 44) Between 31 and 60 min: 5.9 % (n = 30) Longer than 60 min: 3.8 % (n = 19) Do you intend to quit smoking? I don’t smoke/not applicable: 71.6 % (n = 362) Within 6 months from now: 10.3 % (n = 52) Not within 6 months from now: 18.1 % (n = 92) Table note: Percentages are based on weighted data. Counts are based on the n = 506 respondents with direct patient contact. Percentages for “Do you smoke cigarettes” do not count up to 100 % as 0.5 % of the respondents indicated “I don’t know” – those answers have been labelled as missing in the scale analyses Attitudes The scale has a good reliability (Cronbach’s α = 0.90; one factor with eigenvalue >1, based on unweighted data). In general, respondents held positive attitudes towards addressing smoking among patients. The majority of the respondents (76 %) indicated that patients should be encouraged more to quit smoking. Respondents were also supportive of providing more training to their direct colleagues to provide cessation support (71 %) and the involvement of mental health facilities in providing cessation support to patients (69 %). Respondents were notably less supportive of general smoking bans in mental health facilities (39 %) nor considered cessation support currently to be an important task in the ward they worked on. Subjective norms The scale has an acceptable reliability (α = 0.71; one factor with eigenvalue >1, based on unweighted data). The majority of the respondents thinks that smoking policies in their centre prescribe that patients should quit smoking (66 %), while a minority thinks the smoking policy in their centre is very prohibitive of smoking (41 %). Perceived behavioural control The scale’s reliability is acceptable but somewhat low (α = 0.65; one factor with eigenvalue >1, based on unweighted data). The majority of the respondents feel capable to provide cessation support (66 %). Intention The scale has a good reliability (α = 0.80; one factor with eigenvalue >1, based on unweighted data). Two thirds of the respondents wants to provide support, however only a minority (35 %) intends to actually do so over the next year. Past behaviour This scale showed an acceptable reliability (α = 0.71; one factor with eigenvalue >1, based on unweighted data). More than half of the respondents (55 %) indicates that he or she has never helped a patient quit smoking during his or her career. Only 8 % of the respondents has helped more than three patients quit smoking in the last year. Respondent’s smoking behaviour This scale has a good reliability of α = 0.82; one factor with eigenvalue >1, based on unweighted data). Twenty-eight percent of the respondents is a regular smoker, which is somewhat more than the average proportion of smokers in the general population in the Netherlands in 2014 (23 %) [30]. Model fit using unweighted data As a first step in the model fitting phase, a model including all latent variables possibly associated with the intention to provide (more) cessation support in the near future was estimated (model in the top left in Fig. 2). Model 1 showed suboptimal fit to the data, as indicated by the chi-square statistic (χ2 (529) = 4273, p < 0.001), and the RMSEA (0.109). The CFI (0.989) indicated the model fitted the data well, however some of the estimated variances were negative; another indication of model misfit. As can be partly observed from the standardized parameter estimates in Fig. 2, the regression coefficient estimates from subjective norm to intention, and from smoking behaviour to intention were found to be weak and non-significant (SN- > INT = 0.199, SE = 0.179, p = 0.267; SMO- > INT = −0.037, SE = 0.154, p = 0.812). These two latent variables were therefore removed in Model 2. This resulted in a lower chi-square statistic (χ2 (327) = 3498, p < 0.001), but did not improve the RMSEA (0.127) nor the CFI (0.936). For the third model presented in Fig. 2, covariances have been specified based on modification indices. Six covariances that led to an expected significant change in chi-square (with α = 0.01; power = 0.9; delta ≥ 0.2) were added to the model: three between PB and PBC, PB and ATT, PBC and ATT and three and three covariances between pairs of items. These were the items “Important theme to discuss” and “Involvement mental health care important”; “No smoking near patients” and “Ban in clinic” and the PBC items “It is possible in work routine” and “I miss the skills”. Compared to Model 1 and 2, Model 3 showed a better fit to the data: χ2 (322) = 790, p < 0.001); RMSEA = 0.049; CFI = 0.991. All three models were identified. In none of the models, any samples were lost due to non-convergence or other analysis problems.Fig. 2 Initial Model 1, Model 2 with non-significant paths removed, and optimized Model 3. Figure note: SN: subjective norm, PBC: perceived behavioural control; ATT: behavioural attitude; PB: past behaviour (providing cessation support to patients); SMO: current smoking behaviour; INT: intention to provide cessation support to patients in the near future; Model 1: Initial model with all possible relevant constructs included; Model 2: Initial model after insignificant paths and constructs have been removed; Model 3: Final model based on model 2 with covariances specified based on modification indices; The numbers in the paths between two latent constructs are standardized parameter estimates. Covariance matrices for the three models are available as Additional file 2. Analyses are based on unweighted data Model fit using weighted data As it was not directly possible to include the survey weights in the SEM analysis using the DWLS estimator, we chose to test the sensitivity of the model fitting results to the weighted characteristics of the data. The optimized model (Model 3) fitted originally to the unweighted data was fitted iteratively (1000 times) to the bootstrapped datasets, to compare the median fit - with 95 % confidence intervals (CIs) under the weighted bootstrapping approach to the result under the unweighted approach. Figure 3 presents the model with standardized parameter estimates from the median fitting dataset, and the two models fitted using the upper and lower 95 % CI datasets (fit was evaluated using the chi-square statistic).Fig. 3 Median model, 95 % CI worst fitting, and 95 % CI best fitting model. Figure note: PBC: perceived behavioural control; ATT: behavioural attitude; PB: past behaviour (providing cessation support to patients); INT: intention to provide cessation support to patients in the near future; Model ‘Median fit’: Final optimized model identified using bootstrapped data (1000 iterations) using the median fitting data (based on χ 2): 50 % of the bootstrapped datasets fitted the model better, 50 % fitted the model worse; Model ‘95 % CI worst fit’: 97.5 % of the bootstrapped datasets fit the model better, 2.5 % fit the model worse (based on χ 2); Model ‘95 % CI best fit’: 2.5 % of the bootstrapped datasets fit the model better, 97.5 % fit the model worse (based on χ 2); The numbers in the paths between two latent constructs represent standardized parameter estimates, Covariance matrices for the three presented fits are available as Additional file 2 The fit indices for the bootstrapped data with median model fit were χ2 (322) = 1188, p < 0.001); RMSEA = 0.067; CFI = 0.983; the ‘95 % CI worst fit’ indices were χ2 (322) = 1518, p < 0.001); RMSEA = 0.079; CFI = 0.975; and the ‘95 % CI best fit’ indices were χ2 (322) = 953, p < 0.001); RMSEA = 0.057; CFI = 0.988. The median model replicated the pattern and strength of the associations observed when unweighted data were used. Discussion The findings of this study indicated that in general, mental health staff in the Netherlands support encouraging patients more to quit smoking. The majority of the staff members feels capable to provide cessation support if needed, however only a minority of them intends to actually provide support over the next year. More than half of them have no experience in helping a patient quit smoking. Theoretically derived constructs associated with intentions to provide smoking cessation support to patients were identified. Attitudes towards providing cessation support, perceived behaviour control and past experience in providing support were strongly associated with the intention to provide future support. For subjective norms toward smoking (cessation) for patients and respondents own smoking behaviour we found limited evidence of an association with intention. The limited association between subjective norms and intention is in line with previous findings. In the meta-analysis by Armitage and Conner [15] it is reported that subjective norm is more weakly correlated with intention than attitude and perceived behavioural control. A number of possible explanations for this limited correlation have been suggested. Some argue that the lack of association between the two indicates that intentions are influenced primarily by intra-personal factors and not as much by what others are perceived to think or do [14, 31]. Another explanation is in the distinction between injunctive norms (i.e. what significant others think the person ought to do) and descriptive norms (i.e. what significant others themselves do). The subjective norm component of the TPB is an injunctive social norm (perceived social pressure, in this case: perceived strictness of smoking policies), while the results of a meta-analysis based upon 14 TPB studies involving a total sample size of N = 5810, covering a wide range of behavioural domains, provides strong evidence in support of the predictive validity of descriptive norms, over injunctive norms [31]. In addition, the inconsistent findings in the literature regarding the impact of strict smoking policies on smoking prevalence among patients and staff of mental health institutes [32–34] are also in line with the theoretically derived finding that there is only a weak link between subjective norms/policy regarding smoking cessation support and the intention to provide cessation support. The absence of a direct association between staff smoking behaviour and their intentions to provide cessation support has some precedents in the literature, although findings are mixed. A large cross-sectional survey of 3482 nurses working in 35 hospitals in the USA, did not find differences between smoking and non-smoking nurses in the likelihood that nurses asked patients about smoking, gave cessation advice, assessed willingness to quit, assisted in quitting or recommended medications/referred to a quit line [13]. In a recent study performed in Czech Republic, the same author found mixed results [35]. In a study by Slater and colleagues [36], 1074 smoking nurses rated the need for and potential of the nurse’s role in patients’ smoking cessation lower than non-smokers and ex-smokers. However, smoking and ex-smoking nurses rated their responsibility to help patients who wanted to quit higher than non-smokers. Limitations The reported findings in this study and its implications should be interpreted in the light of the limitations. A first limitation of this study is that the survey is cross-sectional in nature, thereby hampering the possibility to (longitudinally) model the impact of the evaluated constructs on actual behaviour (i.e. provision of smoking cessation support). Based on a meta-analysis that included 47 experimental tests of intention–behaviour relations, it is known that a medium-to-large change in intention (d = 0.66) leads only to a small-to-medium change in behaviour (d = 0.36) [37]. Thus, intention has a significant impact on behaviour, but the size of this effect is considerably small. A second limitation of this study is that the sample is comprised of self-referred respondents from mental health institutes. Therefore, the representativeness of the sample is a matter for debate. An assumption underlying the presentation of the results as potentially generalizable to the wider population of mental health care providers is that the associations between variables in this study would also have been found in a representative sample of mental health workers. Survey weights were calculated in order to optimize the representativeness of our sample regarding type of organization the participant worked for, number of inhabitants of the province where the professional worked, gender, age, part time factor and type of function. A third limitation is that although the fit of Model 3 and the bootstrapped model (Median) was acceptable or nearly acceptable according to common cut-off points for CFI (≥.95) [38] and RMSEA (<.07) [39], there still was considerable misfit of the model to the data, as evidenced by the chi-square statistic. In addition, in order to construct Model 3 from Model 2, modification indices were used to identify data-driven model optimisations in the form of the inclusion of six covariance paths to the model. Moreover, although the covariances between the latent constructs and measurement items have face validity, it should be acknowledged that post hoc modifications to models, for example based on modification indices, should be done sparingly and only when the modifications are plausible [27, 40]. Implications For many years the mental health treatment community tolerated or even encouraged smoking [4]. To date, mental health professionals and treatment organisations respond differently to this challenge. Although progress has been made in recent years, many (45.3 % in the current sample) mental health workers have never addressed their patients’ smoking behaviour. This study has some implications for future interventions to further promote these cessation support activities among mental health staff. Based on our results, it is best to address staff attitudes towards providing support, and to increase their perceived behavioural control towards supporting patients to quit smoking. The third identified correlate of intention, past cessation support behaviour, cannot directly be influenced. It should be acknowledged that changes in clinicians’ behaviour tend not to happen overnight [41]. There is however some evidence that an implementation strategy to support mental health professionals in providing smoking cessation support should focus on changing attitudes and perceived behavioural control, based on our and previous [20, 21] findings. Based on findings and frameworks developed in the implementation science discipline, features such as including a focus on engaging stakeholders and iterative Deming cycles (“plan-do-check-act”) in addition to understanding and targeting determinants of behaviour are key to bring about change in professionals’ behaviour [41]. Conclusion This study demonstrated that attitudes, perceived behavioural control and past behaviour of mental health workers are the strongest correlates of intention toward providing smoking cessation support to patients among the five theory-derived constructs tested in a SEM approach. Subjective norms and the mental health workers’ smoking behaviour were found to be notably less strong correlates of intention. These findings are to a great extent in line with previous findings and underline inconsistencies in the literature regarding the association between health workers’ smoking behaviour and their intentions to help patients stop smoking. Based on our findings, an implementation strategy to provide mental health care patients with smoking cessation support should best target staff attitudes and perceived behavioural control. Additional files Additional file 1: List of original survey items and response options. (XLSX 17 kb) Additional file 2: Covariance matrices of the fitted structural equation models. (XLSX 50 kb) Abbreviations ATTAttitudes BScBachelor of science CFIComparative fit index CIConfidence interval DWLSDiagonally weighted least squares INTIntention PBPast cessation support behaviour PBCPerceived behavioural control RMSEARoot mean square error of approximation SDStandard deviation SEMStructural equation modelling SMOSmoking behaviour SNSubjective norms TPBTheory of planned behaviour Acknowledgements Time to write this manuscript was provided by the Trimbos Institute, the Netherlands Institute of Mental Health and Addiction. Since 2013, the Netherlands Expertise Centre for Tobacco Control (NET) is hosted by the Trimbos Institute. NET is funded by the Netherlands Ministry of Health, Welfare and Sport. Funding This work was supported by the Netherlands Ministry of Health, Welfare and Sport. Availability of data and materials The datasets generated during and/or analysed during the current study are not publicly available due to organisation policy but are available from the corresponding author on reasonable request. Authors’ contributions MB, RB, ML designed and performed the survey study underlying this article, RG advised in the design and performance of the study, MB performed the analyses for this article, MB, RB, PH, RG, ML wrote the article and commented on earlier drafts. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate All participants provided informed consent before participating in the survey, in line with the Dutch Medical Research Involving Human Subjects Act. Based on previous consultation with the Netherlands’ Central Committee on Research Involving Human Subjects, survey research as performed for this study is exempted from medical ethics approval. ==== Refs References 1. Callaghan RC Veldhuizen S Jeysingh T Orlan C Graham C Kakouris G Remington G Gatley J Patterns of tobacco-related mortality among individuals diagnosed with schizophrenia, bipolar disorder, or depression J Psychiatr Res 2014 48 1 102 110 10.1016/j.jpsychires.2013.09.014 24139811 2. de Leon J Diaz FJ A meta-analysis of worldwide studies demonstrates an association between schizophrenia and tobacco smoking behaviors Schizophr Res 2005 76 2–3 135 157 10.1016/j.schres.2005.02.010 15949648 3. Kalman D Morissette SB George TP Co-morbidity of smoking in patients with psychiatric and substance use disorders Am J Addict 2005 14 2 106 123 10.1080/10550490590924728 16019961 4. Schroeder SA Morris CD Confronting a neglected epidemic: tobacco cessation for persons with mental illnesses and substance abuse problems Annu Rev Public Health 2010 31 297 314 10.1146/annurev.publhealth.012809.103701 20001818 5. World Health Organisation WHO Report on the Global Tobacco Epidemic, 2015 2015 Geneva World Health Organisation 6. U.S. Department of Health and Human Services The health consequences of smoking: 50 years of progress. A report of the surgeon general 2014 Atlanta U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health 7. Tsoi DT Porwal M Webster AC Interventions for smoking cessation and reduction in individuals with schizophrenia Cochrane Database Syst Rev 2013 2 CD007253 23450574 8. Kerr S Woods C Knussen C Watson H Hunter R Breaking the habit: a qualitative exploration of barriers and facilitators to smoking cessation in people with enduring mental health problems BMC Public Health 2013 13 221 221 232 10.1186/1471-2458-13-221 23497231 9. Smoking and mental health. A joint report by the Royal College of Physicians and the Royal College of Psychiatrists; 2013. http://archive.is/J9ibj. 10. Roberts E Evins AE McNeill A Robson D Efficacy and acceptability of pharmacotherapy for smoking cessation in adults with serious mental illness: A systematic review and network meta-analysis Addiction 2015 25756739 11. Blankers M, Buisman R, ter Weijde W, van Laar M. Rookbeleid in de GGZ: Een verkenning van beleid en praktijk tijdens verblijf in geïntegreerde GGZ-instellingen, verslavingszorginstellingen en RIBW’s. [Smoking policy in mental health care facilities: Exploration of policy and practise for inpatients of integrated mental health care institutions, substance abuse treatment centres, and regional institutions for sheltered housing] Utrecht: Trimbos-instituut; 2015. http://archive.is/bzdxH. 12. Dome P Lazary J Kalapos MP Rihmer Z Smoking, nicotine and neuropsychiatric disorders Neurosci Biobehav Rev 2010 34 3 295 342 10.1016/j.neubiorev.2009.07.013 19665479 13. Sarna L Bialous SA Wells M Kotlerman J Wewers ME Froelicher ES Frequency of nurses’ smoking cessation interventions: report from a national survey J Clin Nurs 2009 18 14 2066 2077 10.1111/j.1365-2702.2009.02796.x 19638062 14. Ajzen I The theory of planned behavior Organ Behav Hum Decis Process 1991 50 2 179 211 10.1016/0749-5978(91)90020-T 15. Armitage CJ Conner M Efficacy of the theory of planned behaviour: a meta-analytic review Br J Soc Psychol 2001 40 Pt 4 471 499 10.1348/014466601164939 11795063 16. McEachan RRC Conner M Taylor N Lawton RJ Prospective prediction of health-related behaviors with the Theory of Planned Behavior: A meta-analysis Health Psychol Rev 2011 5 97 144 10.1080/17437199.2010.521684 17. Ajzen I The theory of planned behaviour: reactions and reflections Psychol Health 2011 26 9 1113 1127 10.1080/08870446.2011.613995 21929476 18. Webb TL Sniehotta FF Michie S Using theories of behaviour change to inform interventions for addictive behaviours Addiction 2010 105 11 1879 1892 10.1111/j.1360-0443.2010.03028.x 20670346 19. Perkins MB Jensen PS Jaccard J Gollwitzer P Oettingen G Pappadopulos E Hoagwood KE Applying theory-driven approaches to understanding and modifying clinicians’ behavior: what do we know? Psychiatr Serv 2007 58 3 342 348 10.1176/ps.2007.58.3.342 17325107 20. McCarty MC Hennrikus DJ Lando HA Vessey JT Nurses’ attitudes concerning the delivery of brief cessation advice to hospitalized smokers Prev Med 2001 33 6 674 681 10.1006/pmed.2001.0944 11716666 21. Puffer S Rashidian A Practice nurses’ intentions to use clinical guidelines J Adv Nurs 2004 47 5 500 509 10.1111/j.1365-2648.2004.03129.x 15312113 22. Nederland GGZ Medewerkers@GGZ 2009 – Inventarisatie van personeel in de GGZ 2010 Amersfoort GGZ Nederland 23. Honaker J King G Blackwell M Amelia II: a program for missing data J Stat Softw 2011 45 7 1 47 10.18637/jss.v045.i07 24. R Core Team R: A language and environment for statistical computing 2015 Vienna R Foundation for Statistical Computing 25. Rosseel Y lavaan: An R Package for Structural Equation Modeling J Stat Softw 2012 48 2 1 36 10.18637/jss.v048.i02 26. Hoyle RH Isherwood JC Reporting results from structural equation modeling analyses in archives of scientific psychology Arch Sci Psychol 2013 1 1 14 22 23997997 27. Jackson DL Gillaspy JA Purc-Stephenson R Reporting practices in confirmatory factor analysis: an overview and some recommendations Psychol Methods 2009 14 1 6 23 10.1037/a0014694 19271845 28. Kline RB Principles and practice of structural equation modeling 2010 3 New York Guilford Press 29. Wolf EJ Harrington KM Clark SL Miller MW Sample size requirements for structural equation models: an evaluation of power, bias, and solution propriety Educ Psychol Meas 2013 76 6 913 934 10.1177/0013164413495237 25705052 30. Verdurmen J Monshouwer K van Laar M Factsheet continu onderzoek rookgewoonten 2014 [Fact sheet monitor smoking behaviour 2014] 2014 Utrecht Trimbos-instituut 31. Rivis AJ Sheeran P Descriptive norms as an additional predictor in the theory of planned behaviour: A meta-analysis Curr Psychol 2003 22 218 233 10.1007/s12144-003-1018-2 32. Lawn S Campion J Factors associated with success of smoke-free initiatives in Australian psychiatric inpatient units Psychiatr Serv 2010 61 3 300 305 10.1176/ps.2010.61.3.300 20194408 33. Etter M Khan AN Etter JF Acceptability and impact of a partial smoking ban followed by a total smoking ban in a psychiatric hospital Prev Med 2008 46 6 572 578 10.1016/j.ypmed.2008.01.004 18289658 34. Ratschen E Britton J Doody G McNeill A Smoking attitudes, behaviour and nicotine dependence among mental health acute inpatients: an exploratory study Int J Soc Psychiatry 2010 56 2 107 118 10.1177/0020764008101855 20207674 35. Sarna LP Bialous SA Králíková E Kmetova A Felbrová V Kulovaná S Malá K Roubíčková E Wells MJ Brook JK Tobacco cessation practices and attitudes among nurses in the Czech Republic Cancer Nurs 2015 25730594 36. Slater P McElwee G Fleming P McKenna H Nurses’ smoking behaviour related to cessation practice Nurs Times 2006 102 19 32 37 16711288 37. Webb TL Sheeran P Does changing behavioral intentions engender behavior change? A meta-analysis of the experimental evidence Psychol Bull 2006 132 2 249 268 10.1037/0033-2909.132.2.249 16536643 38. Hu LT Bentler PM Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives Struct Equ Model 1999 6 1 1 55 10.1080/10705519909540118 39. Steiger JH Understanding the limitations of global fit assessment in structural equation modelling Personal Individ Differ 2007 42 5 893 898 10.1016/j.paid.2006.09.017 40. MacCallum RC Hoyle RH Model specification: Procedures, strategies, and related issues Structural equation modeling: Concepts, issues, and applications 1995 Thousand Oaks Sage 41. Gonzales R, Cattamanchi A. Changing clinician behavior when less is more. JAMA Intern Med. 2015;175(12):1921–2. doi:10.1001/jamainternmed.2015.5987.

PMC005xxxxxx/PMC5000433.txt

==== Front BMC Med Res MethodolBMC Med Res MethodolBMC Medical Research Methodology1471-2288BioMed Central London 21610.1186/s12874-016-0216-1Research ArticleThe use of rapid review methods in health technology assessments: 3 case studies http://orcid.org/0000-0002-6185-8321Kaltenthaler Eva e.kaltenthaler@sheffield.ac.uk Cooper Katy k.l.cooper@sheffield.ac.uk Pandor Abdullah a.pandor@sheffield.ac.uk Martyn-St. James Marrissa m.martyn-stjames@sheffield.ac.uk Chatters Robin r.chatters@sheffield.ac.uk Wong Ruth ruth.wong@sheffield.ac.uk School of Health and Related Research (ScHARR), University of Sheffield, Regent Court, 30 Regent Street, Sheffield, S1 4DA UK 26 8 2016 26 8 2016 2016 16 1 10813 1 2016 17 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Rapid reviews are of increasing importance within health technology assessment due to time and resource constraints. There are many rapid review methods available although there is little guidance as to the most suitable methods. We present three case studies employing differing methods to suit the evidence base for each review and outline some issues to consider when selecting an appropriate method. Methods Three recently completed systematic review short reports produced for the UK National Institute for Health Research were examined. Different approaches to rapid review methods were used in the three reports which were undertaken to inform the commissioning of services within the NHS and to inform future trial design. We describe the methods used, the reasoning behind the choice of methods and explore the strengths and weaknesses of each method. Results Rapid review methods were chosen to meet the needs of the review and each review had distinctly different challenges such as heterogeneity in terms of populations, interventions, comparators and outcome measures (PICO) and/or large numbers of relevant trials. All reviews included at least 10 randomised controlled trials (RCTs), each with numerous included outcomes. For the first case study (sexual health interventions), very diverse studies in terms of PICO were included. P-values and summary information only were presented due to substantial heterogeneity between studies and outcomes measured. For the second case study (premature ejaculation treatments), there were over 100 RCTs but also several existing systematic reviews. Data for meta-analyses were extracted directly from existing systematic reviews with new RCT data added where available. For the final case study (cannabis cessation therapies), studies included a wide range of interventions and considerable variation in study populations and outcomes. A brief summary of the key findings for each study was presented and narrative synthesis used to summarise results for each pair of interventions compared. Conclusions Rapid review methods need to be chosen to meet both the nature of the evidence base of a review and the challenges presented by the included studies. Appropriate methods should be chosen after an assessment of the evidence base. Keywords Rapid review methodsHealth technology assessmentSystematic reviewhttp://dx.doi.org/10.13039/501100000664Health Technology Assessment Programmeissue-copyright-statement© The Author(s) 2016 ==== Body Background Systematic reviews have long been a key component of evidence based medicine. The use of methods to expedite systematic reviews is ever increasing due to time and resource constraints as well as policy maker and clinical demand. Systematic reviews typically take at least 12 months to conduct with rapid reviews taking between 3 weeks and 6 months [1]. Although the use of rapid review methods is increasing, there is little agreement as to how they are defined and what methodologies should be used [2]. Tsertsvadze et al. [3] describe three ways in which systematic reviews may be done more quickly: 1) process parallelisation using several reviewers to perform tasks in parallel; 2) innovative technologies to assist with tasks such as study selection, data extraction and risk of bias assessment and 3) modification of systematic review methodologies such as restricting or bypassing one or more steps in the process. Modifications to standard systematic review methods may include: highly refined research questions, limited searching [4], reduced number of reviewers for sifting and data extraction, restricted study design and limited quality assessment [1] and updating existing reviews [5]. The level and detail of analyses and synthesis is often reduced and existing systematic reviews may be summarised [6] or evidence summaries produced [7, 8]. Rapid review methods may be described as those that seek to reduce the time associated with systematic review methods in a way that will have the least impact on the validity or utility of the results. The reviewer must decide which modifications to standard methods will do this, given review objectives, time constraints and other challenges. Rapid reviews are particularly important in the field of health technology assessment (HTA) where they are used to support informed decision making [7]. Rapid review methods are not unique to HTA, although the need for timely evidence to underpin the assessment of new technologies makes them particularly relevant in this context. There are many inherent limitations associated with the use of rapid reviews such as the risk of introducing publication bias due to reduced searching. Grant and Booth [9] suggest that by limiting quality assessment and appraisal of evidence, disproportionate emphasis may be placed on poorer quality studies and lack of attention to synthesis may overlook inconsistencies or contradictions in the data. In addition, rapid reviews are less likely to use external experts and peer review and therefore they have potentially less scrutiny from clinical and methodological experts [4]. Methodological details of the rapid review process are often not mentioned or poorly described in reviews that use these methods [10]. In addition, the limitations associated with the chosen rapid review approach are frequently not discussed [2]. Recommendations for conducting rapid reviews include the need for replicability and transparency of methods [5]. This includes the need for a clearly stated research question, inclusion criteria, search strategies, inter-rater agreement process, data extraction and synthesis methods and conclusions. Also essential is a description of the limitations of the chosen methods. A recent summit of evidence synthesis experts developed a rapid review research agenda which included the need for a rapid review taxonomy and definitions and the need for methodological guidelines [11]. There is little guidance currently available regarding the most suitable methods to use and no indication as to whether all methods are suitable for all rapid reviews. Reviewers are faced with a dilemma as to which rapid review methods to use to best suit the evidence base available and the time constraints of the review they are undertaking. The National Institute for Health Research (NIHR) within the United Kingdom commissions rapid reviews in the form of short reports on a range of topics. We present three case studies of recent rapid reviews undertaken for the NIHR programme undertaken by the authors of this paper. These reviews were each undertaken within 8–12 weeks (from agreement of the final protocol to delivery of draft report) and all three reviews included almost entirely randomised controlled trials (RCTs) due to the limited time available to conduct the reviews. Three reviews from the same programme were chosen as they used identical processes for developing research questions, adherence to PRISMA reporting standards, report templates and peer review. Differences between the three reviews regarding the rapid review methods used arose from the differing evidence bases and challenges for each review. In this paper we present three distinctly different approaches to rapid reviewing and propose issues to consider when selecting appropriate rapid review methods for use in health technology assessments. Methods The three most recently completed NIHR rapid reviews undertaken at the School of Health and Related Research, University of Sheffield, United Kingdom were included in these analyses. The three reports included in the analyses were:The effectiveness of sexual health interventions for people with severe mental illness (SMI): a systematic review [12]. Interventions to treat premature ejaculation: a systematic review short report [13]. Psychological and psychosocial interventions for cannabis cessation in adults: A systematic review short report [14]. These three reviews were chosen as they had distinctly different evidence bases requiring different approaches to rapid review, thus enabling us to explore a range of rapid review approaches. The commissioning brief or questions to be answered by the reviews were set by the NIHR and it was a requirement that all interventions listed in the commissioning brief be included in the reviews. The protocols for all three reviews were developed from the NIHR commissioning briefs, peer reviewed, agreed with NIHR and registered on the PROSPERO website and published on the NIHR website. All three reviews were undertaken by at least two experienced systematic reviewers with an experienced information specialist undertaking the literature searches. All three reviews had extensive searching, data extraction and quality assessment, intensive input from clinical experts and all underwent peer review of the draft report by independent external reviewers as part of the standard NIHR peer review processes. The reviews all adhered to PRISMA reporting standards and are all published as part of the NIHR monograph series. For the present analyses, data were extracted from each of the short reports on the following variables:Rapid review methods used (reporting of outcomes, synthesis methods used). Research question and aims. Number of interventions and comparators. Number of included studies. Perceived reviewing challenges. Rationale for choice of review method in each case. Reported strengths of rapid review methods. Reported weaknesses of rapid review methods. These variables were determined by the whole research team as they were deemed important in order to understand the approach and challenges of the chosen methods in each review. The authors of each of the three reviews contributed to the data extraction and interpretation for the reviews they authored. Results The three reviews included in these analyses all used rapid review methods due to the short time frame of this HTA process. The rapid review methods used in all three reviews were: the use of a focussed research question, partial double sifting of titles and abstracts and partial double data extraction by a second reviewer. Two reviews included only RCTs or existing systematic reviews of RCTs [12] [14] while one review also included one non-randomised trial [13]. The quality of the included reviews has been assessed using the AMSTAR checklist [15] as shown in Table 1. Although rapid review methods were used in these reviews they are of relatively high quality according to the AMSTAR checklist.Table 1 AMSTAR assessment of the three reviews Question Review 1 Sexual health Review 2 Premature ejaculation Review 3 Cannabis cessation 1. Was an 'a priori' design provided? Yes, published protocol with research questions and inclusion criteria. Yes, published protocol with research questions and inclusion criteria. Yes, published protocol with research questions and inclusion criteria. 2. Was there duplicate study selection and data extraction? Yes All abstracts and full text articles assessed by two reviewers, data extracted by one reviewer, checked by another. Yes (partial) Titles and abstracts of citations identified by the searches were screened for potentially relevant studies by one reviewer and a subset checked by a second reviewer (and a check for consistency undertaken). Full texts were screened by two reviewers. One reviewer performed data extraction of each included study. All numerical data were then checked against the original article by a second reviewer. Any disagreements were resolved by a third reviewer. Yes (partial). Titles and abstracts of citations identified by the searches were screened for potentially relevant studies by one reviewer and a 10 % sample checked by a second reviewer (and a check for consistency undertaken). Full texts were screened by two reviewers. One reviewer performed data extraction for each included study. All numerical data were checked against the original article by a second reviewer and any disagreements were resolved through discussion. 3. Was a comprehensive literature search performed? Yes comprehensive searching reported. Yes comprehensive searching reported. Yes comprehensive searching reported. 4. Was the status of publication (i.e. grey literature) used as an inclusion criterion Yes, Grey literature was searched, non-English papers excluded Yes, some Grey literature was searched, non-English papers excluded Yes, Grey literature was searched, non-English papers excluded 5. Was a list of studies (included and excluded) provided? Yes tables of included and excluded studies both included Yes tables of included and excluded studies both included Yes tables of included and excluded studies both included 6. Were the characteristics of the included studies provided? Yes Yes Yes 7. Was the scientific quality of the included studies assessed and documented? Yes Yes Yes 8. Was the scientific quality of the included studies used appropriately in formulating conclusions? Yes Yes Yes 9. Were the methods used to combine the findings of studies appropriate? Yes Yes Yes 10. Was the likelihood of publication bias assessed? No No No 11. Was the conflict of interest included? Yes source of funding for included studies reported. No sources of funding for included studies not reported No sources of funding for included studies not reported AMSTAR score 10/11 9/11 9/11 Scoping searches and protocol development Scoping searches, based on a simple search strategy, were done while developing the protocols to provide a quick overview of the potential evidence base in terms of existing reviews, approximate numbers of relevant studies, relevant interventions and comparators as well as types of outcomes reported. Scoping searches were essential in estimating the number of records that would be retrieved. This informed the development of the protocol and selection of rapid review methods. A brief summary of the comprehensive search strategies and search results for the three reviews can be found in Table 2. Study design filters were necessary in all three reviews so that the total number of records was manageable in the time frame.Table 2 Search comparison table Sexual health of people with severe mental illness [12] Premature ejaculation [13] Cannabis cessation [14] Scoping search date November 2013 July 2013 January 2014 Design of included studies RCTs comparing sexual health interventions with usual care for adults with severe mental illness (SMI) RCTs of interventions for premature ejaculation (data extracted from existing reviews where available), or non-RCT evidence for any treatments where no RCTs exist RCTs of psychological or psychosocial interventions for cannabis cessation in regular users of cannabis Approaches to searching Electronic database; contact experts; reference tracking Electronic database; contact experts; reference tracking Electronic database; contact experts; reference tracking Sources to search Four electronic databases (3 health and 1 subject specific); conference proceedings database; clinical trials registers; UK/International mental health organisations Five electronic databases (2 health, 1 nursing and 1 multidisciplinary); conference proceedings databases; FDA or EMA websites Four electronic databases (3 health and 1 subject specific); conference proceedings database; clinical trials registers; UK/International societies and organisations Search strategy Single strategy (17 statements for Medline) with RCT filter. Included all terms for SMI and focused mental health terms (schizophrenia, schizoaffective or bipolar) combined with various broad sexual health-related terms. Results from the Medline scoping search was compared to see if RCTs from a known Cochrane review had been missed Focused strategy (8 statements for Medline) comprising terms for premature ejaculation with filters to identify RCTs, reviews or guidelines Single strategy (28 statements for Medline) with filters to identify reviews and RCTs. Cannabis terms (comprehensive) combined with broad psychotherapy and behavioural therapy terms (derived from the scoping search). The strategy was developed from three known Cochrane reviews. Title and abstract keywords were incorporated into the strategy and checked to see if known RCTs included in the reviews have been captured by the strategy Challenges Not all mental health condition terms included. Individual sexually transmitted infections or behavioural terms not searched, due to the anticipated large number of irrelevant papers that would be identified No specific intervention terms included due to large number of potential interventions. This increased potential for retrieval of a large number of records, but scoping searches indicated it would be a manageable number Not all individually named psychotherapy or behavioural therapy terms searched due to the anticipated large number of irrelevant papers that would be identified Records retrieved in Medline scoping search RCTs filter 684 RCTs filter 521; systematic reviews filter 653; cohort studies filter 596; guidelines filter 9 RCTs filter 361; systematic reviews filter 36 Main search date December 2013 August 2013 February 2014 Total records from main database search 2586 2283 1079 There was a process of iteration for each of the reviews in the form of specific questions from the reviewers to the policy makers to ensure that the proposed methods would meet the needs of the policy makers. In addition, the policy makers were able to comment on the draft protocols and revisions to the protocol incorporated in the final version. The approaches chosen were in part in order to meet the requirements of the policy makers as set out in the commissioning briefs to look at all relevant interventions. Table 3 summarises the key characteristics of the three included short reports and the rapid review approaches adopted by the review teams.Table 3 Summary characteristics of the three reviews, review challenges and approaches and strengths and limitations of chosen methods Report and no. RCTs Populations, interventions and comparators Review challenges and approaches Strengths and limitations of chosen method Kaltenthaler et al. 2014 [12] Sexual health of people with severe mental illness 13 RCTs Review aims: summarise effectiveness evidence, determine applicability in UK NHS setting and identify key areas for primary research. Population: people with severe mental illness Interventions: strategies to increase knowledge, assess and reduce sexual health risk, change behaviour and develop condom skills Comparators: educational sessions on HIV, money management or substance abuse, health promotion, wait list or no treatment Challenges due to evidence base:  • Wide variation in populations and settings (patients in psychiatric clinics, residential centres and homeless shelters)  • Wide range of outcomes including: biological (sexually transmitted infections, pregnancy), behavioural (number of partners, uptake of services, use of contraception/condoms) and proxy (knowledge, attitudes, behaviours, facilitators and barriers etc.) Approaches:  • Focussed definition of severe mental illness  • Brief summary of results presented, narrative synthesis, grouping of results from included studies by outcome (biological, behavioural and proxy) Strengths:  • Enabled rapid synthesis of a disparate evidence base to ensure policy makers were aware of areas where evidence was available. This informed the design of relevant RCTs Limitations:  • Quantitative data synthesis not generated for use by policy makers (only effect size by intervention and outcome)  • In-depth narrative synthesis not possible  • Non-RCT evidence excluded Cooper et al. (2015) [13] Premature ejaculation 101 RCTs and 1 CT (65 RCTs from existing reviews and 36 new RCTs and 1 new CT reports) Review aims: synthesise effectiveness evidence for behavioural, topical and systemic treatments. Population: men with premature ejaculation Interventions: topical anaesthetics, antidepressants, phosphodiesterase-5 inhibitors, opioid analgesics, behavioural therapies, acupuncture, Chinese medicine Comparators: placebo, wait list, other therapies Challenges due to evidence base:  • Very large number of RCTs (over 100) and existing systematic reviews covering wide range of interventions (several drug classes plus behavioural approaches)  • Several existing systematic reviews Approaches:  • Meta-analysis of primary outcome using data extracted from existing systematic reviews, with new primary study data added  • Narrative synthesis of secondary outcomes Strengths:  • Meta-analysis able to be used for primary outcome (consistent primary outcome)  • Use of data from existing reviews enabled meta-analysis of large dataset in shorter time Limitations:  • Potential for data errors or synthesis errors in original reviews to be repeated in new report  • Methodological quality of studies extracted from existing reviews not assessed separately  • Although use of data from existing reviews saved some time, triangulation of data from multiple reviews was still time-consuming  • Original RCT publications not revisited for data extraction and quality assessment. Cooper et al. (2015) [14] Cannabis cessation 33 RCTs Review aims: summarise effectiveness evidence for psychological and psychosocial interventions and identify key areas for primary research. Population: adults who use cannabis regularly Interventions: cognitive behavioural therapy, motivational interviewing, motivational enhancement therapy, supportive-expressive dynamic psychotherapy, social support groups, case management, contingency management (vouchers as incentive/reward) Comparators: waitlist, treatment as usual, other interventions, assessment only, education controls, written cannabis information, cannabis education Challenges due to evidence base:  • Wide variation in study populations (extent of cannabis dependence), interventions (type, duration) and comparators  • Very little consistency in outcome measures, time points, and statistics reported  • Large number of RCTs for a short report Approaches:  • For each pair of interventions compared, narrative summary of outcomes reported and how many showed a statistically significant effect Strengths:  • Inclusive approach, covering a wide range of populations, interventions and outcomes  • Included all psychosocial or psychological interventions undertaken in the adult, community dwelling population of cannabis users Limitations:  • Detailed numerical outcome data not presented, since outcome measures and statistics reported were so disparate  • Outcome measures in RCTs not converted to consistent measures to compare across studies as not feasible in timeframe CT controlled trial Review 1 The effectiveness of sexual health interventions for people with severe mental illness (SMI) [12] The aim of this review was to evaluate the effectiveness of sexual health interventions for people with SMI, determine their applicability to the UK NHS setting, and to identify key areas for primary research. Thirteen RCTs were included in the review. The challenges for this review were the inclusion of very diverse studies in terms of populations, interventions, comparators and reported outcomes. There was a large volume of non RCT, uncontrolled study evidence identified in the initial scoping searches which was not included as this was considered to be lower quality evidence. A narrative synthesis approach was chosen as meta-analysis was deemed impossible due to the considerable study heterogeneity. Difficulties were encountered with defining “severe mental illness” which includes different conditions (such as major depression) in some countries but not others. The approach taken was to briefly report all relevant outcomes; grouped into categories including biological, behavioural and proxy (such as barriers and facilitators) categories. Information on effectiveness of interventions was presented by reporting the p-values for outcomes as reported in the individual studies as well as the study authors’ conclusions. The strengths of this approach were that only higher quality evidence was presented; a thorough assessment of quality was undertaken and key details for each included study were readily accessible, particularly information on outcomes. By using a focussed and previously agreed definition of severe mental illness, only studies with populations directly relevant to the needs of the policy makers were included. Limitations of the chosen approach were the exclusion of non-RCT evidence meaning that some information might have been lost, particularly that related to the description of interventions. It was also only possible to report limited quantitative outcome data. However, due to the nature of the evidence base more extensive quantitative synthesis would most likely not have been possible. Our chosen approaches allowed us to fulfil the review objectives in that an overview of the effectiveness of the included interventions was provided, despite not being able to report full quantitative data. In addition, areas for future research and trial design were identified and information on applicability to the UK NHS was provided. Review 2 Interventions to treat premature ejaculation [13] The aim of the review was to systematically review the evidence for the clinical effectiveness of behavioural, topical and systemic treatments for premature ejaculation. The main challenge for this review was the large number of interventions and very large number of relevant RCTs (over 100). The approach taken was to use meta-analysis where appropriate using data extracted from existing systematic reviews. Data from newer primary studies were added to the meta-analyses. The review included 102 controlled trials. Data from 65 of these RCTs were extracted from existing reviews and 37 directly from additional RCT publications. Meta-analysis was possible as most studies used the same primary outcome (intra-vaginal ejaculatory latency time), in contrast to the other two short reports discussed here where outcome measures varied greatly across studies. Secondary outcomes were more varied and narrative synthesis was used for these. This approach had limitations as there was the potential for incorporation of synthesis errors from the original reviews. The methodological quality of studies extracted from the existing reviews was not assessed separately due to time constraints. Although the use of data from existing reviews saved some time, triangulation of data from multiple reviews was still time-consuming. The approaches chosen allowed us to provide up to date quantitative evidence on the effectiveness of a range of treatments for premature ejaculation. Review 3 Psychological and psychosocial interventions for cannabis cessation in adults [14] The aim of the review was to systematically review the effectiveness of psychological and psychosocial interventions for cannabis cessation in adults who use cannabis regularly and to identify key areas for primary research. This review included 33 RCTs. The cannabis cessation review included input from a service user who had previously received similar interventions to those included in the review. This individual provided feedback on the review protocol and final report; specifically, the included interventions, outcome measures and the lay person summary. The challenges associated with this review included a wide variation in study populations, such as the extent of cannabis dependence, interventions (type, duration, group or individual) and comparators. There was also very little consistency with regard to outcome measures, time points of measurement and statistics reported. Broad inclusion criteria were used as the commissioners requested the review to be inclusive of all relevant evidence. The approach taken was to present a narrative summary of outcomes reported for each intervention and comparison, stating how many showed a statistically significant effect. This enabled inclusion of the many relevant studies within the time constraints of the review. Due to the significant study heterogeneity, meta-analysis was not considered suitable. The approaches chosen in this review provided an overview of effectiveness of the interventions although it was not possible to report the quantitative data in full. This approach differed from that used in the first short report (sexual health interventions) in that due to the larger number of trials more limited outcome data was presented. Points to consider when determining a rapid review approach Based on the diverse range of approaches used in these three rapid reviews we have identified issues that are important to reflect on when planning a rapid review. Table 4 outlines a checklist of some items that should be considered when choosing a rapid review method, based on these case studies.Table 4 Checklist of items to consider when determining a rapid review approach 1. Assess the current evidence base-It is important to have an understanding of the evidence available before deciding which rapid review methods are most appropriate. Some points to consider are:   • Scoping searches - These are useful to estimate an approximate number of anticipated relevant studies.   • Existing systematic reviews - What are the search dates for the review (s) and the question answered by the review (s)? What is the methodological quality of the review(s)? This can be assessed using appropriate checklists. Did the review report a quality assessment of included studies? Consider using reported data to incorporate in a meta-analysis with newly identified studies.   • Summary of existing reviews - The findings of identified reviews could be presented plus a summary of any new studies using narrative synthesis. 2. Consider presentation of evidence-The complexity of the evidence base should be taken into account and an assessment made as to how much data should be presented and in what format. Some points to consider are:   • Meta-analysis Does the data support the use of meta-analysis?   • Outcome data Can limited data on outcomes be reported?   • Grouping of outcomes Can relevant outcomes be grouped to assist the reader in understanding the evidence base? 3. Ensure clear communication with policy makers - It is important that there is a common understanding between reviewers and policy makers as to the purpose of the review and the questions to be answered. Some points to consider are:   • In depth analysis Is it preferable to the policy maker to present an in depth analysis of a smaller selection of studies?   • Brief overview Is it preferable to the policy maker to present less information from a wider range of studies?   • Highlight gaps in the evidence Will it be helpful to the policy maker to highlight gaps in the evidence to inform future research? 4. Clearly report rapid review methods used - It is crucial that the reader understands what rapid review methods have been used and the impact this may have on the findings of the review. Points to consider are:   • Description of methods-Have the rapid review methods been transparently reported highlighting differences from standard systematic review methods?   • Discussion of limitations Have the potential limitations and biases of chosen methods been described. It is difficult to plan the review approach until there is a clear understanding of the type, amount and variation of evidence available and scoping searches are very useful for this. The incorporation of good quality existing systematic reviews may be considered. Presenting summaries of existing reviews and new studies or extracting data from existing reviews and incorporating into a meta-analysis with data from new studies are all options for consideration. Attention should be given to the most appropriate way to present the evidence, assessing both the amount of data to be presented and the most appropriate format. The level and type of evidence presented must be acceptable to both reviewers and policy makers. Clear communication with policy makers is crucial to ensure that the review being undertaken will address the question under consideration. This is especially important when using rapid reviews as the limited timeframe and available resources mean that there will be a trade-off between different aspects of the review such as thoroughness of searching, breadth of the research question and depth of analysis. It is essential that the methods used are clearly reported so that the reader is aware of the potential biases and limitations associated with chosen methods. Methods should be reported in enough detail so as to be reproducible and transparent. Discussion This paper presents three distinctly different approaches to the rapid review of evidence to address a pre-defined research question within a limited time frame. The approaches included a brief summary and grouping of results for the sexual health review [12], meta-analysis incorporating data from existing reviews and new RCTs for the premature ejaculation review [13] and reporting of a narrative summary of significant outcomes for each intervention/comparator pair in the cannabis cessation review [14]. All three reviews included other rapid review methods such as the use of a focussed research question, partial double sifting of titles and abstracts and partial double data extraction. Only numerical data was double data extracted in the cannabis and premature ejaculation reviews. Narrative synthesis was used, as opposed to full qualitative synthesis of evidence, in two of the reviews (sexual health interventions [12] and cannabis cessation [14]). Cameron et al. [16] suggest that rapid reviews may benefit from the rigour of external peer review. All three rapid reviews included in this study had expert advisory panels and were peer reviewed by a minimum of two independent external reviewers. They were also deemed to be of relatively high quality using the AMSTAR checklist with scores of 10/11, 9/11 and 9/11. Methods chosen for the rapid review of evidence must be both feasible and appropriate taking into account the requirements of the commissioners, the quantity and nature of the evidence and the time and resources available to do the review. The methods chosen for these three reviews were deemed acceptable to the commissioners as they approved the protocols where the methods were described in detail. The reasons for the commissioning of the reviews varied and this fed into the choice of review methods. For the sexual health review, the majority of evidence was from North America and it was not known how transferable this might be to the UK. The aim of the review was to inform the design of a UK based trial. The aim of the premature ejaculation review was to identify the most effective treatment option or combinations of treatment and the aim of the cannabis cessation review was to evaluate interventions for cannabis use and to identify important evidence gaps that might require further research. In all three reviews alternative methods could have been chosen such as the inclusion of fewer interventions, comparators or fewer outcomes, restriction of study design or the updating of existing systematic reviews. However, all of these options would have meant that either the needs of the commissioners were not met or that the high standards of the NIHR were not adhered to. Policy makers require evidence to make decisions in a timely manner therefore choices need to be made as to how to select, analyse and present the evidence required. We chose the approaches presented here in part because other approaches were not acceptable to the policy makers (limiting the number of interventions considered). Research by Cameron et al. [16] comparing the findings from rapid vs full reviews found no difference in the essential conclusions reached by the reviews. Traditional systematic reviews are more likely to provide greater depth of information than rapid reviews [1] although rapid reviews have been found to meet the needs of knowledge users [8]. It is however important, as suggested by Schünemann and Moja [17], to ensure that guidelines for review conduct and reporting are adhered to. Boundaries between rapid and full systematic reviews are often blurred and many published systematic reviews use rapid methods. The key factors identified in this study for consideration when selecting rapid review approaches include: an assessment of the evidence base, consideration of how to present the evidence, understanding the needs of the policy makers and adequate reporting of methods and their strengths and limitations. The range of reviewing options can then be considered by the reviewer once the size of the evidence base is ascertained including limiting the scope of the review [5] and streamlining processes for full text review and data extraction [1]. The incorporation of good quality existing systematic reviews is potentially very useful [6] [8] and can save valuable time and resources. Previous research has also highlighted the importance of reporting and communication with policy makers. Without a full description of methods the direction and magnitude of any risk of bias cannot be fully assessed [1]. This has real implications for policy makers when making decisions based on rapid reviews and it is essential that limitations are clearly described. Varker et al. [18] advocate the use of a reporting template in rapid reviews in order to ensure a consistent approach. Reviews must be “fit for purpose” so that they reflect the knowledge needs of the commissioning body [10]. Hartling et al. [19] also stress the range of methods employed are both driven and supported by close and ongoing communication between the producer of the review and the end user, which is a very different context from most standard systematic reviews. Feedback from policy makers is crucial to ensure that the use of rapid review methods did not hinder decision making and to determine what approaches are useful to them and which are not. There are several limitations to this research study. Only three rapid reviews were assessed, all from the same institution involving an overlap of reviewers. All three reviews were undertaken for the same HTA programme in the UK. Other programmes may have had other requirements and meant that other approaches were needed. Other reviewers from other organisations may have chosen different review methods. This limits the findings of the research. The authors of this paper were also the reviewers for the three case studies and therefore had in depth knowledge of the reasoning behind choices but other data extractors may have come to different conclusions compromising the replicability of this study. We deliberately looked at our own reviews so that we could provide an in depth description as to how and why decisions were made about which rapid review method to use, and what we found to be the strengths and limitations of these methods. More research is needed in this area to provide guidance to reviewers to enable them to choose the most appropriate rapid review methods. One possible approach would be to select a sample of rapid reviews produced by a range of groups that used the general approaches outlined in these case studies and then use this sample to compare both within and between the three case approaches to help establish best practice in this area. Another possible future research approach would be to compare the rapid review methods used with a full systematic review in the same area. We have recently published a paper comparing our chosen rapid review methods with a full review for one of the interventions from the premature ejaculation review [20]. We found that in the topic area primary outcome data were the same whether the de novo rapid review method or a full review method were employed. However, due to limited reporting across reviews, quality assessment of all RCTs could only be undertaken as part of the full systematic review. Finally, future research could explore the comparison of different rapid review approaches on the same topic, bearing in mind that not all approaches will be relevant and feasible for every topic. We did not receive feedback on whether or not our reviews met the needs of the commissioners, although this was requested. We have suggested this become part of the NIHR peer review process in future. Future research is needed on how best to incorporate feedback from commissioners and policymakers as to how useful the reviews were for decision making as well as potential limitations due to the chosen review methods. There is now a considerable amount of literature available on the use of rapid reviews with details of the methods available and limitations associated with these. There is little guidance on how to choose the most appropriate method for the evidence base identified. It is crucial that commissioners and policy makers have sufficient information to make a judgement on whether or not the chosen review approaches may be considered appropriate and robust. They must also be made aware of what approaches are feasible, bearing in mind the quantity and nature of the evidence as well as the time and resource constraints of the review. Details of the strengths and the limitations of the methods chosen must also be presented to commissioners and policy makers and their potential impact on decision-making. This research goes some way in exploring possible approaches suitable in this context. Conclusions There is no “one size fits all” to the use of rapid review methods. The analyses presented here suggest that the appropriate approach needs to be determined based on the evidence available, time constraints and the needs of policy makers and knowledge users. Authors need to be clear as to what approach was taken and the strengths and limitations of the rapid review methods chosen, how appropriate and robust these choices are and their potential impact on decision-making. Abbreviations HTAHealth technology assessment NIHRNational Institute for Health Research PICOPopulation, intervention, comparator, outcome RCTRandomised controlled trial Acknowledgements Not applicable. Funding This project was funded by the National Institute for Health Research Health Technology Assessment (NIHR HTA) Programme. The three reviews used in this project are published in full in the Health Technology Assessment journal series as projects 13/70/01, 13/12/01 and 12/74/01. Visit the HTA Programme website for more details (www.hta.ac.uk). Availability of data and materials The datasets analysed during the current study are available from the corresponding author on reasonable request. Authors’ contributions EK designed the study, undertook data extraction, drafted the initial version of the manuscript and undertook revisions to the manuscript. KC contributed to the study design, data extraction and revisions to the manuscript. AP contributed to the study design and data extraction and revisions to the manuscript. MMS undertook data extraction and revisions to the manuscript. RC contributed to the study design and revisions to the manuscript. RW contributed information on searching for the three reviews. All authors read and approved the final manuscript. Competing interests The authors of this paper were also the authors of the reviews used as the case studies. Consent for publication Not applicable. Ethics approval and consent to participate Not applicable. ==== Refs References 1. Ganann R Ciliska D Thomas H Expediting systematic reviews: methods and implications of rapid reviews Implement Sci 2010 5 56 10.1186/1748-5908-5-56 20642853 2. Harker J Kleijnen J What is a rapid review? A methodological exploration of rapid reviews in Health Technology Assessments Int J Evid Based Healthc 2012 10 397 410 10.1111/j.1744-1609.2012.00290.x 23173665 3. Tsertsvadze A Chen YF Moher D Sutcliffe P McCarthy N How to conduct systematic reviews more expeditiously? Systematic Reviews 2015 4 160 10.1186/s13643-015-0147-7 26563648 4. Watt A Cameron A Sturm L Lathlean T Blamey S Facey K Hailey D Norderhaug I Maddern G Rapid reviews versus full systematic reviews: an inventory of current methods and practice in health technology assessment Int J Technol Assess Health Care 2008 24 2 133 139 10.1017/S0266462308080185 18400114 5. Abrami PC Borokhovski E Bernard RM Wade CA Tamim R Persson T Bethel EC Hanz K Surkes MA Issues in conducting and disseminating brief reviews of evidence Evidence & Policy 2010 6 3 371 89 10.1332/174426410X524866 6. Chambers D Wilson P A framework for production of systematic review based briefings to support evidence-informed decision making Systematic Reviews 2012 1 32 10.1186/2046-4053-1-32 22775986 7. Khangura S Polisena J Clifford TJ Farrah K Kamel C Rapid review: an emerging approach to evidence synthesis in health technology assessment Int J Technol Assess Health Care 2014 30 1 20 27 10.1017/S0266462313000664 24451157 8. Khangura S Konnyu K Cushman R Grimshaw J Moher D Evidence summaries: the evolution of a rapid review approach Systematic Reviews 2012 1 10 10.1186/2046-4053-1-10 22587960 9. Grant MJ Booth A A typology of reviews: and analysis of 14 review types and associated methodologies Health Info Libr J 2009 26 91 108 10.1111/j.1471-1842.2009.00848.x 19490148 10. Watt A Cameron A Sturm L Lathlean T Babidge W Blamey S Facey K Hailey D Norderhaug I Maddern G Rapid versus full systematic reviews: validity in clinical practice? ANZ Surg 2008 78 1037 1040 10.1111/j.1445-2197.2008.04730.x 11. Polisena J Garritty C Umsheid CA Kamel C Samra K Smith J Vosilla A Rapid review summit: an overview and initiation of a research agenda Systematic Reviews 2015 4 111 10.1186/s13643-015-0096-1 26407674 12. Kaltenthaler E, Pandor A, Wong R. The effectiveness of sexual health interventions for people with severe mental illness: a systematic review. Health Technology Assessment. 2014;18(1). 13. Cooper K, Martyn-St James M, Kaltenthaler E, Dickinson K, Cantrell A. Interventions to treat premature ejaculation: A systematic review short report. Health Technology Assessment. 2015;19(21). 14. Cooper K, Chatters R, Kaltenthaler E, Wong R. Psychological and psychosocial interventions for cannabis cessation in adults: A systematic review short report. Health Technology Assessment. 2015;19(56). 15. Assessing the Methodological Quality of Systematic Reviews http://amstar.ca/Amstar_Checklist.php. Accessed 18 Apr 2016. 16. Cameron A Rapid versus full systematic reviews: an inventory of current methods and practice in Health Technology Assessment. ASERNIP-S Report No. 60 2007 ASERNIP-S Adelaide 17. Schünemann HJ Moja L Reviews: Rapid! Rapid! Rapid! And systematic Systematic Reviews 2015 4 4 10.1186/2046-4053-4-4 25589399 18. Varker T, Forbes D, Dell L, Weston A, Merlin T, Hodson S, O’Donnell M. Rapid evidence assessment: increasing the transparency of an emerging methodology. J Eval Clin Pract. 2015;doi: 10.1111/jep12405 19. Hartling L Guise JM Kato E Anderson J Belinson S Berliner E A taxonomy of rapid reviews links report types and methods to specific decision-making contexts J Clin Epidemiol 2015 68 12 141 148 10.1016/j.jclinepi.2015.05.036 20. Martyn-St JamesM, Cooper K, Kaltenthaler E. Methods for a rapid systematic review and meta-analysis in evaluating selective serotonin reuptake inhibitors for premature ejaculation. Evidence and Policy. 2016;(in press).

PMC005xxxxxx/PMC5000434.txt

==== Front Lipids Health DisLipids Health DisLipids in Health and Disease1476-511XBioMed Central London 30910.1186/s12944-016-0309-1ResearchOne-night sleep deprivation induces changes in the DNA methylation and serum activity indices of stearoyl-CoA desaturase in young healthy men Skuladottir Gudrun Valgerdur +354 525 4825gudrunvs@hi.is 1Nilsson Emil Karl emil.nilsson@neuro.uu.se 2Mwinyi Jessica jessica.mwinyi@neuro.uu.se 2Schiöth Helgi Birgir Helgi.Schioth@neuro.uu.se 21 Department of Physiology, Faculty of Medicine, University of Iceland, Vatnsmyrarvegur 16, IS-101 Reykjavik, Iceland 2 Department of Neuroscience, Functional Pharmacology, Uppsala University, Uppsala, Sweden 26 8 2016 26 8 2016 2016 15 1 13726 5 2016 16 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Sleep deprivation has been associated with obesity among adults, and accumulating data suggests that stearoyl-CoA desaturase 1 (SCD1) expression has a relevant impact on fatty acid (FA) composition of lipid pools and obesity. The aim of this study was to investigate the effect of one-night total sleep deprivation (TSD) on DNA methylation in the 5’-prime region of SCD1, and whether detected changes in DNA methylation are associated with SCD activity indices (product to precursor FA ratios; 16:1n-7/16:0 and 18:1n-9/18:0) derived from serum phospholipids (PL). Methods Sixteen young, normal-weight, healthy men completed two study sessions, one with one-night TSD and one with one-night normal sleep (NS). Sleep quality and length was assessed by polysomnography, and consisted of electroencephalography, electrooculography, and electromyography. Fasting whole blood samples were collected on the subsequent morning for analysis of DNA methylation and FAs in serum PL. Linear regression analyses were performed to assess the association between changes in DNA methylation and SCD activity indices. Results Three CpG sites close to the transcription start site (TSS) of SCD1 (cg00954566, cg24503796, cg14089512) were significantly differentially methylated in dependency of sleep duration (−log10P-value > 1.3). Both SCD-16 and SCD-18 activity indices were significantly elevated (P < 0.05) following one-night TSD, and significantly associated with DNA methylation changes of the three mentioned probes in the 5’ region of SCD1. Conclusion Our results suggest a relevant link between TSD, hepatic SCD1 expression and de-novo fatty acid synthesis via epigenetically driven regulatory mechanisms. Keywords DNA methylationFatty acid compositionMonounsaturated fatty acidsSleep conditionStearoyl-CoA desaturasethe Swedish Research Councilissue-copyright-statement© The Author(s) 2016 ==== Body Background Sleep deprivation has been associated with higher risk of weight gain and development of obesity among children and adults [1–3], which may provoke a higher susceptibility to chronic illnesses, such as diabetes [4, 5], and cardiovascular diseases [6, 7]. Human studies have demonstrated that sleep deprivation alters the central nervous system driven control of both hunger and appetite, provoking excessive food intake [8–10]. However, the knowledge of the complex and multifactorial mechanisms between sleep duration and increased risk of weight gain and obesity is still limited. A recently published study has demonstrated that rhythmic expression patterns of clock and selected clock-controlled genes in human blood cells are in part determined by exogenous factors, such as sleep and fasting state, and in part by the endogenous circadian timing system [11]. Furthermore, new results indicate that acute sleep loss alters the epigenetic and transcriptional profile of core circadian clock genes in key metabolic tissues [12], and that longer habitual sleep duration could ameliorate genetic predisposition to obesity via a favorable dietary profile [13]. The enzyme stearoyl-CoA desaturase (SCD), which is predominantly expressed in the liver, plays a central role in the desaturation of saturated fatty acids (FAs), thus having important implications in the metabolism of FAs and development of obesity. SCD catalyzes the biosynthesis of the monounsaturated FAs palmitoleate (16:1n-7) and oleate (18:1n-9) from the saturated FAs palmitate (16:0) and stearate (18:0), respectively [14]. In sated state the FAs of phospholipids (PL) and triglycerides (TG) in healthy liver originate from nonesterified FAs, dietary FAs incorporated in chylomicrons and FAs synthesized by hepatic de novo lipogenesis from dietary carbohydrates [15–17]. The de-novo FA synthesis in liver is well reflected by SCD activity indices estimated as ratios of 16:1n-7/16:0 and 18:1n-9/18:0 in blood TG as well as in PL [18–21]. Over the last decade there has been much interest in estimating the SCD activity as a putative biomarker for body fat regulation and development of obesity. Studies have shown that there is a very tight and complex regulation of SCD1 gene expression in response to various parameters including hormonal and nutrient factors [22, 23]. Furthermore, elevated expression levels of the human SCD1 gene are found to correlate both with the SCD enzyme activity [24], and obesity [25]. Recently, it has been demonstrated in healthy subjects, who have been also included in the current study that one-night of total sleep deprivation (TSD) alters clock gene regulation, concomitant with deleterious metabolic effects, which are differential across key peripheral metabolic tissues in healthy humans [12]. Hitherto, no study has addressed the effect of sleep deprivation on cytosine DNA methylation of SCD1 that might have implications for SCD activity, endogenous lipid biosynthesis and development of obesity. Therefore, we assessed DNA methylation of selected CpG sites within the SCD1 promoter and studied the associations with SCD activity indices derived from serum PL following both one-night normal sleep (NS) and one-night TSD in young healthy men. Methods Study cohort and design This study was based on prospectively collected data from a randomized crossover within-subject trial designed to examine the effects of one-night TSD on gene expression and DNA methylation of core circadian clock genes in peripheral tissues. The study was registered with ClinicalTrials.gov, number NCT01730742. Details of the study design have been published previously [12]. Sixteen young (age: 23.3 ± 3.4 (mean ± S.D.) years), normal-weight (BMI: 23.64 ± 2.35 kg/m2), and healthy men participated in two sessions of this study. In brief, on day 1 the participants were provided standardized meals including breakfast, lunch, snack, and dinner. The next night nocturnal sleep was either permitted from 22:30 h (lights off) to 06:30 h (lights on; one-night NS) or the participants remained exposed to light under constant supervision from 22:30 h to 06:30 h, remaining bed-restricted and fasted, i.e. one-night TSD. All participants were engaged in both conditions using a within-subject, randomized crossover design, where each condition was separated by 4 weeks. Written, informed consent was obtained from all participants and the regional ethical committee in Uppsala, Sweden approved the study. Sleep quality and length was assessed by polysomnography (Embla A10, Flaga hf, Reykjavik, Iceland), and consisted of electroencephalography, electrooculography, and electromyography. Blood samples were collected in EDTA tubes in the evening (19:30 h) of day 1, and in the morning after the intervention and before a standardized breakfast (07:30 h) of day 2. Fasting whole blood samples for DNA methylation analysis were immediately frozen in a 50/50 mixture of ethanol and dry ice before deposited in −80 °C. Serum samples were separated from fasting whole blood and stored in −80 °C before FA analysis. Sample preparation and methylation analyses Genomic DNA was extracted by robot assisted phenol/chloroform extraction at the Latvian Biomedical Research and Study Centre in Riga, Latvia. Bisulfite conversion of DNA and hybridization to the Illumina 450 K methylation Bead Chip (Illumina, San Diego, USA) was performed at the Science for Life Laboratory (SciLifeLab Uppsala, Sweden). Beta values representing the methylation status (0–100 %) of SCD1, localized on chromosome 10, were generated by GenomeStudio (Illumina, San Diego, USA). Determination of SCD-16 and SCD-18 activity The total serum lipid fraction was extracted with chloroform-methanol (2:1, v ⁄v), using a well-established method [26]. PL were separated on a TLC plate using the solvent system petroleum ether⁄diethyl ether ⁄acetic acid (80:20:1, v⁄v⁄v). The PL FAs were methylated with 14 % boron trifluoride in methanol, and the FA methyl esters analyzed by a HP Series II 5890, series A Gas Chromatograph (Hewlett Packard Co⁄Agilent, Palo Alto, CA, USA). The FA methyl esters were identified and calibrated against commercial standards (Sigma Chemical Co.; Nu-Check-Prep, Elysian, MN, USA). The results were expressed as percentage (%) of total FAs in serum PL. Activity indices of SCD-16 and SCD-18 were determined calculating the ratios of 16:1n-7/16:0 and 18:1n-9/18:0, respectively, derived from serum PL. Statistical analysis Statistical analyses of methylation data were performed with the software package R (version 3.1). Probes having the transcription start site (TSS) of SCD1 as closest TSS were extracted from the chip. Both upstream and downstream probes were included. Information about the exact distance to the TSS was extracted according to Price et al. [27]. A total of 19 SCD probes associated to SCD1 were identified and used in the analysis. Pairwise t-tests were performed to detect differences in DNA methylation between one-night NS and one-night TSD. Subsequently, we performed linear regression analyses regressing TSD induced changes in probe methylation against changes in the activity indices of SCD-16 (16:1n-7/16:0) and SCD-18 (18:ln-9/18:0) derived from serum PL. In all analyses the ratio of neutrophils to leukocytes was taken into account in order to prevent shifting monocyte subpopulations affecting the data (i.e. the regulation of immune and inflammatory responses). Wilcoxon signed-rank test (SPSS software, version 21.0; IBM Corporation, Somers, N.Y., USA) was used to compare intraindividual differences in 16:1n-7/16:0 and 18:ln-9/18:0 ratios before and after the two different sleep conditions. A P-value < 0.05 was considered statistically significant. Results Effect of one-night TSD on cytosine DNA methylation Methylation of three SCD1 probes (cg00954566, cg24503796, cg14089512) close to the TSS was significantly different between one-night NS and one-night TSD in unadjusted analyses (−log10P-value > 1.3; Fig. 1: Left). Other probes, although not significant, displayed a similar trend towards hypermethylation after one-night TSD as in one-night NS. This was especially observed for CpG sites within 2000 bp up- or downstream of the TSS. Here, 11 of 14 probes showed a hypermethylation between 0.1 and 1.8 % after one-night TSD as in one-night NS (Fig. 1: Right).Fig. 1 DNA methylation in relation to SCD1 position. Three of 19 SCD1 probes differ significantly in DNA methylation in dependency of sleep duration (one-night normal-sleep (NS) versus one-night total sleep deprivation (TSD); Left panel −log10 P-value). A similar trend towards hypermethylation was observed for probes close to the transcription start site (TSS) (Right panel Methylated beta-value) Association between cytosine DNA methylation and SCD activity indices CpG sites found to be significantly changed in their methylation pattern after TSD were subsequently investigated in linear regression analyses regressing change of methylation to SCD activity indices. The methylation of three CpG sites was associated with the activity indices of SCD-16, namely cg19191454 (positively associated, (P < 0.001), cg23508052 and cg11311579 (negatively associated, P = 0.01 and P = 0.05, respectively) (Table 1). Furthermore, the probe cg15022173 was negatively associated (P = 0.02), and cg07649988 positively associated (P = 0.05) with the activity indices of SCD-18.Table 1 Association between cytosine DNA methylation and SCD activity indices following one-night total sleep deprivation Probe ID Position Relative TSS (bp) Mean sleep % (sd) Mean wake % (sd) Slope P-value Slope P-value Chr 10 NM_005063 SCD-16 SCD-18 (16:1n-7/16:0) (18:1n-9/18:0) cg19191454 102106597 −175 4.4 (0.9) 4.2 (0.4) 0.016 <0.001 1.434 0.63 cg23508052 102106812 40 2.9 (0.5) 2.8 (0.5) −0.003 0.01 −0.892 0.80 cg11311579 102106755 −17 0.8 (0.2) 1.0 (0.2) −0.004 0.05 −6.887 0.38 cg01270221 102116383 9611 81.9 (1.3) 81.8 (1.2) −0.092 0.07 −0.175 0.93 cg06400428 102107667 895 11.1 (3.5) 10.6 (3.6) 0.015 0.10 0.590 0.48 cg16744911 102108975 2203 82.0 (2.2) 82.4 (1.8) 0.025 0.19 0.213 0.87 cg03440556 102107758 986 27.6 (6.2) 26.6 (6.3) 0.039 0.25 −0.298 0.69 cg12714759 102102352 −4420 21.5 (3.9) 21.4 (3.8) 0.011 0.27 0.270 0.73 cg02237755 102107926 1154 83.1 (3.3) 83.1 (3.7) 0.048 0.55 −0.591 0.48 cg15022173 102101047 −5725 76.3 (2.2) 76.9 (2.4) −0.420 0.56 −2.326 0.02 cg00653847 102100213 −6559 70.9 (4.1) 71.6 (3.5) −0.025 0.61 −0.515 0.63 cg24503796 102107677 905 13.8 (3.2) 14.5 (2.9) −0.018 0.65 0.766 0.53 cg07649988 102106577 −195 3.7 (0.7) 3.6 (0.8) −0.019 0.72 8.026 0.05 cg14089512 102106758 −14 9.6 (1.1) 10.3 (1.2) 0.048 0.75 0.180 0.92 cg07230380 102106359 −413 0.7 (0.6) 0.9 (0.5) 0.004 0.92 1.160 0.75 cg18328965 102107585 813 11.0 (1.7) 10.8 (1.8) 0.221 0.93 1.902 0.14 cg00699831 102116399 9627 87.1 (1.1) 87.2 (1.3) 0.236 0.95 2.899 0.08 cg26351966 102106081 −691 18.2 (2.9) 18.7 (3.2) 0.017 0.96 0.050 0.97 cg00954566 102106406 −366 4.4 (0.9) 4.3 (0.9) −0.002 0.97 0.474 0.90 The table shows selected SCD1 probes and SCD activity indices derived from serum phospholipids of young healthy men. Bold represents a statistically significant association at P < 0.05; software package R (version 3.1) Fatty acid composition and SCD activity indices There were no significant changes of the levels of the SCD precursor and product, 16:0 and 16:1n-7, and 18:0 and 18:1n-9, respectively, in fasting serum PL following one-night NS. However, there were significant changes of individual FA levels, that resulted in elevated (P < 0.05) SCD-16 and SCD-18 activity indices derived from fasting serum PL following one-night TSD (Table 2).Table 2 SCD activity indices before (day 1) and after (day 2) one-night normal-sleep (NS) or total sleep deprivation (TSD) NS (n = 12) TSD (n = 14) Day 1 Day 2 Day 1 Day 2 Fatty acids (% of total FAs)  16:0 26.78 ± 0.44 26.80 ± 0.35 26.58 ± 0.37 27.00 ± 0.24a  16:1n-7 0.41 ± 0.04 0.43 ± 0.04 0.37 ± 0.03 0.40 ± 0.02a  18:0 12.78 ± 0.34 12.00 ± 0.51 12.55 ± 0.22 12.35 ± 0.16a  18:1n-9 9.40 ± 0.33 9.54 ± 0.33 9.51 ± 0.25 9.74 ± 018a Activity indices  SCD16 0.0150 ± 0.0014 0.0159 ± 0.0014 0.0140 ± 0.0009 0.0147 ± 0.0009a  (16:1n-7/16:0)  SCD18 0.7641 ± 0.0351 0.8088 ± 0.0367 0.7591 ± 0.0198 0.7904 ± 0.0170a  (18:1n-9/18:0) The table shows the SCD activity indices derived from serum phospholipids of young healthy men Data are expressed as mean ± SEM a P < 0.05 compared with day 1, Wilcoxon signed-rank test Discussion We demonstrate, for the first time to our knowledge that one-night TSD is significantly associated with increased methylation in the 5’ prime region of the SCD1 gene. Importantly, we show that one-night TSD related methylation changes of SCD1 are significantly associated with changes in SCD-16 and SCD-18 activity indices derived from fasting serum PL followed by one-night TSD in young normal-weight healthy men, thus, describing a novel regulatory pathway by which TSD may influence homeostasis of body fat. Recently it was demonstrated in the subjects of the present study, that a one-night TSD alters the epigenetic and transcriptional profile of core circadian clock genes in key metabolic tissues [12], thus demonstrating TSD systematic changes in methylation of functionally important gene networks. Increased DNA methylation has been, in many cases, associated with a suppressed gene expression [28]. We show that one night TSD is associated with increased methylation in the 5’ prime region of the SCD1, which is in line with a recent study showing that methylation shifts are able to suppress or enhance gene expression, suggesting two different mechanisms of DNA methylation-dependent gene regulation [29]. Our findings that SCD1 methylation differs between one-night NS and one-night TSD strongly support the hypothesis that TSD is connected to epigenetic shifts that might have impact on lipid metabolism [30]. It is known from several experimental studies that changes in the activity of SCD indices estimated from plasma or tissues are accompanied by simultaneous changes in the transcriptional level of the SCD1 [31, 32]. The regulation of SCD1 gene is of considerable physiological importance, as a high SCD activity has been implicated in a wide range of disorders including diabetes, atherosclerosis, cancer, and obesity [14, 33, 34]. Thus, our observation, that one-night TSD is associated with increased methylation of the SCD1, may allow the speculation that TSD has the ability to contribute to the mentioned metabolic diseases via changes in the methylation pattern of genes, such as SCD1. Short sleep duration may increase obesity risk by causing small changes in eating patterns that cumulatively alter energy balance [35, 36]. Several studies have indicated associations between insufficient sleep and alterations in circulating hormones involved in feeding behavior, glucose metabolism, hunger, and appetite, which are probably involved in the development of metabolic disorders, such as obesity and diabetes [9, 10]. Recent study has shown that higher habitual sleep variability, but not habitual sleep duration, is significantly associated with abdominal obesity in adolescents, which can be partially explained by increased caloric intake, especially from carbohydrates [37]. The precise mechanism through which the brain regulates changes in hormone release with sleep deprivation is unknown, but one possibility is increased sympathetic nervous system activity [38]. Several studies have reported that diet affects the FA composition of serum PL [39–41]. Our observation indicates similar dietary habits and lifestyle factors such as cigarette smoking and alcohol consumption among the participants, since there was no significant difference found in FA levels of non-fasting serum PL between the two sleep conditions separated by 4 weeks (day 1). On the other hand, there is no general agreement which blood lipid fraction to use regarding the assessment of hepatic SCD activity indices. The majority of PL synthesis occurs in the endoplasmic reticulum of the liver, where PL associates with other lipids and proteins resulting in lipoproteins that are released into the bloodstream. Thus, we assume that in fasting blood the SCD-16 and SCD-18 activity indices derived from serum PL may mainly reflect hepatic SCD-1 activity. We observed significantly elevated SCD-16 and SCD-18 activity indices derived from fasting serum PL following one-night TSD and before a standardized breakfast. This strengthens the hypothesis that shortened sleep is an additional link to the dysregulation in energy metabolism that might impact on lipid metabolism and weight gain [30, 42]. This study is limited by its relatively small sample size of healthy individuals, which may preclude enough power to assess the interaction between SCD1 methylation and sleep deprivation. Another limitation of our study is that the methylation of SCD1 was measured only at a single time point, i.e., under fasting condition in the morning following each sleep intervention. Thus, conclusions about the causality have to be made carefully. However, our observations allow to suggesting a role of sleep duration and quality in regulating SCD1 expression and development of obesity. Conclusions Our study results indicate that one-night TSD modifies SCD1 methylation, FA levels of serum PL, and, thereby, changes in SCD-16 and SCD-18 activity indices. The exact mechanism how sleep restriction is implicated in SCD1 expression and development of obesity warrants further investigation. Acknowledgements This study was supported by the Swedish Research Council. We would like to thank Lara Bjorgvinsdottir for her assistance with fatty acid analysis. Authors’ contributions HBS and GVS designed the study. EKN was responsible for the DNA methylation analysis. GVS was responsible for the fatty acid analysis. EKN and GVS were responsible for the statistical analyses and interpretation of the data. GVS wrote the initial draft of the manuscript in cooperation with EKN. JM critically reviewed the manuscript. All authors contributed to the final version of the manuscript. Competing interests The authors declare that they have no competing interests. ==== Refs References 1. Taheri S The link between short sleep duration and obesity: we should recommend more sleep to prevent obesity Arch Dis Child 2006 91 881 4 10.1136/adc.2005.093013 17056861 2. Carter PJ Taylor BJ Williams SM Taylor RW Longitudinal analysis of sleep in relation to BMI and body fat in children: the FLAME study Br Med J 2011 342 d2712 10.1136/bmj.d2712 21622518 3. Krueger PM Reither EN Peppard PE Burger AE Hale L Cumulative exposure to short sleep and body mass outcomes: a prospective study J Sleep Res 2015 24 629 38 10.1111/jsr.12327 26211809 4. Marshall NS Glozier N Grunstein RR Is sleep duration related to obesity? A critical review of the epidemiological evidence Sleep Med Rev 2008 12 289 98 10.1016/j.smrv.2008.03.001 18485764 5. Guidolin M Gradisar M Is shortened sleep duration a risk factor for overweight and obesity during adolescence? A review of the empirical literature Sleep Med 2012 13 779 86 10.1016/j.sleep.2012.03.016 22633283 6. Mullington JM Haack M Toth M Serrador J Meier-Ewert H Cardiovascular, inflammatory and metabolic consequences of sleep deprivation Prog Cardiovasc Dis 2009 51 294 302 10.1016/j.pcad.2008.10.003 19110131 7. Moraes W Poyares D Zalcman I de Mello MT Bittencourt LR Santos-Silva R Tufik S Association between body mass index and sleep duration assessed by objective methods in a representative sample of the adult population Sleep Med 2013 14 312 18 10.1016/j.sleep.2012.11.010 23391395 8. Hanlon EC Cauter EV Quantification of sleep behavior and of its impact on the cross-talk between the brain and peripheral metabolism Proc Natl Acad Sci U S A 2011 108 suppl 3 15609 16 10.1073/pnas.1101338108 21852576 9. Cedernaes J Schiöth HB Benedict C Determinants of shortened, disrupted, and mistimed sleep and associated metabolic health consequences in healthy humans Diabetes 2015 64 1073 80 10.2337/db14-1475 25805757 10. Schmid SM Hallschmid M Schultes B The metabolic burden of sleep loss Lancet Diabetes Endocrinol 2015 3 52 62 10.1016/S2213-8587(14)70012-9 24731536 11. Möller-Levet CS Archer SN Bucca G Laing EE Slak A Kabiljo R Lo JCY Santhi N von Schantz M Smith CP Dijk D-J Effects of insufficient sleep on circadian rhythmicity and expression amplitude of the human blood transcriptome Proc Natl Acad Sci U S A 2013 110 E1132 41 10.1073/pnas.1217154110 23440187 12. Cedernaes J Osler ME Voisin S Broman J-E Vogel H Dickson SL Zierath JR Schiöth HB Benedict C Acute sleep loss induces tissue-specific epigenetic and transcriptional alterations to circadian clock genes in men J Clin Endocrinol Metab 2015 100 E1255 61 10.1210/JC.2015-2284 26168277 13. Dashti HS Follis JL Smith CE Tanaka T Cade BE Gottlieb DJ Hruby A Jacques PF Lamon-Fava S Richardson K Saxena R Scheer FA Kovanen L Bartz TM Perälä MM Jonsson A Frazier-Wood AC Kalafati IP Mikkilä V Partonen T Lemaitre RN Lahti J Hernandez DG Toft U Johnson WC Kanoni S Raitakari OT Perola M Psaty BM Ferrucci L Grarup N Highland HM Rallidis L Kähönen M Havulinna AS Siscovick DS Räikkönen K Jørgensen T Rotter JI Deloukas P Viikari JS Mozaffarian D Linneberg A Seppälä I Hansen T Salomaa V Gharib SA Eriksson JG Bandinelli S Pedersen O Rich SS Dedoussis G Lehtimäki T Ordovás JM Habitual sleep duration is associated with BMI and macronutrient intake and may be modified by CLOCK genetic variants Am J Clin Nutr 2015 101 135 43 10.3945/ajcn.114.095026 25527757 14. Ntambi JM The regulation of stearoyl-CoA desaturase (SCD) Prog Lipid Res 1995 34 139 50 10.1016/0163-7827(94)00010-J 7480063 15. Hudgins LC Hellerstein M Seidman C Neese R Diakun J Hirsch J Human fatty acid synthesis is stimulated by a eucaloric low fat, high carbohydrate diet J Clin Invest 1996 97 2081 91 10.1172/JCI118645 8621798 16. Barrows BR Parks EJ Contributions of different fatty acid sources to very low-density lipoprotein-triacylglycerol in the fasted and fed states J Clin Endocrinol Metab 2006 91 1446 52 10.1210/jc.2005-1709 16449340 17. Vedala A Wang W Neese RA Christiansen MP Hellerstein MK Delayed secretory pathway contributions to VLDL-triglycerides from plasma NEFA, diet, and de novo lipogenesis in humans J Lipid Res 2006 47 2562 74 10.1194/jlr.M600200-JLR200 16929033 18. Attie AD Krauss RM Gray-Keller MP Brownlie A Miyazaki M Kastelein JJ Lusis AJ Stalenhoef AF Stoehr JP Hayden MR Ntambi JM Relationship between stearoyl-CoA desaturase activity and plasma triglycerides in human and mouse hypertriglyceridemia J Lipid Res 2002 43 1899 1907 10.1194/jlr.M200189-JLR200 12401889 19. Carpentier YA Portois L Sener A Malaisse WJ Correlation between liver and plasma fatty acid profile of phospholipids and triglycerides in rats Int J Mol Med 2008 22 255 62 18636182 20. Kotronen A Seppänen-Laakso T Westerbacka J Kiviluoto T Arola J Ruskeepää AL Yki-Järvinen H Orešič M Comparison of lipid and fatty acid composition of the liver, subcutaneous and intra-abdominal adipose tissue, and serum Obesity (Silver Spring, Md) 2010 18 937 44 10.1038/oby.2009.326 21. Landry F Chan C-C Huang Z Leclair G Li CS Oballa R Zhang L Bateman K Plasma-based approach to measure target engagement for liver-targeting stearoyl-CoA desaturase 1 inhibitors J Lipid Res 2011 52 1494 9 10.1194/jlr.M013177 21642745 22. Mauvoisin D Mounier C Hormonal and nutritional regulation of SCD1 gene expression Biochimie 2011 93 78 86 10.1016/j.biochi.2010.08.001 20713121 23. Schwenk RW Jonas W Ernst SB Kammel A Jähnert M Schürmann A Diet-dependent alterations of hepatic Scd1 expression are accompanied by differences in promoter methylation Horm Metab Res 2013 45 786 94 10.1055/s-0033-1348263 23803969 24. Ntambi JM Miyazaki M Regulation of stearoyl-CoA desaturases and role in metabolism Prog Lipid Res 2004 43 91 104 10.1016/S0163-7827(03)00039-0 14654089 25. Hulver MW Berggren JR Carper MJ Miyazaki M Ntambi JM Hoffman EP Thyfault JP Stevens R Dohm GL Houmard JA Muoio DM Elevated stearoyl-CoA desaturase-1 expression in skeletal muscle contributes to abnormal fatty acid partitioning in obese humans Cell Metab 2005 2 251 61 10.1016/j.cmet.2005.09.002 16213227 26. Folch J Lees M Stanley GHS A simple method for the isolation and purification of total lipids from animal tissues J Biol Chem 1957 226 497 509 13428781 27. Price ME Cotton AM Lam LL Farré P Emberly E Brown CJ Robinson WP Kobor MS Additional annotation enhances potential for biologically-relevant analysis of the Illumina Infinium HumanMethylation450 BeadChip array Epigenetics Chromatin 2013 6 4 10.1186/1756-8935-6-4 23452981 28. Bird A DNA methylation patterns and epigenetic memory Genes Dev 2002 16 6 21 10.1101/gad.947102 11782440 29. Wan J Oliver VF Wang G Zhu H Zack DJ Merbs SL Qian J Characterization of tissue-specific differential DNA methylation suggests distinct modes of positive and negative gene expression regulation BMC Genomics 2015 16 49 10.1186/s12864-015-1271-4 25652663 30. Killick R Banks S Liu PY Implications of sleep restriction and recovery on metabolic outcomes J Clin Endocrinol Metab 2012 97 3876 90 10.1210/jc.2012-1845 22996147 31. Sampath H Ntambi JM Stearoyl-coenzyme A desaturase 1, sterol regulatory element binding protein-1c and peroxisome proliferator-activated receptor-α: independent and interactive roles in the regulation of lipid metabolism Curr Opin Clin Nutr Metab Care 2006 9 84 8 10.1097/01.mco.0000214564.59815.af 16477170 32. Sjögren P Sierra-Johnson J Gertow K Rosell M Vessby B de Faire U Hamsten A Hellenius M-L Fisher RM Fatty acid desaturases in human adipose tissue: relationships between gene expression, desaturation indexes and insulin resistance Diabetologia 2008 51 328 35 10.1007/s00125-007-0876-9 18030445 33. Mayneris-Perxachs J Guerendiain M Castellote AI Estruch R Covas MI Fitó M Salas-Salvadó J Martínez-González MA Aros F Lamuela-Raventós RM López-Sabater MC Plasma fatty acid composition, estimated desaturase activities, and their relation with the metabolic syndrome in a population at high risk of cardiovascular disease Clin Nutr 2014 33 90 7 10.1016/j.clnu.2013.03.001 23591154 34. Venäläinen T Ågren J Schwab U De Mello VD Eloranta A-M Laaksonen DE Lindi V Lakka TA Cross-sectional associations of plasma fatty acid composition and estimated desaturase and elongase activities with cardiometabolic risk in Finnish children—The PANIC study J Clinical Lipidol 2016 10 82 91 10.1016/j.jacl.2015.09.004 26892124 35. Knutson KL Spiegel K Penev P Cauter EV The metabolic consequences of sleep deprivation Sleep Med Rev 2007 11 163 78 10.1016/j.smrv.2007.01.002 17442599 36. Weiss A Xu F Storfer-Isser A Thomas A Ievers-Landis CE Redline S The association of sleep duration with adolescent’s fat and carbohydrate consumption Sleep 2010 33 1201 9 20857867 37. He F Bixler EO Liao J Berg A Imamura KY Fernandez-Mendoza J Vgontzas AN Liao D Habitual sleep variability, mediated by nutrition intake, is associated with abdominal obesity in adolescents Sleep Med 2015 16 1489 94 10.1016/j.sleep.2015.07.028 26611945 38. Thorp AA Schlaich MP Relevance of sympathetic nervous system activation in obesity and metabolic syndrome J Diabetes Res 2015 2015 341583 10.1155/2015/341583 26064978 39. Holman RT Smythe L Johnson S Effect of sex and age on fatty acid composition of human serum lipids Am J Clin Nutr 1979 32 2390 9 506962 40. Warensjö E Sundström J Lind L Vessby B Factor analysis of fatty acids in serum lipids as a measure of dietary fat quality in relation to the metabolic syndrome in men Am J Clin Nutr 2006 84 442 8 16895896 41. Saadatian-Elahi M Slimani N Chajès V Jenab M Goudable J Biessy C Ferrari P Byrnes G Autier P Peeters PH Ocké M Bueno de Mesquita B Johansson I Hallmans G Manjer J Wirfält E González CA Navarro C Martinez C Amiano P Suárez LR Ardanaz E Tjønneland A Halkjaer J Overvad K Jakobsen MU Berrino F Pala V Palli D Tumino R Vineis P Santucci de Magistris M Spencer EA Crowe FL Bingham S Khaw KT Linseisen J Rohrmann S Boeing H Noethlings U Olsen KS Skeie G Lund E Trichopoulou A Oustoglou E Clavel-Chapelon F Riboli E Plasma phospholipid fatty acid profiles and their association with food intakes: results from a cross-sectional study within the European Prospective Investigation into Cancer and Nutrition Am J Clin Nutr 2009 89 331 46 10.3945/ajcn.2008.26834 19056549 42. Spaeth AM Dinges DF Goel N Effects of experimental sleep restriction on weight gain, caloric intake, and meal timing in healthy adults Sleep 2013 36 981 90 23814334

PMC005xxxxxx/PMC5000435.txt

==== Front J Orthop Surg ResJ Orthop Surg ResJournal of Orthopaedic Surgery and Research1749-799XBioMed Central London 42810.1186/s13018-016-0428-4Research ArticleEradication rates, risk factors, and implant selection in two-stage revision knee arthroplasty: a mid-term follow-up study Hoell Steffen +49-(0)541 966-3030st.hoell@web.de 1Sieweke Anna annasieweke@web.de 2Gosheger Georg georg.gosheger@ukmuenster.de 2Hardes Jendrik Jendrik.Hardes@ukmuenster.de 2Dieckmann Ralf ralf.dieckmann@ukimuenster.de 2Ahrens Helmut helmut.ahrens@ukmuenster.de 2Streitbuerger Arne Arne.Streitbuerger@ukmuenster.de 21 Department of Orthopaedics, Paracelsus Hospital, Am Natruper Holz 69, 49076 Osnabrück, Germany 2 Department of General Orthopaedics and Tumor Orthopaedics, University Hospital, Albert-Schweitzer-Campus 1, 48149 Münster, Germany 26 8 2016 26 8 2016 2016 11 1 9325 6 2016 18 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Two-stage revision (TSR) knee arthroplasty is an established treatment, but failure to control infection still occurs in 4–50 % of cases. The aim of this study was to assess the infection eradication rate, risk factors for failure, and the clinical outcome after two-stage revision knee arthroplasty. Methods This retrospective study included 59 patients who had undergone at least one two-stage revision procedure due to periprosthetic joint infection (PJI). Demographic data, comorbidities, types of implant, and complications were analyzed. Univariate and multivariate logistic regression analysis were used to identify risk factors for failure. Results The infections were controlled in 55 patients (93.2 %). The follow-up period was 4.1 (±2.7) years. Infection control was achieved after the first TSR in 42 patients (71.2 %) and after the second TSR in 13 (76.5 %). The percentage of arthrodesis procedures in patients with infection control increased from 16.75 % after one TSR to 69.2 % after two TSRs. Multivariate logistic regression analysis identified body mass index (BMI) (odds ratio 1.22; 95 % confidence intervals, 1.07 to 1.40; p = 0.004) and smoking (OR 21.52; 95 % CI, 2.60 to 178.19; p = 0.004) as risk factors for failure. Conclusions Two-stage revision protocols can achieve acceptable results even after a second procedure. It is still unclear whether the choice of implant influences failure rates. Risk factors for failure after two-stage revision were identified. Studies with larger sample sizes are needed in order to support these findings and identify further risk factors. To reduce failure rates, programs should be established to treat or minimize risk factors in patients with PJI. Keywords Two-stage revision knee arthroplastyArthrodesisRisk factorPeriprosthetic joint infectionBMINicotine abuseissue-copyright-statement© The Author(s) 2016 ==== Body Background Several studies have identified comorbidities and conditions that increase the rate of periprosthetic joint infection (PJI) after primary hip and knee arthroplasty [1–8]. In two-stage revision (TSR) surgery, protocols involving the implantation of an antibiotic-loaded bone–cement spacer have become the gold standard for treating periprosthetic infections. Radical debridement with explantation of the prosthesis and supportive administration of antibiotics are the most important pillars for controlling PJI [9, 10], but reinfection rates after TSR continue to be high. Reinfection rates reported in the literature range from 4 to 50 % [3, 4, 11–16]. Only a few studies have analyzed the factors that have a negative impact on infection control after TSR [17–20]. In order to minimize failure rates in TSR, evaluated treatment protocols and diagnostic algorithms are needed, and it should be possible to identify patients who are at higher risk of failure. Once risk factors have been identified, further investigations and additional treatments can help reduce failure rates. The aim of the present study was to investigate the extent to which infection can be successfully controlled after two-stage revision knee arthroplasty and identify factors that influence the failure rate. Methods Seventy patients who underwent two-stage revision knee arthroplasty between 2004 and 2008 in our department were identified. The following criteria were used to define PJI: sinus tract communicating with the prosthesis and/or at least two identical positive cultures identified intraoperatively [21, 22]. All infections were defined as delayed or late chronic [23]. Seven patients had died and four patients declined to participate in the study, and a total of 59 patients were therefore included. Their average age at follow-up was 73 years (±9.7), and there were 32 men and 27 women. The patients were all referred to our institution as a tertiary center. The protocol consisted of explantation of the prosthesis with implantation of a fixed antibiotic-loaded bone–cement spacer (Refobacin® Revision bone cement; Biomet Inc., Warsaw, Indiana, USA; 1 g gentamicin and 1 g clindamycin/40 g cement) and at least 14 days of intravenous antibiotic administration, followed by at least 4 weeks of antibiotics orally. If necessary, additional antibiotics were mixed into the spacer, depending on the microbiological results, as an off-label application. All antibiotic treatments were administered in collaboration with the hospital’s Institute of Microbiology. After an interval of 14 days without antibiotics, C-reactive protein (CRP) was measured in serum. The second-stage procedure was performed 9–12 weeks after explantation. The criteria for reimplantation were no sinus track, no signs of local inflammation, and serum CRP values that had declined during the period since explantation. The definition of infection control was no subsequent surgical intervention for infection at the time of follow-up. Eleven potential risk factors were documented from the demographic data, comorbidities, and postoperative complications. The criteria for arthrodesis (n = 18 patients) were an insufficient extension mechanism and/or clearly compromised capsule and soft-tissue conditions, with a high risk of postoperative wound healing problems and limited function. The indication for arthrodesis was based on the personal judgment and experience of the surgeon and the patient’s consent. None of the patients underwent additional soft-tissue coverage with local muscle flaps. Statistical analysis Means plus or minus standard deviation (SD), ranges, and proportions were calculated to analyze the different characteristics in the cases of two-stage knee revision. Statistical significance was assessed using the chi-squared test, Fisher’s exact test, Student’s t test, and the Mann-Whitney U test. The probability of failed infection control was modeled using univariate binary logistic regression. Odds ratios, the corresponding 95 % confidence intervals, and Wald-type p values were calculated. In a second step, variables were selected in a stepwise fashion, applying backward selection to variables in the univariate logistic regression. All inferential statistics are intended to be exploratory, not confirmatory, and are interpreted accordingly. The comparison-wise type 1 error rate is controlled instead of the experiment-wise error rate. The local significance level was set to 0.05. No adjustment for multiple testing was performed. Statistical analyses were performed using IBM SPSS® Statistics for Windows, version 21 (IBM Corporation, Armonk, NY, USA). Results Infection control was achieved in 55 patients (93.2 %). The follow-up period was 4.1 years (±2.7 years). Infection control was achieved after the first TSR in 42 patients (71.2 %) and after the second TSR in 13 patients (76.5 %). There were no significant differences between the first and second TSRs (p > 0.05). The percentage of arthrodesis in patients with infection control increased from 16.75 % after one TSR to 69.2 % after two TSRs. The average time from reimplantation to reinfection was 2.3 years (range 0.6–3.7 years). The amputation rate when infection could not be controlled was 6.8 % (4/59); amputations were required in one patient with an arthrodesis and three with revision endoprostheses. Figure 1 shows the clinical course for all of the patients. The risk factors investigated and the results of the univariate logistic regression are listed in Table 1. Although patients who had Staphylococcus epidermidis at the first revision had the highest failure rate (35.3 %), statistical analysis was not performed due to the small number of cases. Table 2 presents the results of the multivariate logistic regression analysis after variable selection. Table 3 shows the organisms that were cultured in patients with recurrent infections and the choice of implant. Identical bacteria were found at the second TSR in eight of the 17 patients concerned (47.1 %).Fig. 1 Flowchart of all patients Table 1 Potential risk factors for faiure that were investigated with univariable logistic regression Infection controlled after the first TSR Fialure after the first TSR P value Odds ratio CI (95 %) Sinus present 0.008 5.24 1.55–17.65  Yes n = 19 N = 10 N = 9  No n = 40 N = 35 N = 5 Diabetes 0.009 6.65 1.62–27.3  Yes n = 11 n = 4 n = 7  No n = 48 n = 38 n = 10 Smoking 0.018 8.33 1.43–48.54  Yes n = 7 N = 3 N = 4  No n = 52 N = 42 N = 10 BMI >30 0.033 5.74 1.15–28.62  Yes n = 37 n = 24 n = 13  No n = 22 n = 17 n = 5 Periprosthetic fracture 0.034 3.57 1.1–11.57  Yes n = 23 n = 14 N = 9  No n = 36 n = 28 N = 8 Wound healing problems 0.061 3.16 0.95–10.55  Yes n = 17 N = 10 N = 7  No n = 42 N = 35 N = 7 Corticosteriods 0.076 8.38 0.8–87.11  Yes n = 4 N = 2 N = 2  No n = 55 N = 43 N = 12 Immune suppression 0.191 5.2 0.44–61.67  Yes N = 1 N = 2  No N = 45 N = 11 Postoperative hematoma 0.418 1.67 0.48–5.8  Yes n = 16 N = 11 N = 5  No n = 43 N = 29 N = 14 Blood transfusion 0.458 2.37 0.24–23.1  Yes n = 44 N = 28 N = 16  No n = 15 N = 14 N = 1 Tumor disease 0.986 1.02 0.18–5.91  Yes n = 7 N = 5 N = 2  No n = 52 N = 40 N = 12 P value, significance level was set to 0.05 Table 2 Comorbid conditions or patterns that were identified by variable selection as risk factors in a multivariable logistic regression P Odds ratio CI (95 %) Body mass index (kg/m2) 0.004 1.22 1.07–1.40 Nicotine abuse 0.004 21.52 2.60–178.19 Table 3 Patients with recurrent infection Patients Culture during the first TSR Culture during the second TSR Implant after the first TSR Outcome after the second TSR 1 Staph aureus Staph aureus Revision prosthesis Revision prosthesis 2 Staph aureus Staph aureus Revision prosthesis Revision prosthesis 3 Staph aureus Staph aureus Revision prosthesis Revision prosthesis 4 Staph epi Staph epi Revision prosthesis Arthrodesis 5 Staph epi Staph epi Revision prosthesis Arthrodesis 6 Staph epi Staph epi Revision prosthesis arthrodesis 7 Staph epi Staph epi Revision prosthesis Revision prosthesis 8 Staph capitis Staph capitis Revision prosthesis Arthrodesis 9 Staph epi Staph haemolyticus Revision prosthesis Arthrodesis 10 Staph epi MRSA Revision prosthesis Arthrodesis 11 Staph epi MRSA Revision prosthesis Amputation 12 Staph epi MRSA Revision prosthesis amputation 13 Enterobacter faecalis E. coli Revision prosthesis Arthrodesis 14 Staph aureus Staph epi Revision prosthesis amputation 15 Staph simulans Staph epi Revision prosthesis Arthrodesis 16 Staphylococcus hominis Streptococcus acidominimus Candida albicans arthrodesis amputation 17 Staphylococcus epidermidis/Klebsiella oxytoca/Pseudomonas aeruginosa/Enterococcus faecalis Candida albicans Revision prosthesis Arthrodesis Discussion Periprosthetic joint infection (PJI) is one of the most severe complications that occur in patients who undergo total knee arthroplasty (TKA). Two-stage revision is still the gold standard for treatment of PJI, although one-stage revisions may achieve similar results in special conditions [24–28]. Nevertheless, reinfection rates vary from 4 to 50 % [3, 4, 11–16]. Among the patients included in the present study, successful treatment was achieved in 55 (93.2 %) after a mean follow-up period of 4.1 years. There were no differences in the success rates between patients who underwent one TSR procedure and those with two procedures. Lower eradication rates have been reported in the literature after a second TSR [29], but a high rate of arthrodesis in the second TSRs might be an explanation for this. Isiklar et al. recommended arthrodesis instead of multiple revisions in patients with chronic infections, in order to avoid amputation [15]. Other studies have also reported higher rates of infection control with arthrodesis in comparison with revision prostheses [2, 11, 16, 30]. In contrast to these results, a 50 % failure rate after septic arthrodesis was reported in 2015 [31]. In view of the small numbers of arthrodeses, statistical analysis was not carried out in the present study and no conclusions can therefore be drawn on whether or not arthrodesis is in fact associated with lower reinfection rates. It has to be discussed if allograft reconstruction of the extensor mechanism is an alternative instead of arthrodesis. Although it is known that allograft reconstructions show high rates of complications the benefit of a better mobility must be considered. In a study from 2016 in 26 knees, 69 % of the allografts could be retained at a follow-up of 68 months with a reoperation rate of 58 % [32, 33]. However, it is not only the type of treatment administered that is responsible for the clinical outcome. It is known from several studies that comorbidities and other conditions can have a negative influence on infection rates after primary arthroplasty [1–8]. The causes of failure after TSR are rarely reported [17–20]. The most frequent potential risk factors for failure were analyzed in the patients included in the present study. Among the comorbid conditions present, diabetes was identified as a risk factor, with an OR of 6.65 (95 % CI, 1.62 to 27.30) in the univariate analysis. Another study published in 2015 also found that diabetes had a significantly higher prevalence in the group with reinfections [19]. By contrast, Sakellariou et al. did not find any significant differences in a univariate analysis of 110 patients with TSR [18]. Among the local conditions that were present, fistulas were found to be a relevant factor in the univariate analysis. This finding is supported by a study also published in 2015, in which fistulas were associated with recurrent infection even in the multivariate logistic regression analysis [20]. A medical history including periprosthetic fracture around the knee was identified as a risk factor for failure after the first TSR. In an earlier study, our group showed that septic failure of revision arthroplasty with megaprostheses was strongly associated with a medical history of periprosthetic fractures around the knee [13]. Suzuki et al. investigated the influence of surgical procedures in the region of the knee joint. They observed significantly more frequent infections with open reduction and internal fixation after trauma to the knee joint and when osteosynthesis material remained in situ [7]. Two risk factors were identified in the multivariate logistic regression analysis in the present study: body mass index (BMI) and smoking. An increase in the BMI by one point showed an increased risk for failure of about 22 %. However, Mortazavi et al. did not observe any association between BMI and persistent PJI after two-stage TKA [25]. Kubista et al. distinguished between BMI scores of <25, 25–35, and >35. No significant differences were observed between these groups with regard to the rates of persistent infection after two-stage TKA [34]. In a matched-cohort study, patients with a BMI >40 kg/m2 had a 22 % risk for reinfection in comparison with patients with a BMI <30 kg/m2, at 4 % [35]. In two-stage revision hip arthroplasty, obesity has also been found to be a significant risk factor for failure [36]. Higher rates of recurrent infection have also been reported among smokers, with a 71.4 % rate of persistent or recurrent infection after the first two-stage replacement in comparison with only 23.1 % in nonsmokers [6, 7, 37]. These results were confirmed in the present cohort. The study has several limitations. As all of the patients were referred to the department, it was not possible to record all relevant factors. For example, the number of previous revision procedures was unclear and could not be analyzed, although it is known that this factor has a negative influence on complication rates [13, 17]. Due to the relatively small number of patients who underwent arthrodesis, statistical analysis was not useful. The wide variety of bacteria identified also made it impossible to carry out statistical analysis. Conclusions Two-stage revision (TSR) protocols can achieve acceptable results even when they are repeated. Amputation rates can be kept low. It is still unclear whether the choice of implant has an influence on failure rates. Risk factors for failure after two-stage revision have been identified, but studies with larger numbers of patients are needed in order to support these findings and identify further risk factors. In order to reduce failure rates, programs should be established for treating or minimizing risk factors in patients with periprosthetic joint infection. Abbreviations BMIBody mass index CIConfidence interval CRPC-reactive protein OROdds ratio PJIPeriprosthetic joint infection SDStandard deviation TKATotal knee arthroplasty TSRTwo-stage revision Acknowledgements None. Funding The study was not funded. Availability of data and material Not applicable. Authors’ contributions SH made substantial contribution to conception and design; AS made substantial contribution to acquisition of the data and critical revision of the manuscript; GG made critical revision of the manuscript for content; JH made contribution to analysis and interpretation of the data; RD made substantial contribution to acquisition of the data; HA made contribution to analysis and interpretation of the data. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Consent for publication was obtained from the participants. Ethics approval and consent to participate Ethics approval was not necessary, as confirmed by the local ethics committee at the University of Münster, Germany (ref. 2016-309-f-N). ==== Refs References 1. Berbari EF Hanssen AD Duffy MC Steckelberg JM Ilstrup DM Harmsen WS Osmon DR Risk factors for prosthetic joint infection: case-control study Clin Infect Dis 1998 27 1247 54 10.1086/514991 9827278 2. Dowsey MM Choong PFM Obese diabetic patients are at substantial risk for deep infection after primary TKA Clin Orthop Relat Res 2009 467 1577 81 10.1007/s11999-008-0551-6 18841430 3. Jämsen E Huhtala H Puolakka T Moilanen T Risk factors for infection after knee arthroplasty. A register-based analysis of 43,149 cases J Bone Joint Surg Am 2009 91 38 47 10.2106/JBJS.G.01686 19122077 4. Jämsen E Varonen M Huhtala H Lehto MUK Lumio J Konttinen YT Moilanen T Incidence of prosthetic joint infections after primary knee arthroplasty J Arthroplasty 2010 25 87 92 10.1016/j.arth.2008.10.013 19056210 5. Lai K Bohm ER Burnell C Hedden DR Presence of medical comorbidities in patients with infected primary hip or knee arthroplasties J Arthroplasty 2007 22 651 6 10.1016/j.arth.2006.09.002 17689771 6. Peersman G Laskin R Davis J Peterson M Infection in total knee replacement: a retrospective review of 6489 total knee replacements Clin Orthop Relat Res 2001 392 15 23 10.1097/00003086-200111000-00003 11716377 7. Suzuki G Saito S Ishii T Motojima S Tokuhashi Y Ryu J Previous fracture surgery is a major risk factor of infection after total knee arthroplasty Knee Surg Sports Traumatol Arthrosc 2011 19 2040 4 10.1007/s00167-011-1525-x 21541707 8. Winiarsky R Barth P Lotke P Total knee arthroplasty in morbidly obese patients J Bone Joint Surg Am 1998 80 1770 4 9875934 9. Zimmerli W Trampuz A Ochsner PE Prosthetic-joint infections N Engl J Med 2004 351 1645 54 10.1056/NEJMra040181 15483283 10. Koyonos L Zmistowski B Della Valle CJ Parvizi J Infection control rate of irrigation and débridement for periprosthetic joint infection Clin Orthop Relat Res 2011 469 3043 8 10.1007/s11999-011-1910-2 21553171 11. Bengtson S Knutson K The infected knee arthroplasty. A 6-year follow-up of 357 cases Acta Orthop Scand 1991 62 301 11 10.3109/17453679108994458 1882666 12. Gill GS Joshi AB Long-term results of cemented, posterior cruciate ligament retaining total knee arthroplasty in osteoarthritis Am J Knee Surg 2001 14 209 14 11703032 13. Höll S Schlomberg A Gosheger G Dieckmann R Streitbuerger A Schulz D Hardes J Distal femur and proximal tibia replacement with megaprosthesis in revision knee arthroplasty: a limb-saving procedure Knee Surg Sports Traumatol Arthrosc 2012 20 2513 8 10.1007/s00167-012-1945-2 22392068 14. Iorio R Williams KM Marcantonio AJ Specht LM Tilzey JF Healy WL Diabetes mellitus, hemoglobin A1C, and the incidence of total joint arthroplasty infection J Arthroplasty 2012 27 726 9.e1 10.1016/j.arth.2011.09.013 22054905 15. Isiklar ZU Landon GC Tullos HS Amputation after failed total knee arthroplasty Clin Orthop Relat Res 1994 299 173 8 8119014 16. Kramhøft M Bødtker S Carlsen A Outcome of infected total knee arthroplasty J Arthroplasty 1994 9 617 21 10.1016/0883-5403(94)90115-5 7699374 17. Pelt CE Grijalva R Anderson L Anderson MB Erickson J Peters CL Two-stage revision TKA is associated with high complication and failure rates Adv Orthop 2014 2014 659047 10.1155/2014/659047 25610660 18. Sakellariou VI Poultsides LA Vasilakakos T Sculco P Ma Y Sculco TP Risk factors for recurrence of periprosthetic knee infection J Arthroplasty 2015 30 1618 22 10.1016/j.arth.2015.04.005 25891435 19. Cha MS Cho SH Kim DH Yoon HK Cho HS Lee DY Two-stage total knee arthroplasty for prosthetic joint infection Knee Surg Relat Res 2015 27 82 9 10.5792/ksrr.2015.27.2.82 26060606 20. Massin P, Delory T, Lhotellier L, Pasquier G, Roche O, Cazenave A, Estellat C, Jenny JY. Infection recurrence factors in one- and two-stage total knee prosthesis exchanges. Knee Surg Sports Traumatol Arthrosc. 2015. [Epub ahead of print] 21. Parvizi J Gehrke T Chen AF Proceedings of the international consensus on periprosthetic joint infection Bone Jt J 2013 95-B 1450 2 10.1302/0301-620X.95B11.33135 22. Virolainen P Lähteenmäki H Hiltunen A Sipola E Meurman O Nelimarkka O The reliability of diagnosis of infection during revision arthroplasties Scand J Surg 2002 91 178 81 12164519 23. Zimmerli W Moser C Pathogenesis and treatment concepts of orthopaedic biofilm infections FEMS Immunol Med Microbiol 2012 65 158 68 10.1111/j.1574-695X.2012.00938.x 22309166 24. Mittal Y Fehring TK Hanssen A Marculescu C Odum SM Osmon D Two-stage reimplantation for periprosthetic knee infection involving resistant organisms J Bone Joint Surg Am 2007 89 1227 31 10.2106/JBJS.E.01192 17545425 25. Mortazavi SMJ Vegari D Ho A Zmistowski B Parvizi J Two-stage exchange arthroplasty for infected total knee arthroplasty: predictors of failure Clin Orthop Relat Res 2011 469 3049 54 10.1007/s11999-011-2030-8 21866421 26. Romanò CL Gala L Logoluso N Romanò D Drago L Two-stage revision of septic knee prosthesis with articulating knee spacers yields better infection eradication rate than one-stage or two-stage revision with static spacers Knee Surg Sports Traumatol Arthrosc 2012 20 2445 53 10.1007/s00167-012-1885-x 22270671 27. Sherrell JC Fehring TK Odum S Hansen E Zmistowski B Dennos A Kalore N Consortium PI The Chitranjan Ranawat Award: fate of two-stage reimplantation after failed irrigation and débridement for periprosthetic knee infection Clin Orthop Relat Res 2011 469 18 25 10.1007/s11999-010-1434-1 20582495 28. Wroblewski BM One-stage revision of infected cemented total hip arthroplasty Clin Orthop Relat Res 1986 211 103 7 3769248 29. Stammers J Kahane S Ranawat V Miles J Pollock R Carrington RWJ Briggs T Skinner JA Outcomes of infected revision knee arthroplasty managed by two-stage revision in a tertiary referral centre Knee 2015 22 56 62 10.1016/j.knee.2014.10.005 25467934 30. Wu CH Gray CF Lee GC Arthrodesis should be strongly considered after failed twostage reimplantation TKA Clin Orthop Relat Res 2014 472 3295 304 10.1007/s11999-014-3482-4 24488752 31. Röhner E Windisch C Nuetzmann K Rau M Arnhold M Matziolis G Unsatisfactory outcome of arthrodesis performed after septic failure of revision total knee arthroplasty J Bone Joint Surg Am 2015 97 298 301 10.2106/JBJS.N.00834 25695981 32. Bauman RD Lewallen DG Hanssen AD Limitations of structural allograft in revision total knee arthroplasty Clin Orthop Relat Res 2009 467 3 818 24 10.1007/s11999-008-0679-4 19130161 33. Ricciardi BF, Oi K, Trivellas M, Lee YY, Della Valle AG, Westrich GH, Survivorship of extensor mechanism allograft reconstruction after total knee arthroplasty. J Arthroplasty. 2016. [Epub ahead of print] 34. Kubista B Hartzler RU Wood CM Osmon DR Hanssen AD Lewallen DG Reinfection after two-stage revision for periprosthetic infection of total knee arthroplasty Int Orthop 2012 36 65 71 10.1007/s00264-011-1267-x 21553042 35. Watts CD Wagner ER Houdek MT Osmon DR Hanssen AD Lewallen DG Mabry TM Morbid obesity: a significant risk factor for failure of two-stage revision total knee arthroplasty for infection J Bone Joint Surg Am 2014 96 10.2106/JBJS.M.01289 25232084 36. Houdek MT Wagner ER Watts CD Osmon DR Hanssen AD Lewallen DG Mabry TM Morbid obesity: a significant risk factor for failure of two-stage revision total hip arthroplasty for infection J Bone Joint Surg Am 2015 97 326 32 10.2106/JBJS.N.00515 25695985 37. Singh JA Houston TK Ponce BA Maddox G Bishop MJ Richman J Campagna EJ Henderson WG Hawn MT Smoking as a risk factor for short-term outcomes following primary total hip and total knee replacement in veterans Arthritis Care Res 2011 63 1365 74 10.1002/acr.20555

PMC005xxxxxx/PMC5000436.txt

==== Front Parasit VectorsParasit VectorsParasites & Vectors1756-3305BioMed Central London 174710.1186/s13071-016-1747-3ResearchThe intracellular bacterium Anaplasma phagocytophilum selectively manipulates the levels of vertebrate host proteins in the tick vector Ixodes scapularis Villar Margarita margaritam.villar@uclm.es 1López Vladimir vladimirlopez6@gmail.com 1Ayllón Nieves nieves.ayllon@uclm.es 1Cabezas-Cruz Alejandro cabezasalejandrocruz@gmail.com 2López Juan A. juan.lopez@cnic.es 3Vázquez Jesús jesus.vazquez@cnic.es 3Alberdi Pilar maria.alberdi@uclm.es 1de la Fuente José jose_delafuente@yahoo.com 141 SaBio. Instituto de Investigación en Recursos Cinegéticos IREC-CSIC-UCLM-JCCM, Ronda de Toledo s/n, 13005, Ciudad Real, Spain 2 University Lille, CNRS, Inserm, CHU Lille, Institut Pasteur de Lille, U1019 - UMR 8204 - CIIL - Centre d’Infection et d’Immunité de Lille, F-59000 Lille, France 3 Centro Nacional de Investigaciones Cardiovasculares (CNIC), Madrid, Spain 4 Department of Veterinary Pathobiology, Center for Veterinary Health Sciences, Oklahoma State University, Stillwater, OK 74078 USA 25 8 2016 25 8 2016 2016 9 1 46711 5 2016 11 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background The intracellular bacteria Anaplasma phagocytophilum are emerging zoonotic pathogens affecting human and animal health, and a good model for the study of tick-host-pathogen interactions. This tick-borne pathogen is transmitted by Ixodes scapularis in the United States where it causes human granulocytic anaplasmosis. Tick midguts and salivary glands play a major role during tick feeding and development, and in pathogen acquisition, multiplication and transmission. Vertebrate host proteins are found in tick midguts after feeding and have been described in the salivary glands of fed and unfed ticks, suggesting a role for these proteins during tick feeding and development. Furthermore, recent results suggested the hypothesis that pathogen infection affects tick metabolic processes to modify host protein digestion and persistence in the tick with possible implications for tick physiology and pathogen life-cycle. Methods To address this hypothesis, herein we used I. scapularis female ticks fed on uninfected and A. phagocytophilum-infected sheep to characterize host protein content in midguts and salivary glands by proteomic analysis of tick tissues. Results The results evidenced a clear difference in the host protein content between tick midguts and salivary glands in response to infection suggesting that A. phagocytophilum selectively manipulates the levels of vertebrate host proteins in ticks in a tissue-specific manner to facilitate pathogen infection, multiplication and transmission while preserving tick feeding and development. The mechanisms by which A. phagocytophilum manipulates the levels of vertebrate host proteins are not known, but the results obtained here suggested that it might include the modification of proteolytic pathways. Conclusions The results of this study provided evidence to support that A. phagocytophilum affect tick proteolytic pathways to selectively manipulate the levels of vertebrate host proteins in a tissue-specific manner to increase tick vector capacity. Investigating the biological relevance of host proteins in tick biology and pathogen infection and the mechanisms used by A. phagocytophilum to manipulate host protein content is essential to advance our knowledge of tick-host-pathogen molecular interactions. These results have implications for the identification of new targets for the development of vaccines for the control of tick-borne diseases. Electronic supplementary material The online version of this article (doi:10.1186/s13071-016-1747-3) contains supplementary material, which is available to authorized users. Keywords AnaplasmaTickIxodesProteomicsHemoglobinImmunologyMinisterio de Economia y CompetitividadBFU2011-23896Seventh Framework Programme (EU)ANTIGONE project number 278976University of Castilla - La ManchaResearch PlanVillar Margarita issue-copyright-statement© The Author(s) 2016 ==== Body Background Infectious diseases transmitted by arthropod vectors such as ticks constitute a growing burden for human and animal health worldwide [1–4]. Anaplasma phagocytophilum (Rickettsiales: Anaplasmataceae) are tick-borne intracellular bacteria that infect vertebrate host granulocytes causing human, canine and equine granulocytic anaplasmosis and tick-borne fever of ruminants [5–8]. Human granulocytic anaplasmosis (HGA) is the second most common tick-borne disease in the United States, and tick-borne fever is an established and economically important disease of sheep in Europe [6, 9]. This emerging zoonotic pathogen is transmitted by Ixodes ticks of which the major vector species are I. scapularis in the United States and I. ricinus in Europe [2]. In ticks, A. phagocytophilum infects and multiply in different tissues including midguts [10] and salivary glands [11]. Anaplasma phagocytophilum is a good model for the study of tick-host-pathogen interactions because recent results have shown that infection affects gene expression and protein levels in vertebrate hosts and ticks [8, 12]. Recently, we proposed that the evolution of I. scapularis-host-A. phagocytophilum molecular interactions involving genetic traits of all parts resulted in conflict and cooperation between them, with mutual beneficial effects for ticks, hosts and pathogens [13]. Furthermore, A. phagocytophilum evolved common strategies for infection of vertebrate host and tick cells that include but are not limited to remodeling of the cytoskeleton, inhibition of cell apoptosis, manipulation of the immune response and control of host cell epigenetics [14]. Tick midguts are the tissue where blood digestion occurs while salivary glands produce and secrete proteins and other molecules that modulate host defenses to enhance blood feeding [15–21]. Additionally, tick midguts and salivary glands play a major role during pathogen acquisition, multiplication and transmission [22, 23]. The functional role of these tissues during tick feeding and pathogen infection is reflected at the transcriptome and proteome levels, with tissue-specific differences between midguts and salivary glands [12, 20, 21]. Vertebrate host proteins are found in tick midguts after feeding and have been described in the salivary glands and saliva of fed ticks [24–30]. Additionally, vertebrate host proteins and particularly alpha and beta-globin chains of hemoglobin have been identified in unfed I. scapularis and Amblyomma americanum nymphs [31] and Rhipicephalus sanguineus adult ticks [32]. Wickramasekara et al. [31] suggested that because blood meal digestion in ticks occurs gradually within midgut cells after endocytosis [19], the combination of slow assimilation and uptake of some host proteins into tick hemolymph might explain the persistence of host blood proteins for months after feeding and molting. In fact, Francischetti et al. [33] did not identify vertebrate host proteins in salivary glands from fed soft ticks Ornithodoros coriaceus, probably associated with the fact that hard ticks feed for several days while soft ticks feed for less than 1 hour, therefore decreasing the probability for host proteins to persist. Nevertheless, Diaz-Martin et al. [26] did find host proteins in the saliva of the soft ticks Ornithodoros moubata. These results suggest a mechanism by which host proteins remain in the tick after blood digestion. For example, host proteins persisting in the tick after molting may serve as a reserve for nutrients until the next infestation and feeding cycle are completed. However, preliminary results in questing R. sanguineus infected with Rickettsia conorii suggested that pathogen infection modify tick digestion processes, thus provoking an increase in the concentration of some host proteins such as hemoglobin ingested with blood meal in infected ticks when compared to uninfected controls [32]. These results led to the hypothesis that pathogen infection affect tick metabolic processes to modify host protein digestion and persistence in the tick, with possible implications for tick physiology and pathogen life-cycle. To address this hypothesis, in this study we used I. scapularis female ticks fed on uninfected and A. phagocytophilum-infected sheep to characterize host proteins present in midguts and salivary glands by mass spectrometry (MS) analysis of the proteome. The results evidenced a clear difference in the sheep host protein content between tick midguts and salivary glands in response to infection and provided evidence to support that A. phagocytophilum selectively manipulates the levels of vertebrate host proteins in the tick vector I. scapularis. Methods Ethics statement Animals were housed and experiments conducted with the approval and supervision of the Oklahoma State University Institutional Animal Care and Use Committee (Animal Care and Use Protocol, ACUP No. VM1026). Ticks and sample preparation Ticks and sample preparation were previously described [12]. Briefly, I. scapularis ticks were obtained from the laboratory colony maintained at the Oklahoma State University Tick Rearing Facility [34]. Adult female I. scapularis were infected with A. phagocytophilum by feeding on a sheep inoculated intravenously with approximately 1 × 107A. phagocytophilum (NY18 isolate)-infected HL-60 cells (90–100 % infected cells). In this model, over 85 % of the ticks become infected with A. phagocytophilum in nymphs, midguts and salivary glands. One hundred adult female ticks were removed from the sheep 7 days after infestation, held in the humidity chamber for 4 days and dissected for DNA, RNA and protein extraction from midguts and salivary glands using the AllPrep DNA/RNA/Protein Mini Kit (Qiagen, Valencia, CA, USA). Midguts and salivary glands were washed in PBS after collection to remove hemolymph-related cells. Uninfected ticks were prepared in a similar way but feeding on an uninfected sheep. Two independent samples were collected and processed for proteomics analysis for each tick tissue. Ten individual female ticks were dissected and samples collected and processed as described above to characterize A. phagocytophilum infection and the mRNA or protein levels of selected genes/proteins after RNA sequencing (RNAseq) or proteomics analyses. Proteomics data collection and analysis Proteins were digested using the filter aided sample preparation (FASP) protocol [35]. The FASP method allows processing total SDS lysates of essentially any class of protein from biological samples of any origin, thus solving the long-standing problem of efficient and unbiased solubilization of all cellular proteins irrespective of their subcellular localization and molecular weight. Samples were dissolved in 50 mM Tris-HCl pH8.5, 4 % SDS and 50 mM DTT, boiled for 10 min and centrifuged. Protein concentration in the supernatant was measured by the Direct Detect system (Millipore, Billerica, MA, USA). About 150 μg of protein were diluted in 8 M urea in 0.1 M Tris-HCl (pH 8.5) (UA), and loaded onto 30 kDa centrifugal filter devices (FASP Protein Digestion Kit, Expedeon, TN, USA). With this method, the sample is solubilized in 4 % SDS, then retained and concentrated into microliter volumes in an ultrafiltration device. The filter unit then acts as a ‘proteomic reactor’ for detergent removal, buffer exchange, chemical modification and protein digestion. Notably, during peptide elution, the filter retains high-molecular-weight substances that would otherwise interfere with subsequent peptide separation [35]. The denaturation buffer was replaced by washing three times with UA. Proteins were later alkylated using 50 mM iodoacetamide in UA for 20 min in the dark, and the excess of alkylation reagents were eliminated by washing three times with UA and three additional times with 50 mM ammonium bicarbonate. Proteins were digested overnight at 37 °C with modified trypsin (Promega, Madison, WI, USA) in 50 mM ammonium bicarbonate at 40:1 protein:trypsin (w/w) ratio. The resulting peptides were eluted by centrifugation with 50 mM ammonium bicarbonate (twice) and 0.5 M sodium chloride. Trifluoroacetic acid (TFA) was added to a final concentration of 1 % and the peptides were finally desalted onto C18 Oasis-HLB cartridges and dried-down for further analysis. For stable isobaric labeling, the resulting tryptic peptides were dissolved in Triethylammonium bicarbonate (TEAB) buffer and labeled using the 4-plex iTRAQ Reagents Multiplex Kit (Applied Biosystems, Foster City, CA, USA) according to manufacturer’s protocol. Briefly, each peptide solution was independently labeled at room temperature for 1 h with one iTRAQ reagent vial (mass tag 114, 115, 116 or 117) previously reconstituted with 70 μl of ethanol. Reaction was stopped after incubation at room temperature for 15 min with diluted TFA, and peptides were combined. Samples were evaporated in a Speed Vac, desalted onto C18 Oasis-HLB cartridges and dried-down for further analysis. Labeled peptides were loaded into the liquid chromatography (LC)-MS/MS system for on-line desalting onto C18 cartridges and analyzing by LC-MS/MS using a C-18 reversed phase nano-column (75 μm I.D. × 50 cm, 2 μm particle size, Acclaim PepMap 100 C18; Thermo Fisher Scientific, Waltham, MA, USA) in a continuous acetonitrile gradient consisting of 0–30 % B in 145 min, 30–43 % B in 5 min and 43–90 % B in 1 min (A = 0.5 % formic acid; B = 90 % acetonitrile, 0.5 % formic acid). A flow rate of c.200 nl/min was used to elute peptides from the reverse phase nano-column to an emitter nanospray needle for real time ionization and peptide fragmentation on a QExactive mass spectrometer (Thermo Fisher Scientific). For increasing proteome coverage, iTRAQ-labeled samples were run at least twice. For peptide identification, all spectra were analyzed with Proteome Discoverer (version 1.4.0.29, Thermo Fisher Scientific) using a Uniprot databases containing all sequences from Ruminantia and Ixodida (April 14, 2014). For database searching, parameters were selected as follows: trypsin digestion with 2 maximum missed cleavage sites, precursor and fragment mass tolerances of 2 Da and 0.02 Da, carbamidomethyl cysteine as fixed modification and methionine oxidation as dynamic modifications. For iTRAQ labeled peptides, N-terminal and Lys iTRAQ modification was added as a fixed modification. Peptide identification was validated using the probability ratio method [36] and false discovery rate (FDR) was calculated using inverted databases and the refined method [37] with an additional filtering for precursor mass tolerance of 12 ppm. Only peptides with a confidence of at least 99 % were used to quantify the relative abundance of each peptide determined as described previously [38]. Protein quantification from reporter ion intensities and statistical analysis of quantitative data were performed as described previously using QuiXoT [39, 40]. The intensity of the reporter peaks was used to calculate the fitting weight of each spectrum in the statistical model as described previously [39]. Outliers at the scan and peptide levels and significant protein-abundance changes were detected from the Z-values (the standardized variable used by the model that expresses the quantitative values in units of standard deviation) by using a FDR threshold of 1 % as described previously [39]. Results are the mean of two replicates. The gene ontology (GO) analysis was performed using Uniprot (http://www.uniprot.org) annotations. Characterization of the digestion of sheep host hemoglobins in tick midguts and salivary glands Sheep hemoglobin alpha 1/2 (P68240) and beta (P02075) peptides detected by MS analysis with 1 % FDR in midguts and salivary glands from uninfected and A. phagocytophilum-infected ticks were used for analysis. For this analysis, a new database search of MS spectra was performed with the same parameters described above but selecting “no enzyme” instead of “trypsin” digestion to identify the non-tryptic peptides present in the samples. The preferred cleavage sites for hemoglobinolytic enzymes Trypsin, Leucine aminopeptidase, Legumain, Cathepsin B, Cathepsin C, Serine carboxypeptidase were determined by searching against the MEROPS Peptidase Database (https://merops.sanger.ac.uk, release 10.0) (see in Additional file 1: Dataset S1). Characterization of the I. scapularis proteolytic and heme transport pathways mRNA and protein levels in response to A. phagocytophilum infection The quantitative transcriptomics and proteomics data for midguts and salivary glands from uninfected and A. phagocytophilum-infected I. scapularis were obtained from previously published results and deposited at the Dryad Digital Repository database with the dataset identifier http://dx.doi.org//10.5061/dryad.50kt0 [12]. The analysis of the tick proteolytic pathway included the genes/proteins annotated as protease, proteinase, peptidase, and its inhibitors [41]. Determination of hemoglobin protein levels by ELISA Proteins were extracted from midguts and salivary glands dissected from individual uninfected and A. phagocytophilum-infected I. scapularis female ticks using the AllPrep DNA/RNA/Protein Mini Kit (Qiagen, Inc. Valencia, CA, USA) according to manufacturer instructions. Extracted proteins were resuspended in PBS with 0.5 % Triton X-100 and protein concentration was determined with the Pierce BCA Protein Assay Kit (Thermo Scientific, San Jose, CA, USA) using bovine serum albumin (BSA) as standard. Hemoglobin protein levels were determined by ELISA (Cloud-Clone Corp., Houston, TX, USA) following manufacturer instructions. Optical density values were converted to μg/ml hemoglobin using the assay standard curve and regression analysis. Hemoglobin values were compared between groups by one-tailed Student’s t-test for samples with unequal variance (P < 0.05; n = 2 biological replicates). Determination of tick mRNA levels by real-time RT-PCR The expression of selected genes was characterized using total RNA extracted from individual I. scapularis female midguts and salivary glands obtained from uninfected and A. phagocytophilum-infected samples as previously described [12]. All ticks were confirmed as infected or uninfected by real-time PCR analysis of A. phagocytophilum msp4 DNA. Real-time RT-PCR was performed on RNA samples with gene-specific oligonucleotide primers (see Additional file 2: Table S1) using the iScript One-Step RT-PCR Kit with SYBR Green and the iQ5 thermal cycler (Bio-Rad, Hercules, CA, USA) following the manufacturer’s recommendations. A dissociation curve was run at the end of the reaction to ensure that only one amplicon was formed and that the amplicons denatured consistently in the same temperature range for every sample. The mRNA levels were normalized against tick cyclophilin and ribosomal protein S4 as described previously using the genNorm method (ddCT method as implemented by Bio-Rad iQ5 Standard Edition, Version 2.0) [12]. Normalized Ct values were compared between infected and uninfected tick samples by Student’s t-test with unequal variance (P < 0.05; n = 3–17 biological replicates). Immunofluorescence assay (IFA) Female ticks fed on A. phagocytophilum-infected and uninfected sheep and fixed with 4 % paraformaldehyde in 0.2 M sodium cacodylate buffer were embedded in paraffin and used to prepare sections on glass slides as previously described [12]. The paraffin was removed from the sections through two washes in xylene and the sections were hydrated by successive 5 min washes with a graded series of 100 %, 96 and 65 % ethanol and finally with distilled water. Next, the slides were treated with Proteinase K (Dako, Barcelona, Spain) for 7 min, washed with 0.1 % PBS-Tween 20 (Sigma-Aldrich, St. Louis, MI, USA) and blocked with 2 % bovine serum albumin (BSA; Sigma-Aldrich) in PBS-Tween 20 during 1 h at room temperature. The slides were then incubated overnight at 4 °C with rabbit anti-Cathepsin L (mature region No. pab0213-0; Covalab, Villeurbanne, France) antibodies diluted 1:1000 in 2 % BSA/PBS-Tween 20. This antibody was previously shown to recognize tick Cathepsin L by Western blot [42]. After 3 washes with PBS-Tween 20, the slides were incubated for 1 h with goat-anti-rabbit IgG conjugated with FITC (Sigma-Aldrich) diluted 1:160 in 2 % BSA/PBS-Tween 20. Finally, after two washes with PBS the slides were mounted on ProLong Diamond Antifade Mountant with DAPI reagent (Thermo Scientific™, Madrid, Spain). The sections were examined using a Leica SP2 laser scanning confocal microscope (Leica, Wetzlar, Germany) and IgGs from rabbit preimmune serum were used as controls. Proteomics data Data are available via Peptide Atlas (http://www.peptideatlas.org) with identifier PASS00854. Results The vertebrate host protein content differs between tick midguts and salivary glands in response to A. phagocytophilum infection Infection with A. phagocytophilum affects gene expression and protein production in ticks in a tissue-specific manner, but the effect on host protein content in different tick tissues has not been characterized. To address this question, sheep host proteins present in midguts and salivary glands were characterized in uninfected and A. phagocytophilum-infected I. scapularis female ticks. A total of 1,753 sheep host proteins were identified in fed adult female ticks, of which 473 were identified with more than one peptide per protein in at least one of the samples (see Additional file 2: Table S1). Of these, 1,151 (364 identified with more than one peptide per protein) and 1,282 (414 identified with more than one peptide per protein) proteins were identified in tick midguts and salivary glands, respectively (see Additional file 3: Dataset S2). Of the host proteins identified with more than one peptide per protein, 388 proteins were found in both midguts and salivary glands (see Additional file 3: Dataset S2). Sheep host proteins showing statistically significant protein abundance changes on the basis of Zq, or standardized log2-ratio of A. phagocytophilum-infected versus non-infected samples (Zq > 2, Zq < −2), were selected among proteins identified with more than one peptide in at least one of the samples [12, 39]. A total of 48 (6 underrepresented and 42 overrepresented) and 50 (36 underrepresented and 14 overrepresented) differentially represented sheep host proteins were found in tick midguts and salivary glands, respectively (see Additional file 3: Dataset S2). Of them, only 8 proteins were found in both midguts and salivary glands. The GO analysis of differentially represented sheep host proteins showed that most biological processes (BPs) were found in both tick midguts and salivary glands (Fig. 1a, b). However, immune response, other and oxygen transport BPs contained 68 % of the differentially represented sheep host proteins in tick midguts while in salivary glands the most represented BPs were other, unknown, oxygen transport and translation, containing 64 % of the differentially represented sheep host proteins (Fig. 1a, b). These results evidenced a clear difference in the host protein content between tick midguts and salivary glands in response to A. phagocytophilum infection. Additionally, although the total number of differentially represented sheep host proteins was similar between tick tissues, 88 % of the differentially represented host proteins in midguts were overrepresented while in salivary glands only 28 % of the differentially represented proteins were overrepresented in infected ticks when compared to uninfected controls (Fig. 1c, d). Furthermore, while sheep host stress response and transcription/DNA replication proteins were overrepresented in midguts and salivary glands, proteins in the lipid metabolism, oxygen transport and immune response BPs were overrepresented in midguts and underrepresented in salivary glands in response to A. phagocytophilum infection (Fig. 1c, d).Fig. 1 Tissue-specific effect of A. phagocytophilum infection on sheep host proteins represented in ticks. The results demonstrated a clear difference in the sheep host protein content between tick midguts and salivary glands in response to A. phagocytophilum infection. a Biological processes of differentially represented sheep host proteins in infected female tick midguts. b Biological processes of differentially represented sheep host proteins in infected female tick salivary glands. c Number of underrepresented and overrepresented sheep host proteins in different biological processes in infected female tick midguts when compared to uninfected controls. d Number of underrepresented and overrepresented sheep host proteins in different biological processes in infected female tick salivary glands when compared to uninfected controls Sheep host heat shock and chromatin-related proteins are overrepresented in response to A. phagocytophilum infection in tick midguts and salivary glands To characterize the putative physiological role of the host proteins differentially represented in tick midguts and salivary glands in response to A. phagocytophilum infection, we first focused on sheep host stress response and transcription/DNA replication proteins that were overrepresented in both tick tissues (Fig. 1c, d). The results showed that two sheep heat shock proteins (HSPs), HSP60 and HSP70, were overrepresented in infected tick midguts and salivary glands, respectively when compared to uninfected controls (Table 1). In the transcription/DNA replication BP, three host proteins involved in chromatin structure and function were overrepresented in response to A. phagocytophilum infection in tick midguts and salivary glands (Table 1).Table 1 Sheep host stress response, transcription/DNA replication, lipid metabolism, and immune response proteins differentially represented in tick midguts and salivary glands in response to A. phagocytophilum infection ID Description Log2 (infected/uninfected) fold change Function Midguts Salivary glands Stress response proteins  P31081 HSP60 +2.9 ns Response to cold  P0CB32 HSP70 ns +1.4 Heat shock response Transcription/DNA replication  P62803 Histone H4 +3.1 ns Chromatin structure  P68432 Histone H3.1 +3.1 ns Chromatin structure  F1MN93 TOP1 uncharacterized protein ns +2.7 Chromatin binding Lipid metabolism  W5QHX9 Phospholipase B +2.3 -1.3 Lipid absorption  Q9GL30 Phospholipase B +1.9 ns Lipid absorption  P15497 Apolipoprotein A-I +1.9 ns Cholesterol transport  Q32PF2 ATP-citrate synthase ns -1.5 Lipid synthesis  Q9TTS3 Acetyl-CoA carboxylase 1 ns -1.6 Lipid synthesis Immune response  W5NQK9 S100A8 +3.8 ns Innate immunity  W5NQJ0 S100A12 +3.2 ns Innate immunity  W5NQH6 S100A9 +3.1 ns Innate immunity  P28783 S100A9 +2.6 ns Innate immunity  D8X187 Serpin peptidase inhibitor clade B ovalbumin member 1 +2.1 ns Innate immunity  P62808 Histone H2B type 1 +3.4 ns Adaptive immunity  W5PGJ7 PYD and CARD Domain-Containing uncharacterized protein +2.9 ns Adaptive immunity  P49928 Cathelin-related peptide SC5 +3.1 ns Anti-bacterial immunity  P82018 Cathelicidin-2 +2.7 ns Anti-bacterial immunity  P50415 Cathelicidin-3 +2.3 ns Anti-bacterial immunity  P79360 Myeloid antimicrobial peptide +2.6 ns Anti-bacterial immunity  W5P7S6 Alpha-1-acid glycoprotein +2.6 nf Regulation of the immune response  W5PLV3 RAB5B uncharacterized protein +2.4 nf Antigen processing and presentation  W5PSQ7 Ig-like uncharacterized protein ns -1.3 Adaptive immunity  G5E513 Ig-like uncharacterized protein ns -1.8 Adaptive immunity  G5E5T5 Ig-like uncharacterized protein ns -2.7 Adaptive immunity  F1MQF6 Apoptosis-associated speck-like protein-containing a CARD nf -1.5 Innate immunity  W5PGJ7 LOC101105208 uncharacterized protein ns -1.6 Anti-bacterial immunity Abbreviations: ID protein (Uniprot; http://www.uniprot.org) accession numbers; +, overrepresented proteins in infected vs uninfected ticks; -, underrepresented proteins in infected vs uninfected ticks; nf, not found; ns, not significant Sheep host proteins involved in lipid metabolism and immune response are overrepresented in tick midguts but underrepresented in salivary glands in response to A. phagocytophilum infection To characterize further the putative physiological role of the host proteins differentially represented in response to A. phagocytophilum infection, we then focused on sheep host proteins in the lipid metabolism and immune response BPs that were overrepresented in midguts and underrepresented in salivary glands in response to A. phagocytophilum infection (Fig. 1c, d). The host lipid metabolism proteins overrepresented in A. phagocytophilum-infected tick midguts when compared to uninfected controls included proteins involved in lipid absorption, transport and excretion (Table 1). In tick salivary glands, sheep host proteins involved in lipid synthesis were underrepresented in infected ticks when compared to uninfected controls (Table 1). In tick midguts, sheep host immune response proteins that were overrepresented in response to A. phagocytophilum infection included proteins involved in innate immunity (including several S100 proteins), adaptive immunity, anti-bacterial immunity, regulation of the immune response, and antigen processing and presentation (Table 1). In tick salivary glands, sheep host immunoglobulin (Ig)-like proteins and proteins involved in innate and anti-bacterial immunity were underrepresented in response to A. phagocytophilum infection (Table 1). To confirm the origin of selected differentially represented proteins (Table 1), all of the peptides used to identify the proteins sharing tryptic peptides with I. scapularis proteins were revised to show the sequence of the peptides that are exclusive for host-derived proteins (Table 2).Table 2 Identification of host-derived proteins with identical tryptic peptides to I. scapularis tick homologues ID Description Unique host-derived peptides P31081 HSP60 ALMLQGVDLLADAVAVTMGPK VGGTSDVEVNEK VGGTSDVEVNEKKDR P0CB32 HSP70 FDLTGIPPAPR RKELEQVCNPIITK P68432 Histone H3.1 RVTIMPKDIQLAR SAPATGGVK SAPATGGVKKPHRYRPGTVALR F1MN93 TOP1 uncharacterized protein AGNEKEEGETADTVGCCSLR HLQDLMEGLTAK P62808 Histone H2B type 1 AMGIMNSFVNDIFER EIQTAVRLLLPGELAK EIQTAVR ESYSVYVYK SRKESYSVYVYK STITSREIQTAVRLLLPGELAK STITSREIQTAVR VLKQVHPDTGISSK W5PLV3 RAB5B uncharacterized protein TAMNVNDLFLAIAK P62803 Histone H4 None Q9TTS3 Acetyl-CoA carboxylase 1 None To confirm the origin for selected differentially represented proteins (Table 1), all of the peptides used to identify the proteins sharing tryptic peptides with I. scapularis proteins were revised. The peptides unique for host-derived proteins are shown. For protein P62803, we could not define the origin due to 100 % homology between sheep and tick proteins. For protein Q9TTS3, all peptides used for identification were identical in both host and tick-derived proteins Anaplasma phagocytophilum infection impacts on vertebrate host hemoglobin content in tick midguts and salivary glands The identification of differentially represented sheep host proteins in A. phagocytophilum-infected I. scapularis midguts and salivary glands suggested the question about the origin of these proteins. The analysis of cell compartment GO showed that over 50 % of the proteins were extracellular or associated with blood cells (Fig. 2a; see Additional file 3: Dataset S2). Nevertheless, other proteins were localized in the cell cytoplasm (Fig. 2a; see Additional file 3: Dataset S2), probably associated with host blood cells ingested by ticks during feeding. Most of the host proteins in the hemoglobin complex and blood microparticle classification were sheep hemoglobins in the oxygen transport BP represented in both tick midguts and salivary glands (Fig. 2b). These hemoglobins were overrepresented in midguts and underrepresented in salivary glands of A. phagocytophilum-infected ticks when compared to uninfected controls (Fig. 2c), a result that was corroborated by an independent analysis using a specific ELISA test (Fig. 2d).Fig. 2 Sheep host hemoglobin levels vary in a tissue-specific manner in response to A. phagocytophilum infection in ticks. a Cell compartment classification of differentially represented sheep host proteins in infected female tick midguts and salivary glands. b Venn diagram of the sheep host hemoglobin differentially represented in infected vs uninfected tick tissues. c Differential host hemoglobin protein representation in response to A. phagocytophilum infection in tick midguts and salivary glands. d Hemoglobin levels in tick midguts and salivary glands from A. phagocytophilum-infected and uninfected ticks determined by ELISA in individual tick protein extracts, represented as the mean + standard deviation (SD) and compared between samples from infected and uninfected ticks by Student’s t-test with unequal variance (P < 0.05; 2 biological replicates) Anaplasma phagocytophilum manipulates host protein content through modification of tick proteolytic pathways To aid in the probable mechanism responsible for the differential representation of host proteins in midguts and salivary glands of A. phagocytophilum-infected ticks when compared to uninfected controls, the tissue-specific effect of infection was characterized on tick hemoglobinolytic enzymes and other proteases. The transcriptomics and proteomics data used in this study was previously validated by real-time RT-PCR and Western blot or immunofluorescence for selected genes and proteins, respectively [12, 43]. Nevertheless, 5 selected genes coding for hemoglobin digesting enzymes differentially regulated in response to A. phagocytophilum infection were used for analysis by real-time RT-PCR in individual tick midguts and salivary glands (see Additional file 2: Figure S1). As previously discussed [12], the differences observed between the results of both analyses that were evident in tick midguts considering the absence of transcriptomics data for some genes in salivary glands, could be attributed to intrinsic variation in gene expression and the fact that approximately 85 % of the ticks used for RNAseq were infected [44], while for real-time RT-PCR all ticks were confirmed uninfected or infected with A. phagocytophilum before analysis. At the protein level, an antibody recognizing tick Cathepsin L was used to corroborate proteomics results by IFA. Similar to proteomics analysis (Table 3), the results showed protein underrepresentation in the salivary glands of A. phagocytophilum-infected ticks when compared to uninfected controls (Fig. 3). Furthermore, although proteomics data were not available, Cathepsin L was overrepresented in tick midguts in response to infection (Fig. 3). Therefore, considering these results, the analysis of the differential expression/representation of tick hemoglobinolytic enzymes in response to A. phagocytophilum infection was presented by pondering mRNA (transcriptomics RNAseq and real-time RT-PCR) and protein (proteomics) data (see Additional file 2: Figure S2).Table 3 Differential expression/representation of enzymes involved in tick hemoglobinolytic and heme transport pathways in response to A. phagocytophilum infection ID Description Log2 (infected/uninfected) fold change (mRNA/protein) Role in hemoglobin digestion Midguts Salivary glands Tick hemoglobinolytic pathway  EF428204 Cathepsin D nf/ns ns/nf Primary cleavage  A4GTA5  HQ615697 Cathepsin D2 nf/ns ns/nf Primary cleavage  E7E820  ISCW000202 Legumain -0.6/ns ns/ns Primary cleavage  B7P6S9  ISCW015983 Legumain -1.3/nf nf/nf Primary cleavage  B7P2C6  ISCW000076 Cathepsin L -7.3/nf nf/nf Primary and secondary cleavage  B7P3N8  JX502821 Cathepsin L nf/nf nf/-2.9 Primary and secondary cleavage  J9QSA1  ISCW000080 Cathepsin B +0.5/ns nf/-2.1 Secondary and tertiary cleavage  B7P3P1  EU551624 Cathepsin B nf/ns nf/-2.0 Secondary and tertiary cleavage  B7SP39  ISCW013346 Cathepsin B -1.6/ns ns/-2.5 Secondary and tertiary cleavage  B7QCU7  ISCW000078 Cathepsin B +0.2/ns -4.3/-1.8 Secondary and tertiary cleavage  B7P3P0  ISCW003494 Cathepsin C +0.3/ns -2.6/+2.2 Tertiary cleavage  B7PEB4  ISCW001779 Leucine aminopeptidase +2.2/ns ns/nf Tertiary cleavage  B7P2N4  ISCW023735 Leucine aminopeptidase +2.0/ns +1.1/ns Tertiary cleavage  B7QLQ7  ISCW001780 Leucine aminopeptidase +1.4/ns +0.5/ns Tertiary cleavage  B7P2N5  ISCW013904 Serine carboxipeptidase +1.0/ns ns/ns Tertiary cleavage  B7QLB7  ISCW024536 Serine carboxipeptidase -0.5/ns ns/nf Tertiary cleavage  B7Q049  ISCW024751 Serine carboxipeptidase -0.7/ns ns/ns Tertiary cleavage  B7QD81  ISCW024883 Serine carboxipeptidase -4.5/nf -2.4/nf Tertiary cleavage  B7QK83  ISCW007492 Serine carboxipeptidase -1.7/ns ns/ns Tertiary cleavage  B7PTE5  ISCW003059 Serine carboxipeptidase -1.2/nf ns/nf Tertiary cleavage  B7PC00 Tick heme transport pathway  ISCW001847 Heme-responsive gene 1 (HRG1) -0.8/nf ns/nf Heme transporter  B7P8M4  ISCW021709 Heme-binding lipoprotein (HELP) +3.3/ns -0.2/ns Heme transporter  B7Q406  ISCW013727 Vitellogenin 1 (VG1) ns/ns ns/nf Heme transporter  B7QJ67  ISCW021228 Vitellogenin 2 (VG2) -1.1/nf ns/nf Heme transporter  B7Q7E5 Transcriptomics RNAseq and proteomics data from A. phagocytophilum-infected and uninfected tick samples were obtained from Ayllón et al. [12]. Except for Cathepsins D and D2, which were included to show that these proteins were identified in the proteomics analysis but were not significantly different between infected and uninfected samples, only genes/proteins with statistically significant differences in at least one of the analyses (transcriptomics or proteomics) and samples (midguts or salivary glands) were included. Abbreviations: ID, gene (GenBank; http://www.ncbi.nlm.nih.gov) and protein (Uniprot; http://www.uniprot.org) accession numbers +, upregulated/overrepresented genes/proteins in infected vs uninfected ticks; -, downregulated/underrepresented genes/proteins in infected vs uninfected ticks; nf, not found; ns, not significant Fig. 3 Characterization of Cathepsin L protein levels by IFA. Representative images of IFA of midguts and salivary glands of uninfected and A. phagocytophilum-infected adult female I. scapularis. Tick tissues were stained with rabbit anti-Cathepsin L (mature region No. pab0213-0; Covalab, Villeurbanne, France) antibodies (green, FITC) or DAPI (blue), and images were superimposed after staining (right panels). Preimmune control serum-treated samples showed similar results for uninfected and infected ticks. Uninfected and infected samples stained with anti-Cathepsin L antibodies showed higher protein levels in infected midguts while Cathepsin L was underrepresented in infected salivary glands when compared to uninfected controls (arrowheads). Scale-bars: 10 μm The results suggested that in midguts from A. phagocytophilum-infected ticks when compared to uninfected controls, the hemoglobin primary cleavage was inhibited after Legumain underrepresentation while hemoglobin secondary and tertiary cleavages were probably not affected (Fig. 4 and Table 3). The hemoglobinolytic enzymes were also found in tick salivary glands, suggesting a role in hemoglobin digestion in this tissue (Table 3). In the salivary glands of infected ticks when compared to uninfected controls, the results suggested that hemoglobin primary and secondary cleavages were inhibited because Cathepsins L and B were underrepresented in response to infection while the hemoglobin tertiary cleavage was probably not affected (Fig. 4 and Table 3). The analysis of sheep hemoglobin alpha 1/2 (P68240) and beta (P02075) peptides identified by MS in tick midguts and salivary glands showed the presence of potential cleavage sites for trypsin (used in protein digestion for MS analysis), Legumain, Cathepsin B, Cathepsin C, Leucine aminopeptidase and Serine carboxipeptidase, therefore providing additional support for the activity of these enzymes in both tick tissues (Fig. 5a; see Additional file 1: Dataset S1).Fig. 4 The levels of enzymes involved in the tick hemoglobinolytic pathway vary in a tissue-specific manner in response to A. phagocytophilum infection. Differential expression/representation of tick hemoglobinolytic enzymes in response to A. phagocytophilum infection was obtained from Ayllón et al. [12] and represented by pondering mRNA (transcriptomics RNAseq and real-time RT-PCR) and protein (proteomics) data. In tick midguts, the hemoglobinolytic pathway operating in the endosomal digestive vesicle was revised by Sojka et al. [19]. In tick salivary glands, these enzymes are also produced and may function under different conditions. Abbreviations: AMP, hemoglobin-derived antimicrobial peptides (Hemocidins and other); HRG1, Heme-responsive gene 1; HELP, Heme-binding lipoprotein; VG1, Vitellogenin 1; VG2, Vitellogenin 2 Fig. 5 The digestion of sheep host hemoglobin varies between tick midguts and salivary glands in a tissue-specific manner in response to A. phagocytophilum infection. a Sheep hemoglobin alpha 1/2 (P68240) and beta (P02075) peptides detected by MS analysis with 1 % FDR in midguts and salivary glands from uninfected and A. phagocytophilum-infected ticks. Peptides detected in uninfected and infected (blue), infected (red), or uninfected (green) tick midguts and in uninfected and infected (underlined), infected (bold), or uninfected (italics) tick salivary glands are shown. Hemoglobin protein coverage by detected peptides is highlighted in green. The preferred cleavage sites for Trypsin and hemoglobinolytic enzymes are shown over P1 amino acid for Trypsin (t), Leucine aminopeptidase (l), Legumain (g), Cathepsin B (b), Cathepsin C (c), and Serine carboxypeptidase (s) (see Additional file 3: Dataset S2). b The number of protease genes/proteins different from hemoglobinolytic enzymes and differentially expressed/represented in response to A. phagocytophilum infection in tick midguts and salivary glands were extracted from transcriptomics and proteomics data [12]. c The number of protease inhibitor genes/proteins differentially expressed/represented in response to A. phagocytophilum infection in tick midguts and salivary glands were extracted from transcriptomics and proteomics data [12] In the midguts of ticks infected with A. phagocytophilum, Heme-responsive gene 1 protein (HRG1) was underrepresented, while heme transport proteins Heme-binding lipoprotein (HELP) and Vitellogenin 2 (VG2) but not Vitellogenin 1 (VG1) were overrepresented and underrepresented, respectively, in response to infection (Fig. 4 and Table 3). Furthermore, sheep host blood coagulation factors Annexin A3 (Q3SWX7; overrepresented in infected tick midguts and involved in blood anti-coagulation as a Phospholipase 2 inhibitor), and Fibrinogen gamma-B, and uncharacterized protein APOH (P12799 and W5Q268; underrepresented in infected tick salivary glands and involved in blood coagulation) were differentially represented in infected ticks when compared to uninfected controls (see Additional file 3: Dataset S2), resulting in the inhibition of blood coagulation in both tick tissues. In addition to hemoglobinolytic enzymes, other tick proteases were upregulated/overrepresented while protease inhibitors were down-regulated or did not change in midguts and salivary glands of A. phagocytophilum-infected ticks when compared to uninfected controls (Fig. 5b, c). However, as shown for the hemoglobinolytic enzymes (Fig. 3), the tick proteases differentially regulated in response to infection were predominantly different between midguts and salivary glands (see Additional file 2: Table S2). Discussion The characterization of sheep host proteins in the midguts and salivary glands of uninfected and A. phagocytophilum-infected I. scapularis female ticks showed tissue-specific differences in response to infection. Vertebrate host proteins in the transcription, lipid metabolism, immune response and oxygen transport (hemoglobins) were previously found to be highly abundant in the saliva of engorged I. scapularis ticks [24, 30]. The authors suggested that ticks have evolved mechanisms to selectively secrete host proteins in the saliva to aid in the feeding process [30]. Furthermore, anti-microbial peptides (AMP) such as S100 proteins [45] highly abundant in the saliva of engorged I. scapularis were proposed to function in clearing microbes from the feeding site to preserve ticks [30]. In A. phagocytophilum-infected ticks, host proteins from some of these pathways such as stress response and transcription were overrepresented in tick midguts and salivary glands, supporting the existence of a mechanism to facilitate tick feeding that was enhanced in response to infection. However, other host proteins overrepresented in A. phagocytophilum-infected tick midguts probably reflected the host response to infection. For example, proteins in the immune response BP are upregulated at the transcriptional level in sheep infected with A. phagocytophilum [46]. Nevertheless, immune response proteins were underrepresented in infected tick salivary glands when compared to uninfected controls, suggesting that A. phagocytophilum selectively manipulates the levels of host proteins to facilitate pathogen infection, multiplication and transmission. The infection with A. phagocytophilum modulates lipid metabolism in vertebrate host cells and bacteria incorporate host cholesterol for survival [47–49]. In tick cells, A. phagocytophilum infection inhibits lipid metabolism through down-representation of tick proteins [50]. The overrepresentation of sheep host proteins involved in lipid absorption, transport and secretion in midguts and the underrepresentation of lipid synthesis proteins in salivary glands of infected ticks when compared to uninfected controls may constitute an additional mechanism by which A. phagocytophilum selectively manipulates lipid metabolism to enhance infection and multiplication in tick tissues. Although ticks contain genes that encode heme synthesis enzymes, recent results demonstrate that they do not synthesize heme but obtain heme from the vertebrate host hemoglobin in the midgut and from tick heme transporters HELP/VG1/VG2 in other tissues [50–52]. Recently, Hajdusek et al. [50] proposed that the heme produced after host hemoglobin digestion is transported outside the endosomal digestive vesicle by HRG1 and subsequently detoxified in the hemosome or transported by HELP, VG1 and VG2 to other tick tissues such as salivary glands. However, as shown here and in previous reports [30], the presence of active tick hemoglobinolytic enzymes in the salivary glands and secreted in the saliva of engorged I. scapularis suggests the possibility that host hemoglobin may be also digested under different conditions to provide heme in the salivary glands. Although heme may not contribute to the cellular iron pool in ticks [52], the results reported here suggested that A. phagocytophilum affects hemoglobin primary cleavage in tick midguts and salivary glands, probably to reduce the production of hemoglobin-derived AMP to facilitate pathogen multiplication [50]. Furthermore, although A. phagocytophilum infection did not affect most of the enzymes involved in hemoglobin secondary and tertiary cleavage in tick midguts, the underrepresentation of HGR1 suggested a mechanism to reduce heme release into the cytoplasm of midgut cells (Fig. 4). This mechanism is probably manipulated by A. phagocytophilum to facilitate infection through reduction of the antimicrobial oxidative burden caused by reactive oxygen species (ROS) generated after heme release [19, 53, 54]. Furthermore, the inhibition of blood coagulation may be a mechanism driven by tick and/or A. phagocytophilum to facilitate tick feeding and pathogen multiplication. As recently proposed [13], these mechanisms may have evolved to guarantee A. phagocytophilum infection, multiplication and transmission while preserving tick life cycle. Once shown that A. phagocytophilum selectively manipulates the levels of vertebrate host proteins in ticks in a tissue-specific manner, the next question was why these proteins were selected among all host proteins ingested by ticks during blood feeding? Some of these host proteins such as hemoglobins, S100 and Ig-like proteins that were overrepresented in tick midguts and underrepresented in salivary glands have a crucial role during tick feeding and pathogen infection. In feeding ticks, host hemoglobins are the source of heme and AMP, which together with other immune system proteins such as S100 and Ig-like proteins may be essential for tick feeding and antimicrobial response to control microbe levels in tick tissues [19, 30, 45, 50, 53, 54]. Additionally, most of these proteins are highly conserved among major domestic and natural vertebrate hosts for A. phagocytophilum and I. scapularis (see Additional file 2: Table S3). Therefore, these results suggested that the mechanisms responsible for the selective manipulation of vertebrate host proteins by A. phagocytophilum infection in tick tissues are evolutionary conserved. The physiological significance of these findings was addressed by responding to the questions recently proposed by Sojka et al. [19] for a better understanding of how ticks handle the blood meal. Among these questions they proposed to address if the same tick enzyme machinery process hemoglobin and other vertebrate host proteins and the role of blood digestion and chemical reduction-oxidation reaction balance on pathogen infection and transmission in the tick midgut. The results of our study showed that tick hemoglobinolytic enzymes are present and active in both midguts and salivary glands of fed ticks and therefore may be involved in the digestion of hemoglobin and other host proteins. Although the effect of hemoglobin digestion and ROS production on pathogen infection and transmission was not directly addressed in our study, the results suggested that A. phagocytophilum selectively manipulate these and other processes to facilitate pathogen infection, multiplication and transmission. The results reported here suggested that the mechanism used by A. phagocytophilum to selectively manipulate the levels of vertebrate host proteins in a tissue-specific manner is through modification of tick proteolytic pathways. How A. phagocytophilum modify tick proteolytic pathways is not known, but may include the regulation of gene expression through epigenetic mechanisms recently shown to be affected by pathogen infection in I. scapularis [43]. These epigenetic mechanisms are probably controlled by secreted bacterial effectors [55–58]. However, future experiments should address the physiological significance of tick proteolytic pathways during A. phagocytophilum infection and multiplication in midguts and salivary glands. Conclusions In summary, the results of this study corroborated that vertebrate host proteins are present in the midguts and salivary glands of fed female I. scapularis. To our knowledge, the results presented here showed for the first time that A. phagocytophilum selectively manipulates the levels of vertebrate host proteins in the tick vector to facilitate pathogen infection, multiplication and transmission while preserving tick feeding and development (Fig. 6). The mechanisms by which A. phagocytophilum manipulates the levels of vertebrate host proteins are not known, but may include modification of proteolytic pathways by affecting tick epigenetics and other biological processes.Fig. 6 Proposed functional significance for the host proteins selectively manipulated by A. phagocytophilum in tick midguts and salivary glands. These results suggested that A. phagocytophilum selectively manipulates the levels of vertebrate host proteins in the tick midguts (MG) and salivary glands (SG) to facilitate pathogen infection, multiplication and transmission while preserving tick feeding and development Despite the growing burden that A. phagocytophilum and other tick-borne pathogens represent for human and animal health worldwide, effective control measures have not been developed [59]. Investigating the biological relevance of host proteins in tick biology and pathogen infection and the mechanisms used by A. phagocytophilum to manipulate host protein content is essential to advance our knowledge of tick-host-pathogen molecular interactions. These results have implications for the identification of new targets for the development of vaccines for the control of tick-borne diseases. Additional files Additional file 1: Dataset S1. Digestion of sheep host hemoglobin in tick midguts and salivary glands. (XLS 55 kb) Additional file 2: Figure S1. Analysis of RNAseq results by real-time RTPCR. Figure S2. Pondering mRNA and protein data. Table S1. Genes and oligonucleotide primers selected for gene expression analysis by real-time RT-PCR. Table S2. Differential expression of highly differentially regulated tick protease genes in response to A. phagocytophilum infection. Table S3. Percent homology for selected differentially represented host proteins in tick midguts and salivary glands in response to A. phagocytophilum infection. (PDF 337 kb) Additional file 3: Dataset S2. Proteomics results for sheep host proteins in tick tissues. (XLSX 208 kb) Abbreviations HGAhuman granulocytic anaplasmosis MSmass spectrometry RNAseqRNA sequencing FASPfilter aided sample preparation TFAtrifluoroacetic acid TEABTriethyLammonium bicarbonate LCliquid chromatography FDRfalse discovery rate GOgene ontology IFAimmunofluorescence assay BSAbovine serum albumin BPbiological process HSPheat shock protein Igimmunoglobulin HRG1Heme-responsive gene 1 protein HELPHeme-binding lipoprotein VG1Vitellogenin 1 VG2Vitellogenin 2 ROSreactive oxygen species AMPanti-microbial peptides Acknowledgements We would like to acknowledge José Ramón Marín Tébar (University of Castilla - La Mancha, Spain) for technical assistance with immunofluorescence. Funding This research was supported by the Ministerio de Economia y Competitividad (Spain) grant BFU2011-23896 and the European Union (EU) Seventh Framework Programme (FP7) ANTIGONE project number 278976. NA was funded by Ministerio de Economía y Competitividad, Spain. MV was supported by the Research Plan of the University of Castilla - La Mancha, Spain. The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication. Availability of data and material The quantitative transcriptomics and proteomics data for midguts and salivary glands from uninfected and A. phagocytophilum-infected I. scapularis were obtained from previously published results and deposited at the Dryad Digital Repository database with the dataset identifier http://dx.doi.org//10.5061/dryad.50kt0 [12]. Proteomics data for sheep host proteins are available via Peptide Atlas (http://www.peptideatlas.org) with identifier PASS00854. Authors’ contributions JF and MV conceived and supervised the study. VL, NA, ACC, JAL, JV and PA performed the experiments. MV and JF wrote the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate Animals were housed and experiments conducted with the approval and supervision of the Oklahoma State University Institutional Animal Care and Use Committee (Animal Care and Use Protocol, ACUP No. VM1026). ==== Refs References 1. Jones KE Patel NG Levy MA Storeygard A Balk D Gittleman JL Daszac P Global trends in emerging infectious diseases Nature 2008 451 990 994 10.1038/nature06536 18288193 2. de la Fuente J Estrada-Peña A Venzal JM Kocan KM Sonenshine DE Overview: Ticks as vectors of pathogens that cause disease in humans and animals Front Biosci 2008 13 6938 6946 10.2741/3200 18508706 3. Estrada-Peña A Ostfeld RS Peterson AT Poulin R de la Fuente J Effects of environmental change on zoonotic disease risk: an ecological primer Trends Parasitol 2014 30 205 214 10.1016/j.pt.2014.02.003 24636356 4. Gortazar C Reperant LA Kuiken T de la Fuente J Boadella M Martínez-Lopez B Crossing the interspecies barrier: Opening the door to zoonotic pathogens PLoS Pathog 2014 10 e1004129 10.1371/journal.ppat.1004129 24945247 5. Dumler JS Barbet AC Bekker CPJ Dasch GA Palmer GH Ray SC Reorganization of the genera in the families Rickettsiaceae and Anaplasmataceae in the order Rickettsiales: unification of some species of Ehrlichia with Anaplasma , Cowdria with Ehrlichia and Ehrlichia with Neorickettsia , descriptions subjective synonyms of Ehrlichia phagocytophila Int J Syst Evol Microbiol 2001 51 2145 2165 10.1099/00207713-51-6-2145 11760958 6. Stuen S Anaplasma phagocytophilum - the most widespread tick-borne infection in animals in Europe Vet Res Commun 2010 31 79 84 10.1007/s11259-007-0071-y 17682851 7. Rikihisa Y Molecular pathogensis of Anaplasma phagocytophilum Current Microbiol Rev 2011 24 469 489 10.1128/CMR.00064-10 8. Severo MS Pedra JHF Ayllón N Kocan KM de la Fuente J Tang YW Sussman M Liu D Poxton I Schwartzman J Anaplasma Molecular Medical Microbiology 2015 2 London Academic Press-Elsevier 2033 2042 9. Dahlgren FS Mandel EJ Krebs JW Massung RF McQuiston JH Increasing incidence of Ehrlichia chaffeensis and Anaplasma phagocytophilum in the United States, 2000–2007 Am J Trop Med Hyg 2011 85 124 131 10.4269/ajtmh.2011.10-0613 21734137 10. Reichard MV Manzano-Roman R Kocan KM Blouin EF de la Fuente J Snider TA Inoculation of white-tailed deer (Odocoileus virginianus ) with Ap-V1 or NY-18 strains of Anaplasma phagocytophilum and microscopic demonstration of Ap-V1 in Ixodes scapularis adults that acquired infection from deer as nymphs Vector Borne Zoonotic Dis 2009 9 565 568 10.1089/vbz.2008.0106 18973438 11. Sukumaran B Narasimham S Anderson JF DePonte K Marcantonio K Krishnan MN An Ixodes scapularis protein required for survival of Anaplasma phagocytophilum in tick salivary glands J Exp Med 2006 6 1507 1517 10.1084/jem.20060208 16717118 12. Ayllón N, Villar M, Galindo RC, Kocan KM, Šíma R, López JA, et al. Systems biology of tissue-specific response to Anaplasma phagocytophilum reveals differentiated apoptosis in the tick vector Ixodes scapularis. PLoS Genet. 2015;11. 13. de la Fuente J Villar M Cabezas-Cruz A Estrada-Peña A Ayllón N Alberdi P Tick-host-pathogen interactions: conflict and cooperation PLoS Pathog 2016 12 e1005488 10.1371/journal.ppat.1005488 27099928 14. de la Fuente J Estrada-Peña A Cabezas-Cruz A Kocan KM Anaplasma phagocytophilum uses common strategies for infection of ticks and vertebrate hosts Trends Microbiol 2016 3 173 180 10.1016/j.tim.2015.12.001 26718986 15. Ribeiro JM Role of saliva in blood-feeding by arthropods Annu Rev Entomol 1987 32 463 478 10.1146/annurev.en.32.010187.002335 2880553 16. Anderson JM Sonenshine DE Valenzuela JG Exploring the mialome of ticks: an annotated catalogue of midgut transcripts from the hard tick, Dermacentor variabilis (Acari: Ixodidae) BMC Genomics 2008 9 552 10.1186/1471-2164-9-552 19021911 17. Francischetti IM Sa-Nunes A Mans BJ Santos IM Ribeiro JM The role of saliva in tick feeding Front Biosci 2009 14 2051 2088 10.2741/3363 18. Oliveira CJ Sá-Nunes A Francischetti IM Carregaro V Anatriello E Silva JS Deconstructing tick saliva: Non-protein molecules with potent immunomodulatory properties J Biol Chem 2011 286 10960 10969 10.1074/jbc.M110.205047 21270122 19. Sojka D Franta Z Horn M Caffrey CR Mareš M Kopáček P New insights into the machinery of blood digestion by ticks Trends Parasitol 2013 29 276 285 10.1016/j.pt.2013.04.002 23664173 20. Kotsyfakis M Schwarz A Erhart J Ribeiro JM Tissue- and time-dependent transcription in Ixodes ricinus salivary glands and midguts when blood feeding on the vertebrate host Sci Rep 2015 5 9103 10.1038/srep09103 25765539 21. Chmelař J Kotál J Karim S Kopacek P Francischetti IM Pedra JH Kotsyfakis M Sialomes and mialomes: A systems-biology view of tick tissues and tick-host interactions Trends Parasitol 2015 32 242 254 10.1016/j.pt.2015.10.002 26520005 22. Hajdušek O Šíma R Ayllón N Jalovecká M Perner J de la Fuente J Kopáček P Interaction of the tick immune system with transmitted pathogens Front Cell Infect Microbiol 2013 3 26 23875177 23. Mastronunzio JE Kurscheid S Fikrig E Postgenomic analyses reveal development of infectious Anaplasma phagocytophilum during transmission from ticks to mice J Bacteriol 2012 194 2238 2247 10.1128/JB.06791-11 22389475 24. Valenzuela JG Francischetti IM Pham VM Garfield MK Mather TN Ribeiro JM Exploring the sialome of the tick Ixodes scapularis J Exp Biol 2002 205 2843 2864 12177149 25. Madden RD Sauer JR Dillwith JW A proteomics approach to characterizing tick salivary secretions Exp Appl Acarol 2002 28 77 87 10.1023/A:1025342015065 14570118 26. Diaz-Martin V Manzano-Roman R Valero L Oleaga A Encinas-Grandes A Perez-Sanchez R An insight into the proteome of the saliva of the argasid tick Ornithodoros moubata reveals important differences in saliva protein composition between the sexes J Proteomics 2013 80 216 235 10.1016/j.jprot.2013.01.015 23416086 27. Oliveira CJ Anatriello E de Miranda-Santos IK Francischetti IM Sá-Nunes A Ferreira BR Ribeiro JM Proteome of Rhipicephalus sanguineus tick saliva induced by the secretagogues pilocarpine and dopamine Ticks Tick Borne Dis 2013 4 469 477 10.1016/j.ttbdis.2013.05.001 24029695 28. Tirloni L Reck J Terra RM Martins JR Mulenga A Sherman NE Proteomic analysis of cattle tick Rhipicephalus (Boophilus ) microplus saliva: a comparison between partially and fully engorged females PLoS One 2014 9 e94831 10.1371/journal.pone.0094831 24762651 29. Tirloni L Islam MS Kim TK Diedrich JK Yates JR 3rd Pinto AF Mulenga A Saliva from nymph and adult females of Haemaphysalis longicornis : a proteomic study Parasit Vectors 2015 8 338 10.1186/s13071-015-0918-y 26104117 30. Kim TK Tirloni L Pinto AF Moresco J Yates JR 3rd DA Silva Vaz I Jr Mulenga A Ixodes scapularis tick saliva proteins sequentially secreted every 24 h during blood feeding PLoS Negl Trop Dis 2016 10 e0004323 10.1371/journal.pntd.0004323 26751078 31. Wickramasekara S Bunikis J Wysocki V Barbour AG Identification of residual blood proteins in ticks by mass spectrometry proteomics Emerg Infect Dis 2008 14 1273 1275 10.3201/eid1408.080227 18680654 32. Villar M Torina A Nuñez Y Zivkovic Z Marina A Alongi A Application of highly sensitive saturation labeling to the analysis of differential protein expression in infected ticks from limited samples Proteome Sci 2010 8 43 10.1186/1477-5956-8-43 20704695 33. Francischetti IM Meng Z Mans BJ Gudderra N Hall M Veenstra TD An insight into the salivary transcriptome and proteome of the soft tick and vector of epizootic bovine abortion Ornithodoros Coriaceus J Proteomics 2008 71 493 512 10.1016/j.jprot.2008.07.006 18725333 34. Kocan KM de la Fuente J Coburn LA Insights into the development of Ixodes scapularis: A resource for research on a medically important tick species Parasit Vectors 2015 8 592 10.1186/s13071-015-1185-7 26576940 35. Wisniewski J Zougman A Nagaraj N Mann M Universal sample preparation method for proteome analysis Nat Meth 2009 6 359 362 10.1038/nmeth.1322 36. Martínez-Bartolomé S Navarro P Martín-Maroto F López-Ferrer D Ramos-Fernández A Villar M Properties of average score distributions of SEQUEST: the probability ratio method Mol Cell Proteomics 2008 7 1135 1145 10.1074/mcp.M700239-MCP200 18303013 37. Navarro P Vázquez J A refined method to calculate false discovery rates for peptide identification using decoy databases J Proteome Res 2009 8 1792 1796 10.1021/pr800362h 19714873 38. Bonzon-Kulichenko E Martínez-Martínez S Trevisan-Herraz M Navarro P Redondo JM Vázquez J Quantitative in-depth analysis of the dynamic secretome of activated Jurkat T-cells J Proteomics 2011 75 561 571 10.1016/j.jprot.2011.08.022 21920478 39. Navarro P Trevisan-Herraz M Bonzon-Kulichenko E Núñez E Martínez-Acedo P Pérez-Hernández D General Statistical framework for quantitative proteomics by stable isotope labeling J Proteome Res 2014 13 1234 1247 10.1021/pr4006958 24512137 40. Jorge I Navarro P Martinez-Acedo P Nunez E Serrano H Alfranca A Statistical model to analyze quantitative proteomics data obtained by 18O/16O labeling and linear ion trap mass spectrometry: application to the study of vascular endothelial growth factor-induced angiogenesis in endothelial cells Mol Cell Proteomics 2009 8 1130 1149 10.1074/mcp.M800260-MCP200 19181660 41. Barrett AJ McDonald JK Nomenclature: protease, proteinase and peptidase Biochem J 1986 237 935 10.1042/bj2370935 3541905 42. Popara M Villar M Mateos-Hernández L Fernández De Mera IG Marina A Del Valle M Lesser protein degradation machinery correlates with higher BM86 tick vaccine efficacy in Rhipicephalus annulatus when compared to R. microplus Vaccine 2013 31 4728 4735 10.1016/j.vaccine.2013.08.031 23973246 43. Cabezas-Cruz A Alberdi P Ayllón N Valdés JJ Pierce R Villar M de la Fuente J Anaplasma phagocytophilum increases the levels of histone modifying enzymes to inhibit cell apoptosis and facilitate pathogen infection in the tick vector, Ixodes scapularis Epigenetics 2016 11 303 319 10.1080/15592294.2016.1163460 27019326 44. Kocan KM Busby AT Allison RW Breshears MA Coburn L Galindo RC Sheep experimentally-infected with a human isolate of Anaplasma phagocytophilum serve as a host for infection of Ixodes scapularis Ticks Tick-Borne Dis 2012 3 147 153 10.1016/j.ttbdis.2012.01.004 22534515 45. Donato R Cannon BR Sorci G Riuzzi F Hsu K Weber DJ Geczy CL Functions of S100 Proteins Curr Mol Med 2013 13 24 57 10.2174/156652413804486214 22834835 46. Galindo RC Ayoubi P García-Pérez AL Naranjo V Kocan KM Gortazar C de la Fuente J Differential expression of inflammatory and immune response genes in sheep infected with Anaplasma phagocytophilum Vet Immunol Immunopathol 2008 126 27 34 10.1016/j.vetimm.2008.06.004 18640728 47. Lin M Rikihisa Y Ehrlichia chaffeensis and Anaplasma phagocytophilum lack genes for lipid A biosynthesis and incorporate cholesterol for their survival Infect Immun 2003 71 5324 5331 10.1128/IAI.71.9.5324-5331.2003 12933880 48. Manzano-Roman R Almazán C Naranjo V Blouin EF Kocan KM de la Fuente J Expression of perilipin in human promyelocytic cells in response to Anaplasma phagocytophilum infection results in modified lipid metabolism J Med Microbiol 2008 57 159 163 10.1099/jmm.0.47504-0 18201980 49. Severo MS Stephens KD Kotsyfakis M Pedra JH Anaplasma phagocytophilum : deceptively simple or simply deceptive? Future Microbiol 2012 7 719 731 10.2217/fmb.12.45 22702526 50. Villar M Ayllón N Alberdi P Moreno A Moreno M Tobes R Integrated metabolomics, transcriptomics and proteomics identifies metabolic pathways affected by Anaplasma phagocytophilum infection in tick cells Mol Cell Proteomics 2015 14 3154 3172 10.1074/mcp.M115.051938 26424601 51. Gulia-Nuss M Nuss AB Meyer JM Sonenshine DE Roe RM Waterhouse RM Genomic insights into the Ixodes scapularis tick vector of Lyme disease Nat Commun 2016 7 10507 10.1038/ncomms10507 26856261 52. Perner J Sobotka R Sima R Konvickova J Sojka D Oliveira PL Acquisition of exogenous haem is essential for tick reproduction eLife 2016 5 e12318 10.7554/eLife.12318 26949258 53. Graça-Souza AV Maya-Monteiro C Paiva-Silva GO Braz GR Paes MC Sorgine MH Adaptations against heme toxicity in blood-feeding arthropods Insect Biochem Mol Biol 2006 36 322 335 10.1016/j.ibmb.2006.01.009 16551546 54. Toh SQ Glanfield A Gobert GN Jones MK Heme and blood-feeding parasites: friends or foes? Parasit Vectors 2010 3 108 10.1186/1756-3305-3-108 21087517 55. Huang B Troese MJ Howe D Ye S Sims JT Heinzen RA Anaplasma phagocytophilum APH_0032 is expressed late during infection and localizes to the pathogen-occupied vacuolar membrane Microb Pathog 2010 49 273 284 10.1016/j.micpath.2010.06.009 20600793 56. Rikihisa Y Lin M Niu H Type IV secretion in the obligatory intracellular bacterium Anaplasma phagocytophilum Cell Microbiol 2010 12 1213 1221 10.1111/j.1462-5822.2010.01500.x 20670295 57. Rennoll-Bankert KE Garcia-Garcia JC Sinclair SH Dumler JS Chromatin bound bacterial effector AnkA recruits HDAC1 and modifies host gene expression Cell Microbiol 2015 17 1640 1652 10.1111/cmi.12461 25996657 58. Sinclair SH Yegnasubramanian S Dumler JS Global DNA methylation changes and differential gene expression in Anaplasma phagocytophilum -infected human neutrophils Clin Epigenetics 2015 7 77 10.1186/s13148-015-0105-1 26225157 59. de la Fuente J Contreras M Tick vaccines: current status and future directions Exp Rev Vaccines 2015 14 1367 1376 10.1586/14760584.2015.1076339

PMC005xxxxxx/PMC5000438.txt

==== Front Respir ResRespir. ResRespiratory Research1465-99211465-993XBioMed Central London 42210.1186/s12931-016-0422-8ResearchALK gene copy number gains in non-small-cell lung cancer: prognostic impact and clinico-pathological correlations Peretti U. u.peretti@gmail.com 1Ferrara R. robertoferrara86@gmail.com 1Pilotto S. +390458128502+390458128140sara.pilotto.85@gmail.comsara.pilotto@univr.it 1Kinspergher S. stefaniakinspergher@libero.it 1Caccese M. mario.caccese@hotmail.com 1Santo A. antonio.santo@ospedaleuniverona.it 1Brunelli M. matteo.brunelli@univr.it 2Caliò A. calioanna@gmail.com 2Carbognin L. luisa.carbognin@gmail.com 1Sperduti I. isperduti@yahoo.it 3Garassino M. marina.garassino@istitutotumori.mi.it 4Chilosi M. marco.chilosi@univr.it 2Scarpa A. aldo.scarpa@univr.it 25Tortora G. giampaolo.tortora@univr.it 1Bria E. emilio.bria@univr.it 11 Medical Oncology, University of Verona, Azienda Ospedaliera Universitaria Integrata, P.le L.A. Scuro 10, 37124 Verona, Italy 2 Department of Pathology and Diagnostics, University of Verona, Azienda Ospedaliera Universitaria Integrata, P.le L.A. Scuro 10, 37124 Verona, Italy 3 Biostatistics, Regina Elena National Cancer Institute, Via Elio Chianesi 53, 00144 Rome, Italy 4 Istituto Nazionale Tumori, Milan, Italy 5 ARC-NET Applied Research on Cancer Center, University of Verona, P.le L.A. Scuro 10, 37124 Verona, Italy 25 8 2016 25 8 2016 2016 17 1 1059 6 2016 18 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background The correlation between ALK gene copy number gain (ALK-CNG) and prognosis in the context of advanced non-small-cell lung cancer (NSCLC) remains a controversial issue. This study aimed to evaluate the association among ALK-CNG according to Fluorescent In Situ Hybridization (FISH), clinical characteristics and survival in resectable and advanced NSCLC. Methods Clinical and pathological data of patients with resectable and advanced NSCLC were retrospectively collected. Tumor tissues were analyzed for ALK-CNG by FISH, and patients were divided in 3 groups/patterns on the basis of ALK signals: disomic [Pattern A], 3–7 signals [Pattern B], >7 signals [Pattern C]. The association between clinical and pathological features and ALK-CNG patterns was evaluated. Disease/progression-free and overall survival (DFS/PFS and OS) were estimated using the Kaplan-Meyer method. Results A number of 128 (76.6 %) out of the 167 eligible patients were evaluable for ALK-CNG, displaying pattern A, B and C in 71 (42.5 %), 42 (25.1 %) and 15 (9 %) patients, respectively. Gains in ALK-CNG appear to be more frequent in smokers/former smokers than in non-smokers (74.2 % versus 20.4 %, respectively, p = 0.03). Pattern A and C seem more frequently associated with higher T-stage (T3-4), while pattern B appears more represented in lower T-stage (T 1-2) (p = 0.06). No significant differences in survival rate were observed among the above groups. Conclusions A high ALK-CNG pattern might be associated with smoking status and theoretically it might mirror genomic instability. The implications for prognosis should be prospectively investigated and validated in larger patients’ series. Trial registration We confirm that all the study was performed in accordance with relevant guidelines and regulations and that all the protocol (part of a larger project MFAG 2013 N.14282) was approved by the local Ethics Committee of the Azienda Ospedaliera Universitaria Integrata of Verona on November 11st, 2014. Keywords Lung cancerAnaplastic lymphoma kinaseCopy number gainPrognosisClinico-pathological characteristicsItalian Association for Cancer Research (AIRC)My First AIRC Grant n° 14282Bria E. International Association for the Study of Lung Cancer (IASLC)Young Investigational AwardPilotto S. issue-copyright-statement© The Author(s) 2016 ==== Body Background Nowadays, non-small-cell lung cancer (NSCLC) might be considered as a universe of different diseases. Particularly in the context of adenocarcinoma, reliable evidence is available suggesting that cancer development and progression might be led by the addiction from aberrant pathways triggered by genetic abnormalities acting as oncogenic drivers (such as the activating mutation of EGFR or the translocation of ALK). In this setting, the inhibition of these drivers with selective agents has radically changed the natural history of the disease [1–7]. Unfortunately, only a limited subpopulation of lung cancer patients might benefit from this personalized treatment. Hence, the importance of identifying and validating new molecular alterations with prognostic and predictive significance in order to extend the proportion of lung cancer patients who might benefit from targeted drugs. ALK is a versatile oncogene whose role has been recognized in a large variety of tumors through different activation mechanisms, mainly the chromosomal rearrangement with different fusion partners (as the microtubule associated protein EML4 in NSCLC or the nucleophosmin NPM1 in anaplastic large cell lymphoma). In NSCLC, the successful history of ALK inhibitors started with crizotinib and is still ongoing [8]. Crizotinib is an orally available tyrosine kinase inhibitor (TKI) originally designed to target the mesenchymal to epithelial transition process, but it also potently inhibits ALK and ROS1 phosphorylation and signaling [9]. Based on encouraging preclinical data, impressive preliminary results were published from an expansion cohort of a phase I trial [10] and from the PROFILE 1005 phase II trial [11]. In the second-line trial PROFILE 1007, 347 patients with ALK-rearranged NSCLC pretreated with a platinum doublet received either crizotinib or second-line chemotherapy with docetaxel or pemetrexed. Patients benefited from crizotinib in both terms of overall response rate (ORR) (65 % versus 20 %; p < 0.001) and median progression-free survival (PFS) (7.7 versus 3.0 months; HR 0.49; p < 0.001) [6]. Moreover, in the first-line phase III trial PROFILE 1014, 343 treatment-naive patients with ALK-rearranged NSCLC were randomized to receive either crizotinib or standard platinum-based plus pemetrexed first-line chemotherapy. Also in this setting, crizotinib met not only the primary end point, achieving a significantly longer median PFS (10.9 versus 7.0 months; HR 0.45; p < 0.001), but also a significantly higher ORR (74 % versus 45 %; p < 0.001). The more frequent adverse events associated with the administration of crizotinib were vision disorders, diarrhea, edema, increased aminotransferase levels and neutropenia [7]. The main limitations of the crizotinib efficacy are represented by its low brain penetrance and its inactivity against secondary mutations of the ALK gene. Therefore, clinical trials evaluating next generation ALK inhibitors are currently ongoing. Promising response rates and PFS have been reported particularly in crizotinib-refractory ALK-rearranged NSCLC patients treated with ceritinib [12–14], alectinib [15, 16], brigatinib [17] and lorlatinib [18]. Recently, the pre-planned interim analysis of the J-ALEX trial, demonstrated the superiority of alectinib to crizotinib in untreated ALK-rearranged NSCLC patients (median PFS not reached versus 10.2 months; HR 0.34; p < 0.0001) [19]. Another ALK aberration, the gene copy number gain (CNG), has been identified in several tumor types and significantly correlated with poor prognosis and/or advanced disease status. In this regard, ALK-CNG has been reported in the 10 % of renal cell carcinoma patients and the presence of more than 5 copies of the ALK gene was significantly associated with high tumor size, nuclear grade and worse 10-years survival rate [20]. In colorectal cancer patients, ALK-CNG was found in the 3.4 % of non-molecularly selected patients and in the 37 % of RAS-BRAF-PI3KCA wild-type patients [21, 22]. In both studies, the ALK gene copy number increase was significantly associated with poor prognosis. Moreover, in triple wild-type patients, the response rate to cetuximab or panitumumab was significantly higher in the subgroup of disomic ALK (70 %) as compared with ALK-CNG subgroup (32 %), with similar results in term of PFS and overall survival (OS), raising the hypothesis of a possible role of ALK-CNG in resistance to anti-EGFR therapy [21, 22]. ALK-CNG was further identified in the 13 % of patients affected by hepatocellular carcinoma, negative for serum hepatitis B virus DNA, with a significant correlation among ALK-CNG (≥4 copies versus < 4 copies), 3-year PFS rate (27 % versus 42 %) and 3-year OS rate (18 % versus 47 %) [23]. In rhabdomyosarcoma (RMS), ALK-CNG was detected in the 88 % of alveolar RMS and in the 52 % of embryonal RMS (ERMS). In ERMS, specific ALK gain in the primary tumor correlated with metastatic disease and poor 5-year disease-specific OS (62 versus 82 %) [24]. Moreover, ALK-CNG was retrospectively detected in the 47.2 % of patients with inflammatory breast cancer and significantly correlated with worse overall survival (24.9 versus 38.1 months) and recurrence free survival (RFS) after curative mastectomy (12.7 versus 43.3 months) compared to ALK-CNG negative patients [25]. Finally, an aberrant activation of ALK, both by activating mutation and amplification, might drive the tumorigenesis of neuroblastoma (NBL) [26]. In a retrospective analysis of pediatric patients with NBL, ALK amplification due to polyploidy (ALK/CEP2 ratio > 4) represented a negative prognostic factor for OS [27]. Regarding lung cancer, the presence and potential prognostic role of the different ALK gene aberrations (translocation and gene copy number gains) have been widely investigated with debatable results (as more extensively revised in the discussion paragraph). Our study, retrospectively conducted in a large series of NSCLC patients, aimed to investigate the potential correlation between ALK-CNG and clinical features, exploring in particular the prognostic implications of this genetic abnormality in resected and advanced NSCLC patients. Methods Patients and samples All the reported data of this study were obtained from the Lung Verona Database (Verona, Italy). Clinical (gender, familiarity, comorbidity, performance status, smoking history, treatment) and pathological (TNM, disease stage, grading) information has been collected. All the cases were classified according to WHO criteria [28]. Appropriate samples, containing at least 90 % neoplastic cells, have been selected for the interphase cytogenetic and molecular studies. Interphase fluorescence in situ hybridization (FISH) analysis ALK gene status assessment by break-apart probe Interphase cytogenetic analysis was performed by FISH using 5 μm sections from formalin-fixed and paraffin-embedded tissues and a commercially available break-apart ALK kit (Abbott-Vysis) that uses two DNA probes on the ALK gene, one at the 3’ and one at the 5’ regions. The slides were examined using an Olympus BX61 (Germany) with appropriate filters for SpectrumOrange, SpectrumGreen and the UV filter for the DAPI nuclear counterstain. The signals were recorded with a CCD camera (CytoVysion, Olympus) and digitalized by Fluo/D-SIGHT (Menarini/Visia Imaging). A total of 150 neoplastic nuclei were assessed in at least three different areas for surgical specimens. In bioptic samples all neoplastic nuclei (on average 60 nuclei) were evaluated. Nuclei harboring split-signals were scored as positive for ALK rearrangement using the approved cut-off (15 %). Gene copy number was initially scored per each case. Control probes using centromeric alpha-satellite specific for chromosome 2, 3 and 17 (CEP2, CEP3 and CEP17) probes Centromeric alpha-satellite specific for chromosome 2, 3 and 17 (CEP2, CEP3 and CEP17) were used as control probes (Vysis-Abbott, Olympus, Rome, Italy). Control probes polysomy was detected by performing FISH assay on adjacent serial tissue sections. Briefly, each probe was diluted 1:10 in tDenHyb2 buffer (Insitus, Albuquerque, NM). Ten microliters of diluted probe were applied to each slide and cover slips were placed over the slides. Denaturation was achieved by incubating the slides at 80 °C for 10 min in a humidified box; then hybridization was done at 37 °C for 16 h. The cover slips were then removed and the slides were immersed at room temperature in 0.5 x SSC (Superconducting Super Collider) for 2 min and in 2 x SSC for 2 min. The slides were air dried and counterstained with 10 μl DAPI/Antifade (DAPI in Fluorguard, 0.5 μg/ml, Insitus, Albuquerque, NM). Fluorescent in situ signals were evaluated on carcinomatous and normal pulmonary adjacent parenchyma. A total of 150 neoplastic nuclei were assessed in at least three different areas for surgical specimens while 60 neoplastic nuclei were evaluated in bioptic samples. ALK interpretation Mean copy number for the ALK locus (LSI) and the centromeric (CEPs) probes CEP2, CEP3 and CEP17 were primarily evaluated. A mean copy number of centromeric probes were secondly detected to assess ploidy. Ratio between mean copy number of ALK gene and mean copy number of control centromeric probes CEP2, CEP3 and CEP17 was finally scored. Amplification of the ALK locus gene was interpreted when the ratio (LSI/CEPs) was ≥ 2. When increasing gene copy number resulted < 2 after corrections by control probes (LSI/CEPs) the case was interpreted as having gains of chromosome due to polyploidy. Cases were digitalized by using the scan D-Sight/Fluo instrument (VisiaImaging, Florence, Italy). Statistical analysis According to what stated by the Cytogenetics Laboratory of Verona’s University about the frequency distribution of ALK expression as a cytogenetic profile [29], patients were divided in three groups on the base of gene copy number (Pattern A: CNG = 2 [disomic]; Pattern B: CNG 3-7; Pattern C: CNG >7). The Fisher’s Exact Test (with a significance α error of 0.5) was applied to evaluate the correlation of ALK-CNG subgroups with the clinico-pathologic variables in the overall population and the treatment outcome in metastatic patients. In order to correct possible biases due to the disease stage frequencies, patients were divided in resected (those who underwent surgery for an operable disease) and metastatic patients (metastatic from the diagnosis). The prognostic analysis has been performed separately for these two groups. The hazard ratio (HR) and the 95 % confidence intervals (95 % CI) were estimated for each variable using the Cox univariate model. The OS (2-year survival rate) was calculated using the Kaplan Meyer’s method to identify a possible correlation between clinical outcome and expression of ALK. We estimated the disease free survival (DFS) for resected patients and the PFS for patients who underwent systemic treatment for metastatic disease. Finally, Kaplan Meyer curves were compared through Long Rank test and all the analyses were conducted using SPSS 18.0 software. Results Patients’ characteristics A consecutive series of 205 NSCLC (112 biopsies, 93 surgical specimens) were collected from the Lung Verona Database. Thirty-eight patients with locally advanced or metastatic NSCLC harboring EGFR activating mutations and ALK rearranged tumors were excluded from the analysis. We analyzed a total of 167 patients and their clinico-pathological characteristics are summarized in Table 1. One hundred and six men (63.5 %) and 61 women (36.5 %) with a median age of 66 years (ranging from 29 to 85) were included. Among them, 56 patients (33.5 %) were smokers, 68 former smokers (40.7 %), 34 never smokers (20.4 %), while 9 patients (5.4 %) were not evaluable for smoking status; the 60.5 % of patients had familiarity for cancer; 25.7 % had cardiac and respiratory comorbidities. One hundred and one patients (60.4 %) were metastatic at the diagnosis; 26 patients (15.5 %) were stage I, 10 (6.1 %) stage II and 30 (18.0 %) stage III. The predominant histological subtype was adenocarcinoma with 147 cases (88 %), followed by squamous cell carcinoma with 7 cases (4.2 %) and mixed or other histotype were 11 cases (6.6 %). Among those patients evaluable for histologic tumor grade (69/167), 25.1 % were poor differentiated (G3). The majority of patients were in good clinical conditions with PS 0 (59.3 %) and 1 (21.6 %), while only the 7.1 % of patients were PS 2 or PS 3. Forty seven patients (28.1 %) affected by localized disease underwent surgery, in some cases followed by adjuvant chemotherapy; 3 patients were treated with neoadjuvant chemotherapy before surgery. Locally advanced patients were treated with chemo-radiotherapy in 17 cases (10.2 %). Seventy-five patients (44.9 %) with advanced disease were treated with chemotherapy tailored on the basis of histologic type, age, comorbidities and performance status. Overall, 89 patients with advanced disease were treated with I line chemotherapy, among them 3 patients (1.8 %) had complete response (CR), 31 (18.6 %) had partial response (PR), 20 (12 %) had stable disease (SD) and 35 patients (21 %) progressive disease (PD) (Table 1).Table 1 Patients’ characteristics (167 evaluable patients for the clinical analysis) Patients number (%) Gender  Male 106 (63.5)  Female 61 (36.5) Cancer familiarity  Yes 101 (60.5)  No 43 (25.7)  Unknown 23 (13.8) Comorbidity  Yes 52 (31.1)  No 115 (68.9) Performance Status sec. ECOG  0 99 (59.3)  1 36 (21.6)  2–3 12 (7.1)  Unknown 20 (12.0) Histology  Adenocarcinoma 147 (88.0)  Squamous 7 (4.2)  Other/Mixed 13 (7.8) T descriptor according to TNM  1–2 102 (61.1)  3–4 65 (39.9) N descriptor according to TNM  0 36 (21.6)  1 19 (11.4)  2 66 (39.5)  3 46 (27.5) M descriptor according to TNM  0 66 (39.5)  1 101 (60.5) Disease stage  I-II 36 (21.6)  III-IV 131 (78.4) Grading  1 12 (7.2)  2 15 (9.0)  3 42 (25.1)  Unknown 98 (58.7) Smoking status  Current 56 (33.5)  Former 68 (40.7)  Never 34 (20.4)  Unknown 9 (5.4) Starting treatment  Surgery 47 (28.1)  Neoadjuvant 3 (1.8)  Chemo-radiotherapy 17 (10.2)  Chemotherapy 75 (44.9)  Support therapy 25 (15.0) Response rate to 1st line therapy  CR 3 (1.8)  PR 31 (18.6)  SD 20 (12.0)  PD 35 (21.0)  NP 78 (46.6) ALK-CNG  disomic 71 (42.5)  3–7 copies 42 (25.1)  >7 copies 15 (9.0)  NA 39 (23.4) CR complete response, PR partial response, SD stable disease, PD progressive disease, NP not performed, CNG copy number gain, NA not available Analysis of ALK gene copy number status In the overall cohort of patients, 128 were available for ALK-CNG analysis. Fifty-seven cases (34.1 %; 95 % CI 34.2-47.7) showed ALK-CNG. We observed three clustered patterns of fluorescent signals (Fig. 1), applying the technique employed by our group for the study of the 3q chromosomal amplification in squamous cell lung carcinoma [29]. Fifteen patients (9.0 %) showed pattern C (>7 ALK fluorescent signals, ranging from 7 to 12) (Fig. 1-a); 42 patients (25.1 %) showed pattern B (from 3 to 7 ALK-CNG signals) (Fig. 1-b) and, as expected, the largest part of the samples (71 patients, 42.5 %) had pattern A (disomic) (Fig. 1-c) (Table 1). The three ALK patterns were interpreted using the control CEP2, CEP3 and CEP17 probes. Among pattern C, the 13 % (2 cases) showed a ratio ≥ 2 when corrected by CEP3 and CEP17. In the remaining 13 out of 15 cases with pattern C and in all the cases with pattern B the FISH results mirrored the mean number of centromeric signals with a final ratio < 2, classifiable as polyploidy. The only 2 cases showing ALK gene amplification were visible at fluorescent microscope as double minutes rather than homogeneously staining regions pattern.Fig. 1 FISH findings in cell neoplastic nuclei. a Nucleus with >7 ALK fluorescent signals; b Nuclei with polysomy of ALK (from 3 to 7 ALK-CNG signals); c Nuclei with a disomic pattern of ALK Association analysis between clinico-pathological characteristics of patients and ALK-CNG No statistically significant correlations were observed between the three ALK-CNG patterns and the majority of the clinico-pathological characteristics (such as gender, histology, grading, comorbidity, cancer familiarity, performance status sec. ECOG) (Table 2). Nevertheless, the association analysis demonstrated that ALK-CNG, pattern B in particular, is more frequently detected in smokers and former smokers compared to never smoker patients (p = 0.03). Similarly, ALK-CNG seems to correlate with the primary tumor extension (T descriptor). Small tumors (T1 and T2) were more frequently detected in association with ALK-CNG pattern B compared to T3 and T4 tumors (43.1 % versus 22.8 %). No significant association has been observed between ALK-CNG and node (N descriptor) or metastatic (M descriptor) status. Although no statistically significant association has been reported between ALK-CNG and disease stage, early stages (I and II) are more represented in ALK pattern B, whereas in pattern C advanced diseases are more frequently detected (1 case stage I versus 14 cases stage III-IV). No well differentiated diseases (G1) have been observed in ALK pattern C subgroup (versus 8 cases G2-3). No difference in term of response to first line chemotherapy has been observed according to the different ALK-CNG patterns.Table 2 Association analysis between clinico-pathological characteristics and ALK-CNG Clinico-pathological parameters ALK-CNG - patients number (%) p-value Disomic 3–7 >7 Total Gender 128 0.59  Male 41 (51.9) 28 (35.4) 10 (12.7) 79 (100.0)  Female 30 (61.2) 14 (28.6) 5 (10.2) 49 (100.0) Cancer familiarity 110 0.44  No 18 (54.5) 13 (39.4) 2 (6.1) 33 (100.0)  Yes 41 (53.2) 25 (32.5) 11 (14.3) 77 (100.0) Smoking status 120 0.03  Never 20 (80.0) 4 (16.0) 1 (4.0) 25 (100.0)  Current 24 (54.5) 13 (29.5) 7 (15.9) 44 (100.0)  Former 22 (43.1) 23 (45.1) 6 (11.8) 51 (100.0) Comorbidity 121 0.19  No 23 (67.6) 8 (23.5) 3 (8.8) 34 (100.0)  Yes 43 (49.4) 33 (37.9) 11 (12.6) 87 (100.0) Performance Status sec. ECOG 112 0.63  0 43 (57.3) 24 (32) 8 (10.7) 75 (100.0)  1 12 (42.9) 11 (39.3) 5 (17.9) 28 (100.0)  2–3 6 (66.7) 2 (22.2) 1 (11.1) 9 (100.0) Histology 128 0.83  Adenocarcinoma 63 (55.8) 38 (33.6) 12 (10.6) 113 (100.0)  Squamous 3 (50.0) 2 (33.3) 1 (16.7) 6 (100.0)  Other 5 (55.6) 2 (22.2) 2 (22.2) 9 (100.0) T descriptor 115 0.06  1–2 28 (48.3) 25 (43.1) 5 (8.6) 58 (100.0)  3–4 36 (63.2) 13 (22.8) 8 (14.0) 57 (100.0) N descriptor 108 0.52  0 10 (45.5) 10 (45.5) 2 (9.1) 22 (100.0)  1–2–3 47 (54.7) 28 (32.6) 11 (12.8) 86 (100.0) M descriptor 128 0.85  0 29 (58.0) 16 (32.0) 5 (10.0) 50 (100.0)  1 42 (53.8) 26 (33.3) 10 (12.8) 78 (100.0) Disease stage 126 0.14  I–II 12 (48.0) 12 (48.0) 1 (4.0) 25 (100.0)  III–IV 57 (56.4) 30 (29.7) 14 (13.9) 101 (100.0) Grading 54 0.62  1 5 (55.6) 4 (44.4) 0 (0) 9 (100.0)  2 4 (33.3) 6 (50.0) 2 (16.7) 12 (100.0)  3 15 (45.3) 12 (36.4) 6 (18.2) 33 (100.0) Response rate to 1st line therapy 73 0.92  No 28 (62.2) 12 (26.7) 5 (11.1) 45 (100.0)  Yes 17 (60.7) 7 (25.0) 4 (14.3) 28 (100.0) CNG copy number gain Prognostic analysis When all the 128 patients evaluable for the ALK analysis were stratified according to ALK-CNG pattern, no significant difference was observed in terms of survival rate at 2 years (Fig. 2). Among resected patients, although no statistically significant difference was observed, a survival advantage for those patients with ALK-CNG pattern B has been reported (2-year OS 69.5 % versus ≤50 % for pattern A and C) (Fig. 3a). Among metastatic patients, the worst survival was observed for the subgroup of patients with ALK-CNG pattern C (2-year OS 0 % versus 27.0 % and 39.1 % for pattern B and A, respectively) (Fig. 3b). The 1-year DFS among resected patients was increased for those patients with ALK-CNG pattern B (80.0 %) compared to those patients with pattern A (39.8 %) and C (26.7 %) (Fig. 3C). The 1-year PFS among advanced patients did not significantly differ according to the ALK-CNG pattern (15.8 % for pattern A, 19.6 % for pattern B and 38.9 % for pattern C) (Fig. 3D). The univariate analysis confirmed the lack of a statistically significant difference in term of survival according to the ALK-CNG pattern (Table 3). At the univariate analysis, disease stage I-II, tumor size 1-2, negative nodes and surgery were significant predictors for longer DFS in resected patients; whereas performance status 0-1 predicts longer PFS and, together with response to first line chemotherapy and lack of synchronous metastases, longer OS in advanced patients (Table 3).Fig. 2 Kaplan Meyer curves for OS in the overall population stratified according to ALK pattern A, B and C Fig. 3 Kaplan Meyer curves stratified according to ALK pattern A, B and C for OS and DFS in resected patients (a-c) and for OS and PFS in metastatic patients (b-d) Table 3 Univariate analysis Variables Disease free survival Cancer specific survival Overall survival HR 95 % CI P HR 95 % CI P HR 95 % CI P Gender [male versus female] 1.213 0.59–2.48 0.596 1.152 0.67–1.96 0.602 1.063 0.70–1.60 0.770 Age [≤66 versus >66 years] 1.709 0.82–3.55 0.151 1.153 0.68–1.94 0.592 1.191 0.79–1.77 0.390 Smoking status [never versus current/former] 1.478 0.60–1.47 0.395 1.782 0.85 - 3.69 0.121 1.332 0.79–2.23 0.278 Comorbidity [no versus yes] 1.395 0.65–2.96 0.388 - - - - - - ALK-CNGa [Disomic/3-7 versus > 7] 1.695 0.49–5.77 0.399 1.498 0.62–3.57 0.363 1.598 0.85–2.98 0.141 Disease stage [I-II versus III] 13.12 4.59–37.4 <0.001 - - - - - - T descriptor [1–2 versus 3–4] 3.920 1.73–8.86 0.001 - - - - - - N descriptor [0 versus 1] 4.259 1.79–10.2 0.001 - - - - - - Grading [1 versus 2 versus 3] 2.128 3.390 0.42–10.6 0.73–15.6 0.357 0.117 - - - - - - Surgery [yes versus no] 8.425 3.65–19.3 <0.001 - - - - - - Performance Status [0 versus 1 versus 2–3] - - - 2.358 4.581 1.25–4.42 1.81–11.5 0.007 0.001 1.647 3.864 1.00–2.70 1.98–7.52 0.049 <0.001 Synchronous Metastases [no versus yes] - - - 1.119 0.15–8.14 0.912 2.335 1.52–3.58 <0.001 Response rate to 1st line [yes versus no] - - - - - - 2.356 1.31–4.23 0.004 HR hazard ratios, CI confidence intervals, CNG copy number gain aThe results of the univariate analysis are not significant even categorizing individually the variables [disomic versus 3–7 versus > 7] Discussion Our analysis aimed to investigate the potential correlation between ALK-CNG and clinical features, exploring the prognostic implications of ALK-CNG in a retrospective cohort of 167 NSCLC patients. We decided to exclude from the analysis patients harboring EGFR mutant and ALK rearranged NSCLC in order to avoid additional confounding factors reducing the reliability of the prognostic analysis. In fact, to evaluate the prognostic effect of ALK and EGFR alterations, patients should not be treated with the selective inhibitors commonly used in clinical practice, because the benefit deriving from these agents might mask the prognostic value of the biomarker in-study. Overall, ALK-CNG was observed in the 34.1 % of cases. The 25.1 % of the samples showed ALK fluorescent signal ranging from 3 to 7 and, for all of them, the ALK-CNG was due to polyploidy (ALK/CEPs ratio < 2), whereas only 9 % of tumors showed ALK fluorescent signal > 7. Therefore, similarly to what demonstrated by our group in the study of the 3q chromosomal amplification in squamous cell lung carcinoma [29], ALK-CNG seems mostly related to polyploidy (whole DNA reduplication) rather than locus specific gene amplification. As reported in other retrospective studies [24], our results suggest a correlation between ALK-CNG and early stage disease (the majority of T1 tumors clusterize in pattern B, none in pattern C). Moreover, smokers and former smokers showed an increased probability to harbor an ALK-CNG compared to never smokers, which are in the 80 % of cases identified in pattern A. These findings support the hypothesis that ALK-CNG might represent a marker of chromosomal instability appearing early in NSCLC carcinogenesis and potentially triggered by cigarette smoking. Although no definitive conclusions can be drawn from the results of our prognostic analysis, some interesting hypotheses emerged regarding the intrinsic biological features of the three ALK-CNG patterns, potentially reflecting different levels of chromosomal instability. In fact, while for ALK-CNG pattern B a greater 1-year DFS (80.0 %) and 2-year OS (69.5 %) clearly emerged in resected tumors, conversely, for pattern C, the DFS rate was very low (26.7 %) and in the metastatic setting none of these patients survives at 2 year time-point. Regarding the PFS, the greater 1-year PFS observed in patients with ALK-CNG pattern C (38.9 % versus 19.6 % for pattern A and 15.8 % for pattern B) might reflect the higher chemo-sensitivity of these tumors featured by a strong chromosomal instability. Although in our work we calculate a 2-year survival rate, future studies should be performed with long-term survival outcomes in order to reliably evaluate the prognostic role of a biomarker for clinical setting. Similarly to our analysis, some other studies evaluating the incidence and potential prognostic implications of ALK gene aberrations have been performed in lung cancer with debatable results. In the context of a retrospective analysis of 107 NSCLC cases, ALK amplification (>5 copies of ALK per cell in 10 % of analyzed cells) and ALK-CNG (mean copy number of 3–5 in 10 % of cells) were identified in the 10 % and 63 % of NSCLC patients, respectively. Although no significant correlation between ALK and clinico-pathological features or prognosis emerged, a significant association between ALK amplification and early stage has been reported, supporting the hypothesis that ALK amplification might represent an early genetic event in NSCLC and potential marker of genomic instability [30]. In a retrospective analysis of 20 pulmonary sarcomatoid carcinomas (PSC), the frequency of ALK-CNG was significantly higher compared to NSCLC with adenocarcinoma histology (22 % versus 0.02 %) with a mean copy number gain of 7 and a significant association with chromosome 7 (EGFR) and 17 (HER2) polysomy. This finding suggests the implication of ALK-CNG as an oncogenic event in PSC, potentially correlated with epithelial-mesenchymal transition and sarcomatoid differentiation [31]. Future confirmation of the oncogenic role of ALK-CNG could support the design of selective studies exploring the potential activity of the ALK inhibition in PSC. Another study conducted in patients with NSCLC and brain metastases reported ALK-CNG in the 11 % of cases with an interesting increased ALK-CNG in brain metastases compared to primary tumors, supporting the rationale of ALK-CNG as a genetic aberration connected to aggressiveness and metastatic behavior [32]. Moreover, in a large retrospective analysis of 1500 NSCLC patients, ALK-CNG (mean native copy number ranged from 2 to 7) was reported in the 80 % of cases and was significantly more common in ALK non-rearranged tumors compared to the translocated ones (62 % versus 19 %). Furthermore, as observed in our analysis, ALK-CNG was mostly related to polysomy, whereas focal amplification was a rare event (<2 % in ALK non-rearranged tumors) [33]. Regarding ALK inhibitors, ALK-CNG, besides representing a marker of insensitivity to crizotinib, was reported to be a mechanism of resistance to crizotinib in ALK-translocated NSCLC both in vitro [34] and in patients progressing during crizotinib treatment [35]. To summarize, our analysis supports the fact that, although ALK pattern B seems to appear early in the tumorigenesis of NSCLC and might mirror an early genomic instability, an high ALK gene copy number (>7, pattern C) related to polyploidy, could be considered as a cut off to discriminate genomically unstable and smoking-related NSCLC featured by an aggressive biological behavior, frequently in advanced stage and with a poor awaited prognostic outcome. Conclusion Although limited by the retrospective nature, the results of our analysis are able to generate interesting hypotheses regarding the biological behavior and the potential therapeutic implication of ALK genetic aberrations, ALK-CNG in particular. As observed in other studies, our analysis confirms that a high ALK gene copy number gain is not a driver genetic event in lung cancer tumorigenesis but it might represent a marker of chromosome instability, correlated with an aggressive metastatic behavior. In this regard, early pathogenic events probably induce different ALK aberrant NSCLC. The ALK non-translocated tumors are frequently associated with chromosomal instability and ALK-CNG, whereas the ALK-translocated NSCLC presents a low native ALK copy number and the increase in ALK-CNG emerges as a mechanism of resistance to crizotinib treatment. Unfortunately, conversely to ALK and ROS1 translocations that are widely recognized oncogenic drivers in a small subset of NSCLC able to predict sensitivity to specific inhibitors, ALK-CNG represents a candidate mechanism of resistance to this target therapy. Therefore, it is mandatory to clarify the role and mechanism of this genetic aberration in order to identify and validate effective therapeutic approaches in this subpopulation of lung cancer patients harboring ALK-CNG. In this context, our analysis is conceived and designed as a further step towards a more biologically rationale approach to treat NSCLC patients. The implications for prognosis should be prospectively investigated and validated in larger patients’ series. Abbreviations CEPCentromeric alpha-satellite specific for chromosome CNGCopy number gain CRComplete response DFSDisease free survival ERMSEmbryonal RMS FISHFluorescence in situ hybridization IBCInflammatory breast cancer NBLNeuroblastoma NSCLCNon-small cell lung cancer OSOverall survival PDProgressive disease PFSProgression free survival PRPartial response PSCPulmonary sarcomatoid carcinomas RFSRecurrence free survival RMSRhabdomyosarcoma SDStable disease Acknowledgements Not applicable. Funding This work was supported by a specific grant of the Italian Association for Cancer Research (AIRC, My First AIRC Grant n° 14282) and by a Young Investigational Award of the International Association for the Study of Lung Cancer (IASLC). The reported sources of funding allowed the collection, analysis, and interpretation of data. Availability of data and materials The Cytogenetics Laboratory of Verona’s University provided the instruments for FISH analysis described in the method sections. The Lung Verona Database (Verona, Italy) was analyzed to obtain the clinical and pathological data. Authors’ contributions All the authors equally contributed to the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate We confirm that all the experiments were performed in accordance with relevant guidelines and regulations and that all the experimental protocols were approved by the local Ethics Committee of the Azienda Ospedaliera Universitaria Integrata of Verona (protocol number 27 of the 19/11/2014). The informed consent was obtained from all the subjects included in the study. ==== Refs References 1. Maemondo M Inoue A Kobayashi K Gefitinib or chemotherapy for non-small-cell lung cancer with mutated EGFR N Engl J Med 2010 362 2380 2388 10.1056/NEJMoa0909530 20573926 2. Mitsudomi T Morita S Yatabe Y Gefitinib versus cisplatin plus docetaxel in patients with non-small-cell lung cancer harbouring mutations of the epidermal growth factor receptor (WJTOG3405): an open label, randomised phase 3 trial Lancet Oncol 2010 11 121 128 10.1016/S1470-2045(09)70364-X 20022809 3. Mok TS Wu YL Thongprasert S Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma N Engl J Med 2009 361 947 957 10.1056/NEJMoa0810699 19692680 4. Rosell R Carcereny E Gervais R Erlotinib versus standard chemotherapy as first-line treatment for European patients with advanced EGFR mutation-positive non-small-cell lung cancer (EURTAC): a multicentre, open-label, randomised phase 3 trial Lancet Oncol 2012 13 239 246 10.1016/S1470-2045(11)70393-X 22285168 5. Sequist LV Yang JC Yamamoto N Phase III study of afatinib or cisplatin plus pemetrexed in patients with metastatic lung adenocarcinoma with EGFR mutations J Clin Oncol 2013 31 3327 3334 10.1200/JCO.2012.44.2806 23816960 6. Shaw AT Kim DW Nakagawa K Crizotinib versus chemotherapy in advanced ALK-positive lung cancer N Engl J Med 2013 368 2385 2394 10.1056/NEJMoa1214886 23724913 7. Solomon BJ Mok T Kim DW First-line crizotinib versus chemotherapy in ALK-positive lung cancer N Engl J Med 2014 371 2167 2177 10.1056/NEJMoa1408440 25470694 8. Pilotto S Peretti U Novello S PROFILing non-small-cell lung cancer patients for treatment with crizotinib according to anaplastic lymphoma kinase abnormalities: translating science into medicine Expert Opin Pharmacother 2013 14 597 608 10.1517/14656566.2013.778828 23472711 9. Christensen JG Zou HY Arango ME Cytoreductive antitumor activity of PF-2341066, a novel inhibitor of anaplastic lymphoma kinase and c-Met, in experimental models of anaplastic large-cell lymphoma Mol Cancer Ther 2007 6 3314 3322 10.1158/1535-7163.MCT-07-0365 18089725 10. Camidge DR Bang YJ Kwak EL Activity and safety of crizotinib in patients with ALK-positive non-small-cell lung cancer: updated results from a phase 1 study Lancet Oncol 2012 13 1011 1019 10.1016/S1470-2045(12)70344-3 22954507 11. Crino L Kim D Riely GJ Initial phase II results with crizotinib in advanced ALK-positive non-small cell lung cancer (NSCLC): PROFILE 1005 ASCO Meeting Abstracts 2011 29 7514 12. Felip E Orlov S Park K ASCEND-3: A single-arm, open-label, multicenter phase II study of ceritinib in ALKi-naive adult patients (pts) with ALK-rearranged (ALK+) non-small cell lung cancer (NSCLC) ASCO Meeting Abstracts 2015 33 8060 13. Kim DW Mehra R Tan DS Activity and safety of ceritinib in patients with ALK-rearranged non-small-cell lung cancer (ASCEND-1): updated results from the multicentre, open-label, phase 1 trial Lancet Oncol 2016 17 452 463 10.1016/S1470-2045(15)00614-2 26973324 14. Mok T Spigel D Felip E ASCEND-2: A single-arm, open-label, multicenter phase II study of ceritinib in adult patients (pts) with ALK-rearranged (ALK+) non-small cell lung cancer (NSCLC) previously treated with chemotherapy and crizotinib (CRZ) ASCO Meeting Abstracts 2015 33 8059 15. Ou SH Ahn JS De Petris L Alectinib in Crizotinib-Refractory ALK-Rearranged Non-Small-Cell Lung Cancer: A Phase II Global Study J Clin Oncol 2016 34 661 668 10.1200/JCO.2015.63.9443 26598747 16. Shaw AT Gandhi L Gadgeel S Alectinib in ALK-positive, crizotinib-resistant, non-small-cell lung cancer: a single-group, multicentre, phase 2 trial Lancet Oncol 2016 17 234 242 10.1016/S1470-2045(15)00488-X 26708155 17. Kim D-W Tiseo M Ahn M-J Brigatinib (BRG) in patients (pts) with crizotinib (CRZ)-refractory ALK+ non-small cell lung cancer (NSCLC): First report of efficacy and safety from a pivotal randomized phase (ph) 2 trial (ALTA) ASCO Meeting Abstracts 2016 34 9007 18. Solomon BJ Bauer TM Felip E Safety and efficacy of lorlatinib (PF-06463922) from the dose-escalation component of a study in patients with advanced ALK+ or ROS1+ non-small cell lung cancer (NSCLC) ASCO Meeting Abstracts 2016 34 9009 19. Nokihara H Hida T Kondo M Alectinib (ALC) versus crizotinib (CRZ) in ALK-inhibitor naive ALK-positive non-small cell lung cancer (ALK+ NSCLC): Primary results from the J-ALEX study ASCO Meeting Abstracts 2016 34 9008 20. Sukov WR Hodge JC Lohse CM ALK alterations in adult renal cell carcinoma: frequency, clinicopathologic features and outcome in a large series of consecutively treated patients Modern Pathol 2012 25 1516 1525 10.1038/modpathol.2012.107 21. Bavi P Jehan Z Bu R ALK gene amplification is associated with poor prognosis in colorectal carcinoma Br J Cancer 2013 109 2735 2743 10.1038/bjc.2013.641 24129244 22. Pietrantonio F Maggi C Di Bartolomeo M Gain of ALK gene copy number may predict lack of benefit from anti-EGFR treatment in patients with advanced colorectal cancer and RAS-RAF-PI3KCA wild-type status PLoS One 2014 9 e92147 10.1371/journal.pone.0092147 24691006 23. Jia SW Fu S Wang F Shao Q Huang HB Shao JY ALK gene copy number gain and its clinical significance in hepatocellular carcinoma World J Gastroenterol 2014 20 183 192 10.3748/wjg.v20.i1.183 24415871 24. van Gaal JC Flucke UE Roeffen MH Anaplastic lymphoma kinase aberrations in rhabdomyosarcoma: clinical and prognostic implications J Clin Oncol 2012 30 308 315 10.1200/JCO.2011.37.8588 22184391 25. Kim MH Lee S Koo JS Anaplastic lymphoma kinase gene copy number gain in inflammatory breast cancer (IBC): prevalence, clinicopathologic features and prognostic implication PLoS One 2015 10 e0120320 10.1371/journal.pone.0120320 25803816 26. Chen Y Takita J Choi YL Oncogenic mutations of ALK kinase in neuroblastoma Nature 2008 455 971 974 10.1038/nature07399 18923524 27. Wang M Zhou C Sun Q ALK amplification and protein expression predict inferior prognosis in neuroblastomas Exp Mol Pathol 2013 95 124 130 10.1016/j.yexmp.2013.06.002 23797004 28. Travis WD Muller-Hermelink H-K Harris CC Pathology and Genetics of Tumours of the Lung, Pleura, Thymus and Heart 2004 Lyon IARC Press 29. Brunelli M Bria E Nottegar A True 3q chromosomal amplification in squamous cell lung carcinoma by FISH and aCGH molecular analysis: impact on targeted drugs PLoS One 2012 7 e49689 10.1371/journal.pone.0049689 23236352 30. Salido M Pijuan L Martinez-Aviles L Increased ALK gene copy number and amplification are frequent in non-small cell lung cancer J Thorac Oncol 2011 6 21 27 10.1097/JTO.0b013e3181fb7cd6 21107285 31. Pelosi G Gasparini P Cavazza A Multiparametric molecular characterization of pulmonary sarcomatoid carcinoma reveals a nonrandom amplification of anaplastic lymphoma kinase (ALK) gene Lung Cancer 2012 77 507 514 10.1016/j.lungcan.2012.05.093 22705117 32. Preusser M Berghoff AS Ilhan-Mutlu A ALK gene translocations and amplifications in brain metastases of non-small cell lung cancer Lung Cancer 2013 80 278 283 10.1016/j.lungcan.2013.01.019 23453647 33. Camidge DR Skokan M Kiatsimkul P Native and rearranged ALK copy number and rearranged cell count in non-small cell lung cancer: implications for ALK inhibitor therapy Cancer 2013 119 3968 3975 10.1002/cncr.28311 24022839 34. Katayama R Khan TM Benes C Therapeutic strategies to overcome crizotinib resistance in non-small cell lung cancers harboring the fusion oncogene EML4-ALK Proc Natl Acad Sci U S A 2011 108 7535 7540 10.1073/pnas.1019559108 21502504 35. Katayama R Shaw AT Khan TM Mechanisms of acquired crizotinib resistance in ALK-rearranged lung Cancers Sci Transl Med 2012 4 120ra117 10.1126/scitranslmed.3003316

PMC005xxxxxx/PMC5000439.txt

==== Front J Med Case RepJ Med Case RepJournal of Medical Case Reports1752-1947BioMed Central London 102010.1186/s13256-016-1020-6Case ReportAbnormal white matter tracts resembling pencil fibers involving prefrontal cortex (Brodmann area 47) in autism: a case report Hashemi Ezzat ehashemi@ucdavis.edu 12Ariza Jeanelle jariza@ucdavis.edu 12Lechpammer Mirna mlechpammer@ucdavis.edu 1Noctor Stephen C. scnoctor@ucdavis.edu 34Martínez-Cerdeño Verónica (916) 453-2163vmartinezcerdeno@ucdavis.edu 1241 Department of Pathology and Laboratory Medicine, UC Davis, Davis, USA 2 Institute for Pediatric Regenerative Medicine and Shriners Hospitals for Children Northern California, 2425 Stockton Boulevard, Sacramento, California 95817 USA 3 Department of Psychiatry and Behavioral Sciences, UC Davis, Davis, USA 4 MIND Institute, UC Davis School of Medicine, Davis, USA 26 8 2016 26 8 2016 2016 10 1 23729 4 2016 1 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Autism is not correlated with any neuropathological hallmark as the brain of autistic individuals lack defined lesions. However, previous investigations have reported cortical heterotopias and local distortion of the cytoarchitecture of the neocortex in some cases of autism. Case presentation Our patient was a 40-year-old white woman diagnosed at an early age with autism and mental retardation. Pencil fibers were present within the prefrontal cortex (Brodmann area 47) and its composition resembled that of the underlying white matter region. Pencil fibers encompassed most of the extent of the cortical grey matter and were populated by oligodendrocytes, astrocytes, and microglial cells, but not by neurons. Conclusions Here we report a new cytoarchitectural abnormality that has not been previously described in autism. Future pathological examinations should keep in mind the potential presence of pencil fibers within the prefrontal cortex of cases with autism. Keywords AutismPathologyPencil fibersCerebral cortexPrefrontalHumanCase reporthttp://dx.doi.org/10.13039/100000025National Institute of Mental HealthMH094681MH101188Noctor Stephen C. Martínez-Cerdeño Verónica Shriners Hospitals issue-copyright-statement© The Author(s) 2016 ==== Body Background Autism spectrum disorders (ASD) are defined by a pattern of qualitative abnormalities in reciprocal social interaction, communication, and repetitive interest and behavior. Altered functioning in several areas of the brain underlies the social and cognitive phenotype in autism. Regions of the brain in which alterations have been identified include the prefrontal cerebral cortex. Autism is accompanied by altered patterns of connectivity in the adult brain and impairments in brain development including cell generation and migration during prenatal cortical development [1–3]. However, the classical symptoms of autism are not correlated with characteristic neuropathological hallmarks as the brain of autistic individuals lack defined lesions. Previous investigations have reported subcortical, periventricular, hippocampal, and cerebellar heterotopias detected in the brains of 30 % of autistic individuals [4]. Multifocal cerebral dysplasia resulted in local distortion of the cytoarchitecture of the neocortex in four brains (31 % of the examined brains), the entorhinal cortex in two brains (15 %), the hippocampus in four brains (31 %), and the dentate gyrus in two brains (15 %) [4]. These alterations included focal patches of abnormal laminar cytoarchitecture and disorganization of cortical neurons in prefrontal and temporal cortical tissue in children with autism [5]. Here, we show for the first time to the best of our knowledge, the presence of a white matter abnormality resembling pencil fibers in Brodmann area (BA) 47 in a case of autism. Case presentation We report the case of a 40-year-old white woman diagnosed with autism whose clinical features included qualitative abnormalities in reciprocal social interaction and communication, and restricted repetitive and stereotyped patterns of behavior together with mental retardation, as is common in autism [6]. Her clinical history (AN07770) was collected by the Autism Tissue Program (ATP) through a parental interview using the Autism Diagnostic Interview - Revised (ADI-R). Later the ATP collected her brain. The ATP has transitioned to a new autism brain network called Autism BrainNet. Summarizing, our patient was a full term baby who weighed 2.95 kg (6 lb 8 oz). Her physical development is described as being normal. She was diagnosed at an early age with mental retardation. Her language was significantly delayed. No signs of convulsions, vision or hearing problems, or neurological abnormalities were diagnosed. She demonstrated apparent adherence to non-functional routines and several sensory aversions to loud noises. She was medicated with Vistaril (hydroxyzine), Decadron (dexamethasone), and Demulen (ethinyl estradiol). Her maternal family history is notable in that her grandmother had emphysema. Her paternal family history is significant in that her father died at age 53 due to atherosclerosis after undergoing two bypass surgeries. Her magnetic resonance imaging (MRI) demonstrated moderate cerebral and cerebellar atrophic changes, and extensive low intensity in her substantia nigra and basal ganglia bilaterally, probably secondary to iron deposition. At age 38 she experienced mood swings, was leaning to the left and walking on her right toe with her right arm curled in. By the end of that year she was unable to walk and was confined to a wheelchair. At age 39 years she was diagnosed with pantothenate kinase-associated neurodegeneration (PKAN), because of the spasticity in her limbs and feet. She experienced rapid and severe regression of her motor skills and ability to speak and died at age 40 due to respiratory arrest. No pathology report was obtained. Her brain was very small (890 grams). We analyzed prefrontal tissue from areas BA45, 46, 47, and 9. No obvious modification in cell size or density was observed. Of these areas, only BA47 presented with abnormal white matter islands extending into the cortical grey matter. The rest of the brain tissue from this patient was distributed to other research groups. We are not aware of any additional report of pathology in this case. Publications reporting on this case included that by McKavanagh et al. that reported on the temporal cortex (BA40, 41, and planum temporale) and the orbitofrontal cortex (BA11) [7]. Our patient presented with abnormal white matter extensions and islands into the cortical grey matter in BA47. These white matter regions resembled the pencil fibers of the striatum [8] and therefore we referred to them as “pencil fibers.” Cortical pencil fibers have never been described in autism. Pencil fibers encompassed most of the extent of the cortical grey matter, some including all layers from VI to II, but did not extend into layer I (Fig. 1). We performed 14 μm coronal sections of BA47 area and immunostained it with specific cell type markers. We found that oligodendrocytes (SRY (sex determining region Y)-box 10 and oligodendrocyte transcription factor 2; Sox10+ and Olig2+), astrocytes (S100+ and glial fibrillary acidic protein; GFAP+), and microglial cells (ionized calcium-binding adapter molecule 1; Iba1+) were present within the pencil fibers. Some of the microglia were immunopositive for CD68, indicating that they were activated cells (Fig. 2). However, neurons were not present. Pencil fibers  were rich in axonal neurofilament 312 (SMI312+) and neurofilament H non-phosphorylated 132 (SMI32+) fibers. Overall, the cellular composition within the cortical pencil fibers resembled that of the underlying white matter region. We found b-amyloid deposits and tau+ neurofilaments, which are typical of neurodegenerative diseases including PKAN, within the cortical grey matter but not within the white matter pencil fibers. Iron deposits were not detected.Fig. 1 a A block of cerebral cortex tissue isolated from Brodmann area 47. The arrows point to one of the pencil fibers that prominently extend into the cortical grey matter. The inset labeled ‘D’ is shown at higher power in panel (d). b Same image shown in panel (a) in which the outline of the pencil fibers is delineated with a black line. c A section of tissue obtained from the block stained for Nissl substance. Arrows in panel c point to the location of the pencil fiber that is also indicated with arrows in panel a. The inset labeled ‘D’ is shown at higher power in panel (d). d Higher power magnification of the inset shown in panels (a) and (c). This image shows the histology within the pencil fibers. Glial nuclei are observable in the Nissl, but not neurons. Scale bars: a–c 600 μm; d 150 μm Fig. 2 Immunostaining with cell-specific markers in tissue sections that were adjacent to the section shown in Panel 1c. The images were taken within the pencil fibers indicated by the insets in Panels 1a and 1c, and shown in Nissl staining at higher power in Panel 1d. a–f Markers for glial cells show the presence of glial cells within the cortical pencil fibers: Sox 10 and Olig2 for oligodendrocytes, S100 and GFAP for astrocytes, Iba 1 for microglial cells and CD68 for activated microglia. g, h Axonal neurofilament 312 and neurofilament H non-phosphorylated 132 demonstrate the presence of axonal fibers within the pencil fibers. Scale bar: 150 μm Discussion The formation of cortical pencil fibers in this patient most likely arose from abnormal cortical development rather than from a neurodegenerative process. We did not detect any atypical microscopic feature common to neurodegenerative disorders within the pencil fibers such as neuritic plaques and neurofibrillary tangles, dysplastic neurons and/or balloon cells. Therefore, we conclude that pencil fibers in this patient were most likely the product of altered prenatal cortical development linked to autism, rather than the later manifested pathological consequence of PKAN. The presence of pencil fibers within the grey matter of the cerebral cortex disrupts cortical cytoarchitecture, occupying a position in the cortex that would otherwise be populated by excitatory and inhibitory neurons and glial cells. The abnormal laminar cytoarchitecture and cortical pencil fibers could represent a form of cortical dysplasia resulting from defects in cellular migration during prenatal cortical development. Defects in the radial migration of excitatory cortical neurons can produce defects as observed here. In addition, defects in the tangential migration of inhibitory cortical interneurons could also impair the regional position of these cells and disrupt proper formation of the cortex. Dysregulated patterns of progenitor cell division (radial glial cells and/or intermediate progenitor cells) could also modify the final laminar destination of neurons [9–11]. Alternatively, it is possible that severely impaired axonal guidance disrupted white matter development in this cortical area. Accordingly, 88 % of high-risk genes for autism have been found to influence neural induction and early maturation of newly born cells [2, 12]. Conclusions Overall, we show a new cytoarchitectural abnormality, cortical pencil fibers, not previously described in the cerebral cortex of an autistic individual. Future pathological examinations should keep in mind the potential presence of pencil fibers within the prefrontal cortex of cases with autism. Abbreviations ADI-R, Autism Diagnostic Interview - Revised; ASD, autism spectrum disorders; ATP, Autism Tissue Program; BA, Brodmann area; GFAP, glial fibrillary acidic protein; Iba1, ionized calcium-binding adapter molecule 1; MRI, magnetic resonance imaging; Olig2, oligodendrocyte transcription factor 2; PKAN, pantothenate kinase-associated neurodegeneration; S100, protein of low molecular weight characterized by two calcium-binding sites that have helix-loop-helix (“EF-hand type”) conformation; SMI312, axonal neurofilament 312; SMI32, neurofilament H non-phosphorylated 132; Sox10, SRY (sex determining region Y)-box 10 Acknowledgements Not applicable. Funding This study was funded by the National Institutes of Health (MH094681 V. Martínez-Cerdeño and MH101188 Stephen Noctor) and by Shriners Hospitals (V. Martínez-Cerdeño). The funding bodies did not participate in the design of the study, or in the collection, analysis, and interpretation of data, or in writing the manuscript. Availability of data and materials Tissue was obtained from Autism BrainNet that is sponsored by the Simons Foundation and Autism Speaks. The Autism Tissue Program was the predecessor to Autism BrainNet. The data supporting the clinical aspects of this article are available at the Autism BrainNet repository, https://takesbrains.org Authors’ contributions EH found the cytoarchitecture abnormality described here. JA performed experimental work. SCN and ML contributed to the writing of the manuscript. VMC designed and supervised the project and wrote the manuscript. All authors read and approved the final manuscript. Competing interests The authors declare that they have no competing interests. Consent for publication Written informed consent was obtained from the patient’s next-of-kin for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal. ==== Refs References 1. Dajani DR Uddin LQ Local brain connectivity across development in autism spectrum disorder: A cross-sectional investigation Autism Res 2016 9 1 43 54 10.1002/aur.1494 26058882 2. Casanova EL Casanova MF Genetics studies indicate that neural induction and early neuronal maturation are disturbed in autism Front Cell Neurosci 2014 8 397 25477785 3. Martinez-Cerdeno V Camacho J Fox E Miller E Ariza J Prenatal Exposure to Autism-Specific Maternal Autoantibodies Alters Proliferation of Cortical Neural Precursor Cells, Enlarges Brain, and Increases Neuronal Size in Adult Animals Cereb Cortex 2016 26 1 374 83 10.1093/cercor/bhu291 25535268 4. Wegiel J Kuchna I Nowicki K Imaki H Wegiel J The neuropathology of autism: defects of neurogenesis and neuronal migration, and dysplastic changes Acta Neuropathol 2010 119 755 70 10.1007/s00401-010-0655-4 20198484 5. Stoner R Chow ML Boyle MP Sunkin SM Mouton PR Patches of disorganization in the neocortex of children with autism N Engl J Med 2014 370 1209 19 10.1056/NEJMoa1307491 24670167 6. Constantino JN Charman T Diagnosis of autism spectrum disorder: reconciling the syndrome, its diverse origins, and variation in expression Lancet Neurol 2016 15 3 279 91 10.1016/S1474-4422(15)00151-9 26497771 7. McKavanagh R Buckley E Chance SA Wider minicolumns in autism: a neural basis for altered processing? Brain 2015 138 2034 45 10.1093/brain/awv110 25935724 8. Wilson SAK An experimental research into the anatomy and physiology of the corpus striatum Brain 1914 36 427 92 10.1093/brain/36.3-4.427 9. Cunningham CL Martinez Cerdeno V Navarro Porras E Prakash AN Angelastro JM Premutation CGG-repeat expansion of the Fmr1 gene impairs mouse neocortical development Hum Mol Genet 2011 20 64 79 10.1093/hmg/ddq432 20935171 10. Noctor SC Martinez-Cerdeno V Ivic L Kriegstein AR Cortical neurons arise in symmetric and asymmetric division zones and migrate through specific phases Nat Neurosci 2004 7 136 44 10.1038/nn1172 14703572 11. Kriegstein AR Noctor SC Patterns of neuronal migration in the embryonic cortex Trends Neurosci 2004 27 392 9 10.1016/j.tins.2004.05.001 15219738 12. Maximo JO Cadena EJ Kana RK The implications of brain connectivity in the neuropsychology of autism Neuropsychol Rev 2014 24 16 31 10.1007/s11065-014-9250-0 24496901

PMC005xxxxxx/PMC5000440.txt

==== Front Acta Neuropathol CommunActa Neuropathol CommunActa Neuropathologica Communications2051-5960BioMed Central London 36010.1186/s40478-016-0360-1ErratumErratum to: Notch1 hallmarks fibrillary depositions in sporadic Alzheimer’s disease Brai E. 1Raio N. Alina 1Alberi L. lavinia.alberi@unifr.ch 1231 Unit of Anatomy, Department of Medicine, University of Fribourg, Route de Gockel, 1, Fribourg, 1700 Switzerland 2 Unit of Pathology, Department of Medicine, University of Fribourg, Route de Gockel, 1, Fribourg, 1700 Switzerland 3 Swiss Integrative Center for Human Health, Passage du Cardinal, 13B, Fribourg, 1700 Switzerland 25 8 2016 25 8 2016 2016 4 1 9013 7 2016 1 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.issue-copyright-statement© The Author(s) 2016 ==== Body Erratum The authors of the original article [1] would like to make the readers aware that the acknowledgements section of this article was incomplete. The updated text should read as follows: We are particularly grateful to Prof. Francis (King’s College London, UK) for his valuable feedback on our submission for human samples to the Brain Bank for Dementia, UK. We would like to gratefully acknowledge all donors and their families for the tissue provided for this study. Human tissue samples were supplied by the Brains for Dementia Research programme, jointly funded by Alzheimer’s Research UK, the Alzheimer’s Society and the Medical Research Council, and sourced from the Oxford Brain Bank. The Oxford Brain Bank is also supported by the National Institute for Health Research (NIHR) Units. We are sorry for any inconvenience caused. The online version of the original article can be found under doi:10.1186/s40478-016-0327-2. ==== Refs Reference 1. Brai E, et al. Notch1 hallmarks fibrillary depositions in sporadic Alzheimer’s disease. Acta Neuropathol Commun. 2016;4(1):64.

PMC005xxxxxx/PMC5000441.txt

==== Front Front ZoolFront. ZoolFrontiers in Zoology1742-9994BioMed Central London 16910.1186/s12983-016-0169-6ResearchContextual flexibility in the vocal repertoire of an Amazon parrot Montes-Medina Adolfo Christian 1Salinas-Melgoza Alejandro 2Renton Katherine krenton@st.ib.unam.mx 31 Posgrado en Ciencias Biológicas, Instituto de Biología, Universidad Nacional Autónoma de México, Ciudad Universitaria, Mexico City, Mexico 2 Facultad de Biología, Universidad Michoacana de San Nicolás de Hidalgo, Ciudad Universitaria, Morelia, Michoacán Mexico 3 Estación de Biología Chamela, Instituto de Biología, Universidad Nacional Autónoma de México, Apartado Postal 21, San Patricio-Melaque, Chamela, Jalisco CP 48980 Mexico 26 8 2016 26 8 2016 2016 13 1 4029 3 2016 8 8 2016 © The Author(s). 2016 Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Background Understanding the role of avian vocal communication in social organisation requires knowledge of the vocal repertoire used to convey information. Parrots use acoustic signals in a variety of social contexts, but no studies have evaluated cross-functional use of acoustic signals by parrots, or whether these conform to signal design rules for different behavioural contexts. We statistically characterised the vocal repertoire of 61 free-living Lilac-crowned Amazons (Amazona finschi) in nine behavioural contexts (nesting, threat, alarm, foraging, perched, take-off, flight, landing, and food soliciting). We aimed to determine whether parrots demonstrated contextual flexibility in their vocal repertoire, and whether these acoustic signals follow design rules that could maximise communication. Results The Lilac-crowned Amazon had a diverse vocal repertoire of 101 note-types emitted at least twice, 58 of which were emitted ≥5 times. Threat and nesting contexts had the greatest variety and proportion of exclusive note-types, although the most common note-types were emitted in all behavioural contexts but with differing proportional contribution. Behavioural context significantly explained variation in acoustic features, where threat and nesting contexts had the highest mean frequencies and broad bandwidths, and alarm signals had a high emission rate of 3.6 notes/s. Three Principal Components explained 72.03 % of the variation in temporal and spectral characteristics of notes. Permutated Discriminant Function Analysis using these Principal Components demonstrated that 28 note-types (emitted by >1 individual) could be correctly classified and significantly discriminated from a random model. Conclusions Acoustic features of Lilac-crowned Amazon vocalisations in specific behavioural contexts conformed to signal design rules. Lilac-crowned Amazons modified the emission rate and proportional contribution of note-types used in each context, suggesting the use of graded and combinatorial variation to encode information. We propose that evaluation of vocal repertoires based on note-types would reflect the true extent of a species’ vocal flexibility, and the potential for combinatorial structures in parrot acoustic signals. Electronic supplementary material The online version of this article (doi:10.1186/s12983-016-0169-6) contains supplementary material, which is available to authorized users. Keywords Animal communicationLilac-crowned AmazonPsittaciformesSignal design rulesTropical dry foresthttp://dx.doi.org/10.13039/501100003141Consejo Nacional de Ciencia y Tecnología23168598294Montes-Medina Adolfo Christian Salinas-Melgoza Alejandro issue-copyright-statement© The Author(s) 2016 ==== Body Background Knowledge of the vocal repertoire of avian species and the association with behaviour enables further understanding of the function and complexity of vocal communication [1–3]. However, the majority of studies on avian vocal communication have been conducted on passerines, with few studies on non-passerines [4–6], effectively narrowing our understanding of the array of signal design strategies for communication in the animal kingdom. Psittaciformes (parrots) are an interesting model for evaluating the behavioural context of the vocal repertoire as parrots have complex social systems that require a similarly complex communication system, and use acoustic communication in a variety of contexts [7], as well as being vocal learners able to acquire acoustic signals through social interaction [8]. Furthermore, parrots use their tongue to modulate sound independent of the source, analogous to that of humans, indicating a speech-like system in the emission of parrot vocalisations [9, 10]. Early studies of psittacine vocal repertoires classified vocalisations by onomatopoeic sound and visual representation in spectrograms [11–14]. Later studies used parametric description of spectrogram features to categorise vocalisations but lacked statistical quantification to objectively differentiate acoustic signals [15–19]. Some studies have attempted to statistically differentiate vocalisations by their temporal or acoustic properties. Univariate analyses of vocalisations found that acoustic signals used by the Blue-crowned Conure (Aratinga acuticaudata) varied significantly in emission rate of notes per second, particularly for alarm signals [20]. Long-range alarm calls of the Yellow-faced Parrot (Alipiopsitta xanthops) also had a significantly higher emission rate that flight calls, and greater amplitude than sentinel calls [21]. However, few studies have conducted comprehensive statistical analysis of a suite of acoustic traits to reliably differentiate vocalisations. Of these, guttural calls differed from other vocalisations of the Blue-fronted Amazon (Amazona aestiva) in note duration and bandwidth [22], while five call-types of the North Island Kaka (Nestor meridionalis septentrionalis) were differentiated primarily by call length and secondly by maximum frequency [23]. In general, parrots have been found to present short- and long-range vocalisations [19, 21, 24] of notes with 0–6, and up to ten, harmonics [15–17, 20, 24–26]. Studies of vocal communication of Psittaciformes report from five to 15 calls that can be classified in discrete spectrographic or structural categories [12, 13, 17, 19–21, 23–28], where some vocalisations are given in a variety of contexts, but other vocalisations may be specific to a given behavioural context [19–21, 23–25, 27, 29]. In particular, van Horik et al. [23] determined a significant association of 5 calls of the North Island Kaka with three behavioural contexts of paired, perched, and flying, and Zdenek et al. [25] found a significant association of vocal syllables of the Palm Cockatoo (Probosciger aterrimus) with five behavioural contexts. Selection forces may drive the form or characteristics of vocal signals in accordance with signal design rules to attain optimal communication in a given behavioural context [3]. Design rules state that signals used in differing behavioural contexts should present features of range, locatability, duty cycle (duration and repetition rate), sender identification, within-individual variation, and form-content linkage that optimise coding of the information to be conveyed [3]. The importance of vocal communication in psittacine behaviour and social organisation is reflected by the fact that parrots use vocal signals in a variety of behavioural contexts, yet to date no studies have conducted across-function comparisons of parrot acoustic signals in differing social contexts to identify the combination of design features that could optimise communication, and whether these conform to signal design rules. Parrots use vocalisations of contrasting characteristics, with long-range acoustic signals where energy is concentrated at low frequencies, and short-range signals of high frequencies [22, 24, 27], suggesting that some design rules may be at play. Behavioural studies of the Lilac-crowned Amazon (Amazona finschi) indicate that vocal signals are used to coordinate nesting behaviour by the reproductive pair [30]. However, the characteristics of the vocal repertoire, context specificity of vocalisations, and how these conform to signal design rules, are still unknown. Therefore, in the present study we statistically characterised the vocal repertoire of free-living Lilac-crowned Amazons in distinct behavioural contexts. We aimed to determine whether parrots demonstrated contextual flexibility in their vocal repertoire, and whether these acoustic signals follow design rules that could maximise communication. In accordance with signal design rules [3], we hypothesised that alarm vocalisations would be context-specific, having acoustic characteristics of either flee (low, short, single vocalisation) or assembly (loud, broad, repeated) signals. On the other hand, aggressive threat signals should be more complex being either loud or soft, involve counter-calling, and have characteristics to encode information on body size or motivation. Similarly, nesting vocalisations should be directed at a specific receiver or nest-site, have a high duration or repetition rate, with a diverse repertoire, where both the male and female participate. Finally, we expected vocalisations emitted when perched or foraging to be of short-range, with low diversity and repetition rate, so as to maintain contact with conspecifics but avoid detection by potential predators. Methods We recorded vocalisations of free-living Lilac-crowned Amazons in the tropical dry forest of the Chamela-Cuixmala Biosphere Reserve (19o22′N 104o56′W to 19o35′N 105o03′W), on the coast of Jalisco, Mexico. The region has a marked seasonality in rainfall and plant phenology, with precipitation concentrated in 5 months (June to October), and a prolonged dry season [31, 32]. The main vegetation types within the reserve are dense deciduous forest on the hills and slopes, and taller semi-deciduous forest in valleys [33]. Deciduous forest has small trees with a canopy height of 8–12 m, where the majority of plants lose leaf-cover for 5–8 months of the year, whereas semi-deciduous forest has larger trees of 15–30 m height, most of which retain their leaves or drop leaves for 1–3 months of the year [33, 34]. The Lilac-crowned Amazon is endemic to the Pacific coast of Mexico, and nests during the dry season from February to May [30]. Research permits for the study were granted by the Secretaria del Medio Ambiente y Recursos Naturales, Mexico. Vocal recording We recorded vocalisations emitted by parrots at nest-sites, and during opportunistic encounters while they were foraging and resting. Recordings of 61 individuals were made during the morning (07:30–11:00 h) and afternoon (17:00–19:00 h) when parrots are most active [35]. All recordings were made at about 30 m from focal individuals. Parrot vocalisations were recorded with a Marantz PMD 660 or Marantz PMD 670 solid state digital recorders, and a directional ME66/k6 microphone (Sennheiser Electronic) on a shock-mount pistol-grip. Recordings were saved on secure digital or compact flash cards as 16-bitwav files, with a sampling rate of 44.1 kHz or 48 kHz. We then resampled the 48 kHz recordings to standardise them to 44.1 kHz in Goldwave 5.57 (GoldWave Inc.). Recordings were viewed via spectrogram in Raven Pro 1.4 (Cornell Laboratory of Ornithology, New York) with a Hann window size of 592 samples, a 3 dB filter bandwidth of 107 Hz, a frequency grid with discrete Fourier transform size of 1024 samples and grid spacing of 43.1 Hz, and a time grid with a hop size of 59 samples and 90 % overlap, averaging 1 spectra. Vocal analyses We reviewed recordings to extract notes, defined as a continuous sound bordered by a silent interval [36]. Each note was saved in a single file, extracting the note from the original recording with 20 ms of silence at the beginning and the end of the note. ACMM then conducted visual comparison of spectrograms for each note to classify notes in different types. Although we carried out a full account of all notes emitted to evaluate the diversity of acoustic signals, we selected only note-types emitted more than once to describe vocalisations. For statistical analysis we used only note-types that were emitted at least five times across all recordings, and randomly selected five high-quality notes, with low background noise and a high signal-to-noise ratio, for each note-type. We measured five spectrographic variables in Raven Pro 1.4: i) note duration (ms); ii) low frequency (Hz); iii) high frequency (Hz), giving the lower and upper frequency bounds; iv) delta frequency or bandwidth (Hz), being the difference between the upper and lower frequency bounds; and v) number of harmonics. In addition, we used Sound Analysis Pro SA.04 [36, 37] to obtain 6 spectral derivatives for each note: i) mean pitch (Hz), or tone, is a measure of the period of oscillation, or number of cycles made by a sound wave in a unit of time; ii) variance of pitch; iii) mean frequency (Hz), estimates the central tendency of the distribution of power across frequencies; iv) goodness of pitch (Hz), is the peak of the power spectrum for harmonic pitch; v) frequency modulation (deg), is the slope angle of frequency contours; and vi) Weiner entropy, gives a measure of order in the waveform of the sound on a logarithmic scale of 0 (disorder) to minus infinity (complete order). We thereby obtained a total of 11 variables for each note. Vocalisations were associated with nine behavioural contexts [7, 19, 20, 24, 27]: 1) Nesting activity, when the male returned to the nest after foraging, and called the incubating female who vocalised on leaving the nest-cavity to be fed [30]; 2) Threat interactions of agonistic encounters between conspecifics, usually around the nest; 3) Alarm vocalisations emitted in the presence of potential predators; 4) Foraging, emitted by individuals while foraging in trees; 5) Perched, when parrots were perched inactive or at rest in a tree; 6) Take-off, vocalisations emitted seconds before, during and after flight take-off by parrots; 7) Flight, obtained from flying parrots; 8) Landing, vocalisations of parrots on final flight approach to land in a tree; and 9) Food soliciting, begging by nestlings soliciting food from parent birds, and nesting females soliciting food from males. Given the difficulties of capturing and marking free-ranging parrots, we considered individual identification based on nest-site ownership for recordings obtained at nest sites. Reproductive pairs of Lilac-crowned Amazons are highly synchronous in nesting behaviour [30]. Therefore for many of the behavioural contexts we used vocalisations recorded at different nest-sites where we could be confident of individual identification. For recordings of behavioural contexts away from nest-sites (foraging, flight), we used only recordings obtained on different days and those that were sufficiently separated by distance among sites to potentially represent different individuals, considering the daily foraging distances travelled by Lilac-crowned Amazons [38]. We were able to distinguish between male and female parrots at nests, as only the female incubates [30]. Therefore, we described both male and female nesting vocalisations, but as this was not possible for other behavioural contexts we did not separate nesting vocalisations by gender for statistical analyses of acoustic parameters among contexts. We collected 75 h of non-continuous recordings over all behavioural contexts, obtaining a total of 8622 notes emitted by 61 Lilac-crowned Amazon individuals that could be spectrographically classified in 152 note-types. Statistical analyses A third of note-types were emitted only once, and were not considered in further analyses, leaving a total of 8571 notes that comprised 101 note-types emitted at least twice. For each of the nine behavioural contexts, we determined the emission rate of notes per second, calculated as the total number of notes emitted divided by time from when the first note was emitted to the last note for that context, which was used as a measure of intensity of vocal activity during the recording period for each behavioural context. We also determined the frequency of occurrence of the most common note-types emitted by adult parrots across all recordings in each behavioural context, and applied chi-square contingency table analysis to determine whether note-types were associated with a specific behavioural context. We calculated adjusted standardised residuals for each cell [39] to determine which notes were used more than expected in each context. For statistical analyses of acoustic parameters, we eliminated 43 note-types that were emitted less than five times, or had poor quality recordings with a lot of background noise or overlapped other notes. This gave a total of 58 note-types of sufficient sound quality and frequency of emission that were used in statistical parameter analyses among behavioural contexts. We used Generalised Linear Mixed Models (GLMM) fit by maximum likelihood, where we considered behavioural context as a fixed effect, and included individual identity as a random effect across contexts. We excluded the food soliciting context from these analyses, as this included vocalisations of nestlings that were not considered in other contexts. We applied GLMM with log-link using a negative binomial error distribution that showed the best fit to the plot of residuals for the vocal parameters of emission rate, duration, low and high frequencies, and bandwidth. To evaluate number of harmonics among contexts, we employed GLMM with log-link and a Poisson error distribution model that best fit the residuals for the data. We obtained significance values by performing likelihood ratio tests comparing the full model including the effect of behavioural context against the reduced model without the effect. We compared parameter estimates with the intercept, set as the context with lowest mean values. We used the Automatic Differentiation Model Builder (glmmADMB) package [40] to run negative binomial GLMMs, and the lme4 package [41] for the Poisson distribution GLMM, both available in R 3.2.3 [42]. To acoustically discriminate among note-types emitted by adults taking into account individual variation in notes, we selected a data set of 28 note-types that were emitted by more than one individual from the 58 note-types emitted at least five times. We first applied principal component analysis (PCA) on the data set of 28 note-types to convert the 11 spectral and time variables for each note to a reduced set of linearly uncorrelated variables. We then used the Principal Components with eigenvalues >1 in a permuted Discriminant Function Analysis (pDFA) with a nested design that deals with potential non-independence of data when several vocalisations from one individual are included in the data set [43]. Acoustic parameters were considered to aid differentiation among note-types when the observed correct classification was significantly higher (with P < 0.05) than the expected correct classification for the null hypothesis of no discrimination among note-types. The expected distribution of the correctly classified signals under the null hypothesis that note-types cannot be acoustically discriminated was obtained performing 1000 permutations. The pDFA was performed using a script for R version 3.0.1 [42] written by R. Mundry, based on the function lda of the R package MASS [44]. Datasets for the statistical analyses are provided in an additional Excel file (Additional file 1). Descriptive statistics are presented as mean with standard error, applying a P < 0.05 significance level for statistical analysis. Results Design characteristics of vocalisations by behavioural context We provide descriptions of acoustic and behavioural characteristics of each context in an additional Word file (Additional file 2). Overall, a total of 73 note-types were exclusive to a particular behavioural context (Table 1), although the majority of these were emitted infrequently, as shown in Additional file 3: Table S1 that gives the percent emission in nine behavioural contexts for the 101 note-types emitted at least twice (Additional file 3). The greatest variety of 64 note-types were emitted during threat contexts (Table 1). Threat interactions also had the highest number of exclusive notes, where 64 % of note-types emitted during threat interactions were exclusive to this context (Table 1). Vocalisations emitted during threat contexts often involved counter-calling between conspecifics, and were sometimes accompanied by visual displays, such as the wing display where parrots raised both wings in an arc above the body [45], or the tail-fan (Additional file 2). Nesting vocalisations were also highly diverse, with the second-highest number of exclusive notes (Table 1), particularly with regard to male vocalisations where 51.9 % of note-types were exclusively used to call the female from the nest. Among these, the exclusive note Z4 was emitted by males on final approach to the nest, and was spectrographically and acoustically similar to the ‘grr-uíp’ vocalisation reported for the Blue-fronted Amazon [24], as shown in an additional figure (Additional file 4) and audio file (Additional file 5). By comparison, alarm vocalisations given in the presence of avian predators such as the Crane Hawk (Geranospiza caerulecens) and Collared Forest Falcon (Micrastur semitorquatus) had the lowest variety of only nine different note-types, none of which were exclusive to alarms (Table 1). These may be similar to assembly signals as on one occasion we observed 6 Lilac-crowned Amazons flying to congregate with another vocalising pair in response to their alarm calls given on approach by a pair of Collared Forest Falcons.Table 1 Frequency of 101 note-types emitted more than once by Lilac-crowned Amazons in nine behavioural contexts Behavioural context Parrot individuals Behavioural encounters Total mins recorded Total notes emitted Number of note-types Number of exclusive note-types Alarm 6 4 3.2 577 9 0 Threat 32 15 35.2 1534 64 41 Flight 28 24 6.7 525 15 5 Take-off 14 14 3.4 238 14 1 Landing 18 29 29.2 1196 24 5 Perched 18 14 19.4 298 16 1 Foraging 6 12 35 451 14 2 Soliciting food (Adult, Chicks) 10 (8, 2) 5 (4, 1) 8.3 (4.7, 3.6) 193, NA 11 (10, 1) 3 (2, 1) Nesting (Male, Female) 36 (21, 15) 160, 39 92.1 (89.3, 2.8) 3565 (3423, 142) 27 (25, 8) 16 (14, 1) Acoustic parameter analysis for 58 note-types emitted ≥5 times, determined that including behavioural context as a fixed factor in GLMMs significantly explained variations in note duration (GLMM: χ27 = 18.2, P = 0.011), low frequency (GLMM: χ27 = 74.2, P < 0.001), high frequency (GLMM: χ27 = 53.5, P < 0.001), bandwidth (GLMM: χ27 = 46.1, P < 0.001), number of harmonics (GLMM: χ27 = 50.0, P < 0.001), and emission rate (GLMM: χ27 = 75.5, P < 0.001). The behavioural contexts with most distinct acoustic characteristics were nesting and threat interactions (Fig. 1). Nesting vocalisations had on average notes of longer duration, with the greatest number of harmonics compared to other contexts (Fig. 1). In general, most note-types had three to four harmonics (Mean: 3.4 ± 1.76 harmonics, range = 0–13 harmonics, n = 58 note-types), a characteristic of long-range signals in other Amazon parrots [24]. The behavioural contexts of nesting and threat interactions had notes with higher low and high frequencies, and broad bandwidth (Fig. 1). By comparison, alarm vocalisations were distinct in their high emission rate of 3.6 notes/s (Fig. 1), while contexts of perched and foraging had low vocal activity (Fig. 1).Fig 1 Mean acoustic characteristics of vocalisations emitted by Lilac-crowned Amazons in eight behavioural contexts. (a) note duration, (b) low frequency, (c) high frequency, (d) bandwidth, (e) number of harmonics, all calculated from 58 note-types emitted ≥5 times across recordings. (f) emission rate was calculated considering all notes emitted in each behavioural context. Error bars show standard error. Significant from intercept: * P < 0.05, ** P < 0.01, *** P < 0.001. Forage was set as the baseline intercept in GLMM for all variables except note duration, which was set with the perched context Classification by note-type Considering all notes emitted over all behavioural contexts, seven note-types (Fig. 2a) represented more than 80 % of all notes emitted, and are included in additional audio files (Additional files 6, 7, 8, 9, 10, 11, 12, 13, and 14). Of these, note-types C (26.4 % of notes), B (23.7 %), and J4 (11.5 %) were emitted most frequently, and used in the majority of behavioural contexts. Notes B and C were emitted in all behavioural contexts, but with greater percent contribution in contexts where it was necessary to attract attention of the group or individual such as in alarm, threat, flight, landing, and nesting vocalisations (Fig. 3). By comparison, note J4 was used mainly when foraging, soliciting food, just prior to take-off, and when perched at rest (Fig. 3). The frequency of emission for the seven most-common note-types was significantly associated with behavioural context (χ254 = 3984, P < 0.001). In particular, note-type J4 was used significantly more when foraging (57 % of notes emitted; cell z = 30.3), soliciting food (39 %; cell z = 16.7), just prior to take-off (32 %; cell z = 9.8), and when perched (22 %; cell z = 5.0). The other most common note-type emitted by parrots when perched was note D, which comprised just under half of notes emitted when perched (Fig. 3), and was emitted significantly more than expected in this context (cell z = 28.0). Note C was the most common note-type used in threat context (26 %; cell z = 5.9; Fig. 3), with the growl-like note E also emitted more frequently than expected in threat interactions (cell z = 11.4), but used infrequently in other contexts (Fig. 3). Alarm vocalisations were characterised by note B (45 %; cell z = 8.7), although note C2 was also emitted more than expected in alarm contexts (24 %; cell z = 20.3), and these two notes comprised almost 70 % of notes emitted in alarm contexts. Note-type C2 was particularly characteristic of female nesting vocalisations (52 %; cell z = 26.4), while male nesting vocalisations were characterised by note-types C (30 %; cell z = 6.5) and B (27 %; cell z = 6.8).Fig 2 Spectrograms of (a) 7 note-types most frequently emitted by Lilac-crowned Amazons, and (b) female and nestling begging vocalisations to solicit food Fig 3 Percent contribution by behavioural context of the 7 note-types most frequently emitted by Lilac-crowned Amazons. Values above columns denote sample size of notes in each context Principal Components Analysis performed with 11 spectral and temporal variables yielded three principal components with eigenvalues >1 for 28 note-types emitted ≥5 times, where each note-type was emitted by >1 individual (Table 2). These components explained 72.03 % of total variance among notes. The variables with greatest loading on Principal Component 1 were variance of pitch, frequency modulation, goodness of pitch, and duration (Table 2). Component 2 was influenced mainly by mean pitch, low frequency, and Weiner entropy (Table 2). The parameters with greatest weight for Principal Component 3 were bandwidth, high frequency, and number of harmonics (Table 2). Discriminant Function Analysis (DFA) for categories of 28 note-types using these three components determined an observed correct classification of 72.7 % that was significantly larger than the expected correct classification for the null hypothesis (Expected = 24.0 %; P = 0.001). Similarly, when analysis was controlled by individuals in the pDFA for 28 note-types emitted by more than one individual, the 46.9 % observed correct classification was significantly larger than the 9.6 % expected correct classification by chance (P = 0.001).Table 2 Principal Components with eigenvalues >1 for 28 note-types emitted ≥5 times across the study, where each note-type was emitted by >1 individual of the Lilac-crowned Amazon. PC1 PC2 PC3 Overall eigenvalues 3.49 2.67 1.76 Explained variation (%) 26.3 23.4 22.3 Variance Pitch 0.805 0.384 0.211 Frequency Modulation 0.794 −0.325 −0.032 Goodness of Pitch −0.704 0.157 0.293 Duration −0.697 −0.104 0.297 Mean Pitch −0.165 0.857 0.158 Low Frequency 0.083 0.775 −0.058 Weiner Entropy 0.531 −0.659 −0.158 Mean frequency 0.483 0.585 0.133 Bandwidth −0.037 0.176 0.942 High frequency −0.026 0.268 0.916 Harmonics −0.287 −0.266 0.663 Bold text highlights variables with greatest weighting for each component (r >0.60) Discussion High vocal diversity The Lilac-crowned Amazon demonstrated a high diversity of 101 note-types emitted more than once, with 58 note-types emitted at least five times, which is one of the largest vocal repertoires so far recorded for Psittaciformes [12, 13, 19–21, 23–29]. However, most studies report calls [12, 13, 15–17, 19–24, 27, 28], or vocalisations [14, 18] that may comprise a number of notes. Fernández-Juricic et al. [24] describe nine call types for the Blue-fronted Amazon, where just the guttural call has at least 23 different note-types [22], and breeding season songs have 17 note-types [24]. Conversational chattering by the Brown-headed Parrot (Poicephalus cryptoxanthus) also comprises 23 note-types [17]. Similarly, de Moura et al. [19] classified 9 vocalisations comprised of a total of 36 note-types for the Orange-winged Amazon (Amazona amazonica), and May [26] identified 39 acoustic call types for the Grey Parrot (Psittacus erithacus). In a similar approach to the present study, Zdenek et al. [25] identified 27 structurally distinct note-types in the vocal repertoire of the Palm Cockatoo. These studies demonstrate that many parrot species have a high diversity of 23–40 note-types commonly used in the vocal repertoire, with the Lilac-crowned Amazon presenting one of the most diverse vocal repertoires, having 58 commonly used note-types. One explanation for the high diversity of note-types found in our study may be that we have a high recording sample. However, our sample of 75 h of recordings is in the mid-range of that reported by other studies, with lower recording samples of 10 h obtained by Fernández-Juricic and Martella [22] and 30 h by May [26], but larger samples of 100 h reported by Fernández-Juricic et al. [24] and 210 h obtained by Zdenek et al. [25]. Therefore, sample size of recordings is unlikely to explain the high diversity of note-types found in our study. A number of hypotheses may explain this high diversity of vocalisations for the Lilac-crowned Amazon and other parrots, such as acoustic adaptation to forest habitats, social complexity, and ecological characteristics of species. The Lilac-crowned Amazon may require a larger vocal repertoire to facilitate communication in a complex forest habitat [46, 47]. This could be of particular importance in the tropical dry forest where there is dramatic seasonal variation in phenological characteristics of the forest [32], and the two main habitats of deciduous and semi-deciduous forest have differing vegetation structure [33, 34]. According to the acoustic adaptation hypothesis [46–48], variations in habitat structure affect sound transmission [48, 49] leading to the selection of signals structured to transmit with minimal distortion through native habitat [46–48]. Forests are complex environments, and even slight changes in vegetation structure would have effects on sound transmission with vocal learning species able to rapidly adapt acoustic signals [47], potentially leading to greater diversity in vocal communication [3]. Social organisation may also influence vocal repertoire for species such as parrots, with large, complex social groups. The social complexity hypothesis [50] states that groups with complex social systems require more complex communicative systems to regulate interactions and relations among group members. This variety of social interactions may then lead to the development of large vocal repertoires [50–52]. Parrots exhibit complex social systems [7] where for most species the basic unit is the mated pair, but species such as the Lilac-crowned Amazon also form large communal roosts, smaller foraging flocks, and are territorial around nests in the breeding season [30, 53]. Other parrot species with a large diversity of note-types in vocal repertoires, such as the Brown-headed Parrot, Orange-winged Amazon, Blue-fronted Amazon, and Grey Parrot, exhibit similar flexibility in social organisation [17, 19, 24, 26]. This provides many occasions when individuals may switch group affiliation, requiring mechanisms for recognizing individuals, potentially increasing vocal diversity. Parrots also establish dominance hierarchies by vocal communication [7], requiring complex vocal repertoires to maintain this social complexity. A simple, auditory description of vocalisations by three parrot species in Australia appears to indicate that the species with a more complex hierarchy of groups and individuals has the greatest number of distinct auditory signals [54]. Pidgeon [12] also suggests for five Australian parrot species, that species with more agonistic interactions have a greater number of auditory signals. However, no studies have as yet evaluated the social complexity hypothesis with regard to parrot vocal communication. Bradbury and Balsby [55] recently suggest that diet-driven social dynamics may explain extensive vocal learning in Psittaciformes. Parrots consume highly variable plant resources of flowers, fruits, or seeds [56], requiring extensive knowledge of potential food resources, and foraging in flexible flocks over a wide area, where vocal learning with the acquisition of new acoustic signals would facilitate identification of individuals with knowledge of food sources [55]. The Lilac-crowned Amazon has a predominantly granivorous diet [57], uses communal roosts [53], and forms foraging flocks with large home-ranges [38]. This species also undertakes seasonal migrations to track food resources [53, 57] that may require a capacity to learn new vocalisations in different regions, habitats, and social groups [58, 59]. Parrot species with vocal learning have been found to modify their vocalisations on relocation to new sites with new social groups [60], and migratory behaviour is associated with larger song repertoires within genera of passerine birds [61]. Therefore, given that parrots maintain vocal learning ability throughout life [7], individuals may encounter and acquire new elements in the vocal repertoire during long-distance movements, and interchange among foraging flocks, leading to high vocal diversity and a low proportion of context-exclusive notes, as found in the vocal repertoire of the Lilac-crowned Amazon. The Orange-fronted Parakeet (Eupsittula canicularis) also inhabits seasonal tropical dry forest, and exhibits fission/fusion flock dynamics [7], but the species has a smaller vocal repertoire [7, 11, 62] compared to the Lilac-crowned Amazon. Therefore, other factors may be influencing vocal diversity of the Lilac-crowned Amazon. One factor may be the larger ranging areas of the Lilac-crowned Amazon with an average home-range estimated to be 4674 ha [38] compared to 666 ha for the Orange-fronted Parakeet which exhibits range lengths of just 6–9 km [63]. Larger movements by the Lilac-crowned Amazon mean that the species is likely to encounter heterogeneous environmental and social conditions that could promote diversity in the vocal repertoire. Another key ecological difference is that the Orange-fronted Parakeet excavates nest-cavities in arboreal termiteria [11] that are generally abundant resources but with only short-term longevity [64, 65]. By comparison, most parrot species, including the Lilac-crowned Amazon, depend on pre-existing naturally-formed tree-cavities [56] that are limited but long-term resources, and exhibit intense intraspecific competition for nest-sites [45]. High vocal diversity may serve to intimidate conspecifics, particularly competitors for nest-cavities, and may reflect selective pressures for a larger vocal repertoire during territorial defense. In support of this, we found greater vocal diversity of Lilac-crowned Amazons in threat interactions with conspecifics around nests compared to other behavioural contexts. Therefore, we consider that competition with conspecifics for scarce, suitable tree-cavity resources may be a contributing factor increasing social complexity and vocal diversity. Hence, the Lilac-crowned Amazon may have a diverse vocal repertoire given that the species inhabits a heterogeneous, seasonal, forest environment, has complex social dynamics including strong intraspecific competition for nest-sites, ranges over a large area, and undertakes long-distance migrations to alternate habitats and regions, all of which may require vocal adaptation to changing environmental conditions and social complexity. Design characteristics of the vocal repertoire Behavioural context significantly explained variations in acoustic characteristics of vocalisations emitted by the Lilac-crowned Amazon. In accordance with signal design rules, threat vocalisations were on average of short duration, with a high emission rate, broad bandwidth, and frequently involved counter-calling, which may encode information on motivation in threat vocalisations. Threat vocalisations were not of low frequencies that could indicate large body size, but had the highest frequency values of all contexts. This may reflect the short-range aspect of threat signals as sender and receiver are generally in close proximity. Furthermore, parrots frequently combined acoustic threat signals with visual displays that may effectively indicate body size, motivation, and an escalation of aggression. Other parrot species have also been reported to use compound signals of high frequency vocalisations with visual displays in threat context [19, 21, 24]. These features correspond to the design rules for threat displays, where a high vocal diversity of notes emitted by Lilac-crowned Amazons, and their combination with visual displays, would permit encoding of additional information on status, body size, intensity, and motivation during threat interactions [3]. Nesting vocalisations were also of high frequency, being short-range signals used between the nesting pair. Nesting vocalisations had the longest note duration which would increase their duty cycle, or percent of time that the signal is active, and broad bandwidth that may have capacity to carry more information in the signal. These are similar to the design rules for courtship signals having both male and female components that are given in a specific sequence [3], although in this case the pair is already mated. This suggests that nesting vocalisations may have a similar role in coordinating activities of the nesting pair; however, experimental evidence is required to determine the function of acoustic signals in nesting contexts. Alarm signals comprised notes of relatively short duration, with low frequencies, short bandwidth, and had the highest emission rate of 3.6 notes/s. Other parrot species also present alarm vocalisations with high emission rates of short, repeated notes [21, 27, 66]. Wheatcroft [67] determined that various bird species use increased signal repetition rate on approach by a predator, which is recognised as a contextual cue by both adults and nestlings influencing their responses. The low frequencies of alarm signals emitted by the Lilac-crowned Amazon may increase their range through forest habitats as low frequency sounds are less easily absorbed and travel further than high frequency sounds [49]. Features of alarm signals may vary between the extremes of flee and assembly signals, with alert signals having intermediate features [3]. In the case of the Lilac-crowned Amazon, alarm signals given in response to avian predators had features of alert or assembly signals, which are short pulses that are regularly repeated to attract attention and enable location of the sender, rather than flee signals designed to reduce locatability of the sender [3]. Foraging and perched contexts had the lowest emission rates of only 0.3 notes/s that were of low frequencies, short bandwidth, and with few harmonics. This would reduce locatability of individuals where there is a potential cost in attracting predators while individuals are resting, or distracted by foraging. It may also be that vocalisations given in these contexts only need to indicate presence, and therefore do not require greater complexity to communicate more information. It should be noted however, that sample sizes were low for some behavioural contexts, which could be influenced by individual differences, limiting our conclusions on the acoustic characteristics of these contexts. Evidence from playback experiments is also needed to determine the function of acoustic signals used in distinct behavioural contexts. Contextual flexibility in use of notes Note-types could be discriminated by acoustic features, with more than half of all note-types being exclusive to a specific behavioural context, although the seven most common note-types were emitted by Lilac-crowned Amazons in a variety of contexts, but with differing proportional contribution in each context. The common notes B and C were used by Lilac-crowned Amazons with greater frequency in high intensity behaviours of threat, nesting, alarm, and flight, whereas notes J4 and D had greater proportional contribution in low intensity behaviours of foraging, prior to take-off, or when parrots were perched at rest. Threat contexts had the highest variety of notes, and greatest number of exclusive notes, which may reflect a greater complexity of vocalisations. Zdenek et al. [25] suggest that the complex vocal repertoire of the Palm Cockatoo functions in year-round territorial defense. This has been observed in songbirds, where larger song repertoires are more effective at deterring invaders than small or single-song repertoires [68]. In particular, the growl-like note E that was emitted more frequently than expected in threat context may be similar to the soft vocalisations produced by songbirds in aggressive encounters that are reliable indicators of motivation [69–71]. The next most vocally diverse context for Lilac-crowned Amazons was that of nesting vocalisations, particularly with respect to vocalisations of males calling females out of the nest-cavity. The variety of notes and high proportion of exclusive notes emitted by males when calling the nesting female may permit encoding of individual identity, particularly during incubation of eggs or nestlings when the female lacks visual contact with the male from within the nest-cavity. However, it may be that only a small sample of a vocalisation is required for individual recognition, as Mockingbirds (Mimus gilvus) were found to respond to playbacks of conspecifics within seconds, even when presented with only a fraction of the hundreds of song types available per singer [72]. In this sense, the exclusive note Z4 emitted by males on final approach to the nest may contain information on individual identity, alerting the incubating female to her mate’s arrival. The high vocal diversity we found for threat and nesting contexts could be a result of longer recording times, increasing the sample size for these contexts. However, we have equally long recording times for foraging and landing contexts, and these do not show similar vocal diversity, particularly in the number of exclusive notes. Therefore the incorporation of a high number of exclusive notes in threat and nesting contexts may reflect the complexity of interactions and amount of information to be communicated. By comparison, alarm vocalisations had the lowest variety of note-types, and contrary to expectation, these were not context-specific, but consisted in frequent repetition of three commonly used note-types (B, C, and C2). The Japanese Great Tit (Parus minor) has been shown to use acoustically discrete alarm signals for snake predators, but does not use predator-specific alarm signals when mobbing avian predators [73]. Instead of discrete signals for different species of avian predator, birds may vary note repetitions and combinations in compositional syntax to encode information about predator type [73, 74]. It would be interesting therefore to determine whether parrots give different types of alarm signals for terrestrial predators as opposed to avian predators. Finally, in foraging and perched contexts Lilac-crowned Amazons predominantly used note-types J4 and D, which comprised >60 % of notes emitted and were produced more than expected in these contexts. This contextual flexibility in the use of notes across behavioural contexts may suggest that the vocal repertoire contains a large amount of redundancy in acoustic signals [3]. Alternatively, it may indicate that parrots use graded or combinatorial variation to encode information for different contexts, where the compositional syntax, or the way in which notes are combined, is essential for communicating different messages [74–77]. Our findings on note-type composition in different behaviour contexts suggest that Lilac-crowned Amazons use a variety of strategies for acoustic communication. Other parrot species have also been found to emit calls or notes in a variety of behavioural contexts [13, 20, 22–24], although no studies have determined the cross-functional contribution of commonly used notes in differing behavioural contexts. Functional or contextual flexibility in vocalisations has been determined in non-human primates [78], but there is a paucity of evidence to evaluate the existence of this in other animal groups. Nevertheless, some studies have determined that avian species with small repertoires may use combinatorial structures in compositional syntax to achieve greater communicative complexity [74, 75]. Experimental evidence could determine the acoustic strategies and combinatorial structures employed by parrots for communication in different contexts. Conclusions The Lilac-crowned Amazon presents a diverse vocal repertoire of note-types that are used in a variety of behavioural contexts. This may provide more dimensions for encoding information, which could help the Lilac-crowned Amazon to deal with the constraints imposed on communication within a complex social and natural environment. It is important to evaluate not just the acoustic features and types of notes emitted in each behavioural context, but the compositional syntax of notes used in different contexts [74–77]. Therefore, we propose that evaluation of parrot vocal repertoires based on note-types emitted as the basic unit would reflect the potential vocal diversity of each species. Statistical analysis of the acoustic features of notes, their contribution in each behavioural context, and their combinatorial structures, would reflect the true extent of the species’ vocal flexibility. This would enable comparative studies of vocal diversity among psittacine species to evaluate the relationship of vocal repertoire with habitat structure and social organisation. The cross-functional use of vocalisations by parrots in differing behavioural contexts also makes them ideal species for elucidating signal design rules for differing social functions [3]. Understanding the vocal repertoires of free-living Psittaciformes is essential as a foundation for future research on the extensive vocal learning abilities of parrots [55], the use of combinatorial structures in vocal communication, and the parallels with human language development. Additional files Additional file 1: Datasets for statistical analyses by context and note-type. (XLSX 116 kb) Additional file 2: Descriptions of behavioural and acoustic characteristics in each context. (DOCX 15 kb) Additional file 3: Table S1. Percent emission by Lilac-crowned Amazons in nine behavioural contexts for 101 note-types emitted at least twice across all recordings. (DOCX 24 kb) Additional file 4: Spectrogram of note Z4 emitted by males on approach to the nest. (TIF 593 kb) Additional file 5: 16-bit WAV sound file of note Z4 emitted by a male Lilac-crowned Amazon on final approach within sight of the nest. (WAV 116 kb) Additional file 6: 16-bit WAV sound file of note C emitted by a Lilac-crowned Amazon. (WAV 294 kb) Additional file 7: 16-bit WAV sound file of note B emitted by a Lilac-crowned Amazon. (DOCX 15 kb) Additional file 8: 16-bit WAV sound file of note A emitted by a Lilac-crowned Amazon. (DOCX 24 kb) Additional file 9: 16-bit WAV sound file of note C2 emitted by a Lilac-crowned Amazon. (TIF 593 kb) Additional file 10: 16-bit WAV sound file of note E emitted by a Lilac-crowned Amazon. (WAV 46 kb) Additional file 11: 16-bit WAV sound file of note J4 emitted by a Lilac-crowned Amazon. (WAV 20 kb) Additional file 12: 16-bit WAV sound file of note D emitted by a Lilac-crowned Amazon. (WAV 24 kb) Additional file 13: 16-bit WAV sound file of train of begging notes by a female Lilac-crowned Amazon to the male. (WAV 25 kb) Additional file 14: 16-bit WAV sound file of train of begging notes emitted by nestling Lilac-crowned Amazons in presence of the parent birds. (WAV 19 kb) Abbreviations GLMMGeneralised Linear Mixed Models PCAPrincipal Component Analysis PCPrincipal Component DFADiscriminant Function Analysis pDFAPermuted Discriminant Function Analysis The study was conducted in partial fulfilment of a Ph.D. degree by ACMM at the Posgrado en Ciencias Biológicas, of the Universidad Nacional Autónoma de México (UNAM). We are grateful to the Fundación Ecológica de Cuixmala A.C. for logistical support. We thank Karen Esquivel for assistance in the field and with analysis of sound recordings, and are grateful to Lynna Kiere for assistance with GLMM analysis. Research permits were authorised by the Secretaria del Medio Ambiente y Recursos Naturales of Mexico. The Posgrado en Ciencias Biológicas, UNAM, and Estación de Biología Chamela, Instituto de Biología, UNAM provided facilities for the preparation of this manuscript. We are grateful to the anonymous reviewers and Associate Editor for their constructive comments that greatly improved the manuscript. Funding Funding for the research was provided by the Consejo Nacional de Ciencia y Tecnología (CONACyT) through project grants 179877 to KR, and C-965/2014 for ASM, while the Fundación Ecológica de Cuixmala provided logistical support. CONACyT also provided a Doctoral scholarship to ACMM (231685), and postdoctoral grant to ASM (98294). Availability of data and materials The datasets supporting the conclusions of this article are included within the article and additional files (Additional files 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, and 14). Authors’ contributions KR, ASM, and ACMM conceived and designed the study. ACMM and ASM generated field data. ACMM, ASM, and KR analysed and interpreted the data. All authors wrote, revised, and approved the final version of the manuscript for publication. Competing interests The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate The study involved free-ranging individuals which were not manipulated by researchers. Permits for the research were granted by the Secretaria del Medio Ambiente y Recursos Naturales, Mexico. ==== Refs References 1. Kroodsma DE Byers BE The function(s) of bird song Am Zool 1991 31 318 328 10.1093/icb/31.2.318 2. Catchpole CK Slater PJB Bird songs: biological themes and variations 2008 2 Cambridge Cambridge University Press 3. Bradbury JW Vehrencamp SL Principles of animal communication 2011 2 Sunderland Sinauer 4. Bretagnolle V Kroodsma DE Miller EH Acoustic communication in a group of nonpasserine birds, the petrels Ecology and evolution of acoustic communication in birds 1996 Ithaca Cornell University Press 160 178 5. Favaro L Ozella L Pessani D The vocal repertoire of the African Penguin (Spheniscus demersus ): structure and function of calls PLoS ONE 2014 9 e103460 10.1371/journal.pone.0103460 25076136 6. Odom KJ Mennill DJ A quantitative description of the vocalizations and vocal activity of the Barred Owl Condor 2015 112 549 560 10.1525/cond.2010.090163 7. Bradbury JW DeWaal FBM Tyack PL Vocal communication in wild parrots Animal social complexity: intelligence, culture, and individualized societies 2003 Cambridge Harvard University Press 293 316 8. Pepperberg IM Vocal learning in Grey Parrots (Psittacus erithacus ): effects of social interaction, reference, and context Auk 1994 111 300 313 10.2307/4088595 9. Patterson DK Pepperberg IM A comparative study of human and parrot phonation: acoustic and articulatory correlates of vowels J Acoust Soc Am 1994 96 634 648 10.1121/1.410303 7930064 10. Beckers GJL Nelson BS Suthers RA Vocal-tract filtering by lingual articulation in a parrot Curr Biol 2004 14 1592 1597 10.1016/j.cub.2004.08.057 15341748 11. Hardy JW Epigamic and reproductive behavior of the Orange-fronted Parakeet Condor 1963 65 169 199 10.2307/1365664 12. Pidgeon R Call of the Galah Cacatua roseicapilla and some comparisons with four other species of Australian Parrots EMU 1981 81 158 168 10.1071/MU9810158 13. Saunders DA Vocal repertoire and individual vocal recognition in the Short-billed White-tailed Black Cockatoo, Calyptorhynchus funereus latirostris Carnaby Wildlife Res 1983 10 527 536 10.1071/WR9830527 14. Snyder NFR Wiley JW Kepler CB The parrots of Luquillo: natural history and conservation of the Puerto Rican Parrot 1987 1 Los Angeles Western Foundation of Vertebrate Zoology 15. Toyne EP Flanagan JNM Jeffcote MT Vocalizations of the endangered Red-faced Parrot Hapalopsittaca pyrrhops in southern Ecuador Ornitol Neotrop 1995 6 125 128 16. Wirminghaus JO Downs CT Symes CT Dempster E Perrin MR Vocalisations and behaviours of the Cape Parrot Poicephalus robustus (Psittaciformes: Psittacidaes) Durban Mus 2000 25 12 17 17. Taylor S Perrin MR Vocalisations of the Brown-headed Parrot, Poicephalus cryptoxanthus : their general form and behavioural context Ostrich 2005 76 61 72 10.2989/00306520509485474 18. Tovar-Martínez AE Redescubrimiento y notas sobre la ecología y vocalizaciones del Periquito de Todd (Pyrrhura picta caeruleiceps ) en el nororiente de Colombia Ornitologia Colombiana 2010 9 48 55 19. de Moura LN Da Silva ML Vielliard J Vocal repertoire of wild breeding Orange-winged Parrots Amazona amazonica in Amazonia Bioacoustics 2011 20 331 339 10.1080/09524622.2011.9753655 20. Fernández-Juricic E Alvarez EV Martella MB Vocalizations of Blue-crowned Conures (Aratinga acuticaudata ) in the Chancaní reserve, Córdoba, Argentina Ornitol Neotrop 1998 9 31 40 21. de Araújo CB Marcondes-Machado LO Vielliard JME Vocal repertoire of the Yellow-faced Parrot (Alipiopsitta xanthops ) Wilson J Ornithol 2011 123 603 608 10.1676/10-102.1 22. Fernández-Juricic E Martella MB Guttural calls of Blue-fronted Amazons: structure, context and their possible role in short range communication Wilson Bull 2000 112 35 43 10.1676/0043-5643(2000)112[0035:GCOBFA]2.0.CO;2 23. Van Horik J Bell B Burns KC Vocal ethology of the North Island Kaka (Nestor meridionalis septentrionalis ) New Zeal J Zool 2007 34 337 345 10.1080/03014220709510093 24. Fernández-Juricic E Martella MB Alvarez EV Vocalizations of the Blue-fronted Amazon (Amazona aestiva ) in the Chancaní Reserve, Córdoba, Argentina Wilson Bull 1998 110 352 361 25. Zdenek CN Heinsohn R Langmore NE Vocal complexity in the Palm Cockatoo (Probosciger aterrimus ) Bioacoustics 2015 24 253 267 10.1080/09524622.2015.1070281 26. May DL. Vocal repertoire of Grey Parrots (Psittacus erithacus) living in the Congo Basin. Ph.D. Thesis. Tuscon, Arizona: The University of Arizona. 2004. http://arizona.openrepository.com/arizona/handle/10150/280626. Accessed 22 Aug 2016. 27. Martella MB Bucher EH Vocalizations of the Monk Parakeet Bird Behav 1990 8 101 110 10.3727/015613890791784290 28. Chan K Mudie D Variation in vocalisations of the Ground Parrot at its northern range Aust J Zool 2004 52 147 158 10.1071/ZO03009 29. Schwing R Parsons S Nelson XJ Vocal repertoire of the New Zealand Kea parrot Nestor notabilis Curr Zool 2012 58 727 740 10.1093/czoolo/58.5.727 30. Renton K Salinas-Melgoza A Nesting behavior of the Lilac-crowned Parrot Wilson Bull 1999 111 488 493 31. Bullock SH Climate of Chamela, Jalisco, and trends in the south coastal region of Mexico Arch Meteor, Geophy B 1986 36 297 316 10.1007/BF02263135 32. Bullock SH Solis-Magallanes JA Phenology of canopy trees of a tropical deciduous forest in Mexico Biotropica 1990 22 22 35 10.2307/2388716 33. Lott EJ Bullock SH Solis-Magallanes JA Floristic diversity and structure of upland and arroyo forests of coastal Jalisco Biotropica 1987 19 228 235 10.2307/2388340 34. Rzedowski J. Vegetación de México. Comisión Nacional para el Conocimiento y Uso de la Biodiversidad. 2006. http://www.biodiversidad.gob.mx/publicaciones/librosDig/pdf/VegetacionMx_Cont.pdf. Accessed 10 Mar 2016. 35. Salinas-Melgoza A Renton K Seasonal variation in activity patterns of juvenile Lilac-crowned Parrots in tropical dry forest Wilson Bull 2005 117 291 295 10.1676/04-096.1 36. Tchernichovski O Nottebohm F Ho CE Pesaran B Mitra PP A procedure for an automated measurement of song similarity Anim Behav 2000 59 1167 1176 10.1006/anbe.1999.1416 10877896 37. Tchernichovski O Sound Analysis Pro Users manual 2012 38. Salinas-Melgoza A. Dinámica espacio-temporal de individuos juveniles del loro corona lila (Amazona finschi) en el bosque seco de la costa de Jalisco. M.Sc. Thesis. Universidad Nacional Autónoma de México. 2003. http://132.248.9.195/ppt2002/0318559/Index.html. Accessed 10 Mar 2016. 39. Quinn GP Keough MJ Experimental design and data analysis for biologists 2002 Cambridge Cambridge University Press 40. Fournier D Skaug H Ancheta J Ianelli J Magnusson A Maunder MN Nielsen A Sibert J AD Model Builder: using automatic differentiation for statistical inference of highly parameterized complex nonlinear models Optim Methods Softw 2012 27 233 249 10.1080/10556788.2011.597854 41. Bates D, Maechler M, Bolker B, Walker S. lme4: Linear mixed-effects models using Eigen and S4. R package version 1.1-7. 2014. Available http://CRAN.R-project.org/package=lme4. Accessed 1 Jun 2016. 42. R Development Core Team. R: a language and environment for statistical computing. R Foundation for Statistical Computing, 2013. http://www.R-project.org. Accessed 10 Mar 2016. 43. Mundry R Sommer C Discriminant function analysis with nonindependent data: consequences and an alternative Anim Behav 2007 74 965 976 10.1016/j.anbehav.2006.12.028 44. Venables WN Ripley BD Modern Applied Statistics with S 2002 4 New York Springer 45. Renton K Agonistic interactions of nesting and nonbreeding macaws Condor 2004 106 354 362 10.1650/7388 46. Morton ES Ecological sources of selection on avian sounds Am Nat 1975 109 17 34 10.1086/282971 47. Hansen P Vocal learning: its role in adapting sound structures to long-distance propagation, and a hypothesis for its evolution Anim Behav 1979 27 1270 1271 10.1016/0003-3472(79)90073-3 48. Brown TJ Handford P Acoustic signal amplitude patterns: a computer simulation investigation of the acoustic adaptation hypothesis Condor 1996 98 608 623 10.2307/1369573 49. Forrest TG From sender to receiver: propagation and environmental effects on acoustic signals Am Zool 1994 34 644 654 10.1093/icb/34.6.644 50. Freeberg TM Dunbar RIM Ord TJ Social complexity as a proximate and ultimate factor in communicative complexity Phil Trans R Soc B 2012 367 1785 1801 10.1098/rstb.2011.0213 22641818 51. Krams I Krams T Freeberg TM Kullberg C Lucas JR Linking social complexity and vocal complexity: a parid perspective Phil Trans R Soc B 2012 367 1879 1891 10.1098/rstb.2011.0222 22641826 52. Freeberg TM Krams I Does social complexity link vocal complexity and cooperation? J Ornithol 2015 53. Renton K Salinas-Melgoza A Noguera FA Vega JH García AN Quesada M Amazona finschi (Sclater 1864) (Loro corona lila) Historia natural de Chamela 2002 Mexico City Instituto de Biología, Universidad Nacional Autónoma de México 343 344 54. Brereton JLG Esser A Inter-animal control of space Behaviour and environment: the use of space by animals and man 1971 New York Plenum Press 69 91 55. Bradbury JW Balsby TJS The functions of vocal learning in parrots Behav Ecol Sociobiol 2016 56. Renton K Salinas-Melgoza A De Labra-Hernández MA de la Parra-Martínez SM Resource requirements of parrots: nest site selectivity and dietary plasticity of Psittaciformes J Ornithol 2015 156 S73 S90 10.1007/s10336-015-1255-9 57. Renton K Lilac-crowned Parrot diet and food resource availability: resource tracking by a parrot seed predator Condor 2001 103 62 69 10.1650/0010-5422(2001)103[0062:LCPDAF]2.0.CO;2 58. Wright TF Regional dialects in the contact call of a parrot P Roy Soc B: Biol Sci 1996 263 867 872 10.1098/rspb.1996.0128 59. Kleeman PM Gilardi JD Geographic variation of St Lucia Parrot flight vocalizations Condor 2005 107 62 68 10.1650/7568 60. Salinas-Melgoza A Wright TF Evidence for vocal learning and limited dispersal as dual mechanisms for dialect maintenance in a parrot PLoS ONE 2012 7 e48667 10.1371/journal.pone.0048667 23139809 61. Read AF Weary DM The evolution of bird song: comparative analyses Philos T Roy Soc B 1992 338 165 187 10.1098/rstb.1992.0137 62. Cortopassi KA Bradbury JW Contact call diversity in wild Orange-fronted Parakeet pairs, Aratinga canicularis Anim behave 2006 71 1141 1154 10.1016/j.anbehav.2005.09.011 63. Bradbury JW Cortopassi KA Clemmons JR Geographic variation in the contact calls of Orange-fronted Parrots Auk 2001 118 958 972 10.1642/0004-8038(2001)118[0958:GVITCC]2.0.CO;2 64. Brightsmith DJ Use of arboreal termitaria by nesting birds in the Peruvian Amazon Condor 2000 102 529 538 10.1650/0010-5422(2000)102[0529:UOATBN]2.0.CO;2 65. Sanchez Martinez TC Renton K Availability and selection of arboreal termitaria as nest-sites by Orange-fronted Parakeets Aratinga canicularis in conserved and modified landscapes in Mexico Ibis 2009 151 311 320 10.1111/j.1474-919X.2009.00911.x 66. Soobramoney S Perrin M A comparison of the alarm calls of five species of African lovebirds: genus Agapornis T Roy Soc S Afr 2014 69 9 18 10.1080/0035919X.2014.890140 67. Wheatcroft D Repetition rate of calls used in multiple contexts communicates presence of predators to nestlings and adult birds Anim Behav 2015 103 35 44 10.1016/j.anbehav.2015.02.009 68. Krebs J Ashcroft R Webber M Song repertoires and territory defence in the Great Tit Nature 1978 271 539 542 10.1038/271539a0 69. Searcy WA Anderson RC Nowicki S Bird song as a signal of aggressive intent Behav Ecol Sociobiol 2006 60 234 241 10.1007/s00265-006-0161-9 70. Ballentine B Searcy WA Nowicki S Reliable aggressive signaling in swamp sparrows Anim Behav 2008 75 693 703 10.1016/j.anbehav.2007.07.025 71. Hof D Hazlett N Low-amplitude song predicts attack in a North American wood warbler Anim Behav 2010 80 821 828 10.1016/j.anbehav.2010.07.017 72. Botero CA Rivero JM Vehrencamp SL Relative threat and recognition ability in the response of tropical mockingbirds to song playback Anim Behav 2007 73 661 669 10.1016/j.anbehav.2006.10.007 18079978 73. Suzuki TN Communication about predator type by a bird using discrete, graded and combinatorial variation in alarm calls Anim Behav 2014 87 59 65 10.1016/j.anbehav.2013.10.009 74. Suzuki TN Wheatcroft D Griesser M Experimental evidence for compositional syntax in bird calls Nat Commun 2016 75. Hailman JP Ficken MS Ficken RW Constraints on the structure of combinatorial “Chick-a-dee” calls Ethology 1987 75 62 80 10.1111/j.1439-0310.1987.tb00642.x 76. Dahlin CR Wright TF Duets in Yellow-naped Amazons: variation in syntax, note composition and phonology at different levels of social organization Ethology 2009 115 857 871 10.1111/j.1439-0310.2009.01669.x 77. Hailman JP Griswold CK Syntax of Black-capped Chickadee (Parus atricapillus ) gargles sorts many types into few groups: implications for geographic variation, dialect drift, and vocal learning Bird Behav 1996 11 39 57 78. Clay Z Archbold J Zuberbühler K Functional flexibility in wild Bonobo vocal behavior PeerJ 2015 26290789